Compare commits

...

75 Commits

Author SHA1 Message Date
Thomas Pelletier dc72d75f3e Keep separate fn for []interface{} unmarshal 2021-11-13 19:20:20 -05:00
Thomas Pelletier f77775b59e Use less reflection when making slices
```
name                               old time/op    new time/op    delta
UnmarshalDataset/config-2            24.9ms ± 0%    24.6ms ± 0%  -1.09%  (p=0.029 n=4+4)
UnmarshalDataset/canada-2            61.7ms ± 1%    62.1ms ± 3%    ~     (p=1.000 n=5+5)
UnmarshalDataset/citm_catalog-2      24.7ms ± 1%    24.2ms ± 0%  -2.30%  (p=0.008 n=5+5)
UnmarshalDataset/twitter-2           10.9ms ± 2%    10.7ms ± 1%  -1.46%  (p=0.008 n=5+5)
UnmarshalDataset/code-2               108ms ± 0%     106ms ± 0%  -1.91%  (p=0.008 n=5+5)
UnmarshalDataset/example-2            176µs ± 0%     173µs ± 0%  -1.83%  (p=0.008 n=5+5)
Unmarshal/SimpleDocument/struct-2     586ns ± 1%     587ns ± 0%    ~     (p=0.690 n=5+5)
Unmarshal/SimpleDocument/map-2        876ns ± 0%     872ns ± 0%    ~     (p=0.095 n=5+5)
Unmarshal/ReferenceFile/struct-2     49.5µs ± 0%    49.5µs ± 0%    ~     (p=0.222 n=5+5)
Unmarshal/ReferenceFile/map-2        79.6µs ± 0%    79.1µs ± 0%  -0.62%  (p=0.008 n=5+5)
Unmarshal/HugoFrontMatter-2          13.7µs ± 0%    13.5µs ± 0%  -0.91%  (p=0.008 n=5+5)

name                               old speed      new speed      delta
UnmarshalDataset/config-2          42.2MB/s ± 0%  42.7MB/s ± 0%  +1.10%  (p=0.029 n=4+4)
UnmarshalDataset/canada-2          35.7MB/s ± 1%  35.5MB/s ± 3%    ~     (p=1.000 n=5+5)
UnmarshalDataset/citm_catalog-2    22.6MB/s ± 1%  23.1MB/s ± 0%  +2.36%  (p=0.008 n=5+5)
UnmarshalDataset/twitter-2         40.6MB/s ± 2%  41.2MB/s ± 1%  +1.47%  (p=0.008 n=5+5)
UnmarshalDataset/code-2            24.9MB/s ± 0%  25.4MB/s ± 0%  +1.95%  (p=0.008 n=5+5)
UnmarshalDataset/example-2         46.0MB/s ± 0%  46.9MB/s ± 0%  +1.86%  (p=0.008 n=5+5)
Unmarshal/SimpleDocument/struct-2  18.8MB/s ± 1%  18.7MB/s ± 0%    ~     (p=0.651 n=5+5)
Unmarshal/SimpleDocument/map-2     12.6MB/s ± 0%  12.6MB/s ± 0%    ~     (p=0.087 n=5+5)
Unmarshal/ReferenceFile/struct-2    106MB/s ± 0%   106MB/s ± 0%    ~     (p=0.222 n=5+5)
Unmarshal/ReferenceFile/map-2      65.8MB/s ± 0%  66.2MB/s ± 0%  +0.63%  (p=0.008 n=5+5)
Unmarshal/HugoFrontMatter-2        40.0MB/s ± 0%  40.3MB/s ± 0%  +0.92%  (p=0.008 n=5+5)

name                               old alloc/op   new alloc/op   delta
UnmarshalDataset/config-2            5.85MB ± 0%    5.85MB ± 0%    ~     (p=1.000 n=5+5)
UnmarshalDataset/canada-2            75.2MB ± 0%    75.2MB ± 0%    ~     (p=1.000 n=5+5)
UnmarshalDataset/citm_catalog-2      35.0MB ± 0%    35.0MB ± 0%    ~     (p=0.841 n=5+5)
UnmarshalDataset/twitter-2           13.5MB ± 0%    13.5MB ± 0%    ~     (p=0.548 n=5+5)
UnmarshalDataset/code-2              22.0MB ± 0%    22.0MB ± 0%    ~     (p=0.738 n=5+5)
UnmarshalDataset/example-2            203kB ± 0%     203kB ± 0%    ~     (p=0.714 n=5+5)
Unmarshal/SimpleDocument/struct-2      709B ± 0%      709B ± 0%    ~     (all equal)
Unmarshal/SimpleDocument/map-2       1.08kB ± 0%    1.08kB ± 0%    ~     (all equal)
Unmarshal/ReferenceFile/struct-2     19.7kB ± 0%    19.7kB ± 0%    ~     (all equal)
Unmarshal/ReferenceFile/map-2        37.0kB ± 0%    37.0kB ± 0%    ~     (p=0.333 n=4+5)
Unmarshal/HugoFrontMatter-2          7.22kB ± 0%    7.22kB ± 0%    ~     (all equal)

name                               old allocs/op  new allocs/op  delta
UnmarshalDataset/config-2              230k ± 0%      230k ± 0%    ~     (p=0.556 n=4+5)
UnmarshalDataset/canada-2              391k ± 0%      391k ± 0%    ~     (all equal)
UnmarshalDataset/citm_catalog-2        158k ± 0%      158k ± 0%    ~     (p=1.000 n=4+5)
UnmarshalDataset/twitter-2            54.7k ± 0%     54.7k ± 0%    ~     (p=1.000 n=4+5)
UnmarshalDataset/code-2               1.05M ± 0%     1.05M ± 0%    ~     (all equal)
UnmarshalDataset/example-2            1.28k ± 0%     1.28k ± 0%    ~     (all equal)
Unmarshal/SimpleDocument/struct-2      8.00 ± 0%      8.00 ± 0%    ~     (all equal)
Unmarshal/SimpleDocument/map-2         13.0 ± 0%      13.0 ± 0%    ~     (all equal)
Unmarshal/ReferenceFile/struct-2        123 ± 0%       123 ± 0%    ~     (all equal)
Unmarshal/ReferenceFile/map-2           590 ± 0%       590 ± 0%    ~     (all equal)
Unmarshal/HugoFrontMatter-2             130 ± 0%       130 ± 0%    ~     (all equal)
```
2021-11-13 19:20:20 -05:00
Thomas Pelletier b52f6c9823 Remove some allocs for slices in interfaces
```
name                               old time/op    new time/op    delta
UnmarshalDataset/config-2            24.9ms ± 1%    24.9ms ± 0%     ~     (p=0.413 n=5+4)
UnmarshalDataset/canada-2            66.1ms ± 0%    61.7ms ± 1%   -6.63%  (p=0.008 n=5+5)
UnmarshalDataset/citm_catalog-2      25.3ms ± 5%    24.7ms ± 1%   -2.09%  (p=0.032 n=5+5)
UnmarshalDataset/twitter-2           10.9ms ± 2%    10.9ms ± 2%     ~     (p=1.000 n=5+5)
UnmarshalDataset/code-2               108ms ± 0%     108ms ± 0%     ~     (p=0.095 n=5+5)
UnmarshalDataset/example-2            177µs ± 2%     176µs ± 0%     ~     (p=0.841 n=5+5)
Unmarshal/SimpleDocument/struct-2     579ns ± 0%     586ns ± 1%   +1.30%  (p=0.008 n=5+5)
Unmarshal/SimpleDocument/map-2        875ns ± 1%     876ns ± 0%     ~     (p=0.548 n=5+5)
Unmarshal/ReferenceFile/struct-2     49.7µs ± 1%    49.5µs ± 0%     ~     (p=0.095 n=5+5)
Unmarshal/ReferenceFile/map-2        80.4µs ± 0%    79.6µs ± 0%   -0.99%  (p=0.008 n=5+5)
Unmarshal/HugoFrontMatter-2          13.9µs ± 0%    13.7µs ± 0%   -1.70%  (p=0.008 n=5+5)

name                               old speed      new speed      delta
UnmarshalDataset/config-2          42.1MB/s ± 1%  42.2MB/s ± 0%     ~     (p=0.381 n=5+4)
UnmarshalDataset/canada-2          33.3MB/s ± 0%  35.7MB/s ± 1%   +7.11%  (p=0.008 n=5+5)
UnmarshalDataset/citm_catalog-2    22.1MB/s ± 5%  22.6MB/s ± 1%   +2.08%  (p=0.032 n=5+5)
UnmarshalDataset/twitter-2         40.7MB/s ± 2%  40.6MB/s ± 2%     ~     (p=1.000 n=5+5)
UnmarshalDataset/code-2            24.8MB/s ± 0%  24.9MB/s ± 0%     ~     (p=0.103 n=5+5)
UnmarshalDataset/example-2         45.8MB/s ± 2%  46.0MB/s ± 0%     ~     (p=0.841 n=5+5)
Unmarshal/SimpleDocument/struct-2  19.0MB/s ± 0%  18.8MB/s ± 1%   -1.26%  (p=0.008 n=5+5)
Unmarshal/SimpleDocument/map-2     12.6MB/s ± 1%  12.6MB/s ± 0%     ~     (p=0.508 n=5+5)
Unmarshal/ReferenceFile/struct-2    105MB/s ± 1%   106MB/s ± 0%     ~     (p=0.095 n=5+5)
Unmarshal/ReferenceFile/map-2      65.2MB/s ± 0%  65.8MB/s ± 0%   +1.00%  (p=0.008 n=5+5)
Unmarshal/HugoFrontMatter-2        39.3MB/s ± 0%  40.0MB/s ± 0%   +1.73%  (p=0.008 n=5+5)

name                               old alloc/op   new alloc/op   delta
UnmarshalDataset/config-2            5.85MB ± 0%    5.85MB ± 0%   -0.00%  (p=0.008 n=5+5)
UnmarshalDataset/canada-2            76.6MB ± 0%    75.2MB ± 0%   -1.76%  (p=0.016 n=4+5)
UnmarshalDataset/citm_catalog-2      35.3MB ± 0%    35.0MB ± 0%   -0.71%  (p=0.008 n=5+5)
UnmarshalDataset/twitter-2           13.5MB ± 0%    13.5MB ± 0%   -0.19%  (p=0.016 n=4+5)
UnmarshalDataset/code-2              22.3MB ± 0%    22.0MB ± 0%   -1.31%  (p=0.008 n=5+5)
UnmarshalDataset/example-2            204kB ± 0%     203kB ± 0%   -0.34%  (p=0.008 n=5+5)
Unmarshal/SimpleDocument/struct-2      709B ± 0%      709B ± 0%     ~     (all equal)
Unmarshal/SimpleDocument/map-2       1.08kB ± 0%    1.08kB ± 0%     ~     (all equal)
Unmarshal/ReferenceFile/struct-2     19.8kB ± 0%    19.7kB ± 0%   -0.24%  (p=0.008 n=5+5)
Unmarshal/ReferenceFile/map-2        37.3kB ± 0%    37.0kB ± 0%   -0.64%  (p=0.029 n=4+4)
Unmarshal/HugoFrontMatter-2          7.26kB ± 0%    7.22kB ± 0%   -0.66%  (p=0.008 n=5+5)

name                               old allocs/op  new allocs/op  delta
UnmarshalDataset/config-2              230k ± 0%      230k ± 0%   -0.00%  (p=0.000 n=5+4)
UnmarshalDataset/canada-2              447k ± 0%      391k ± 0%  -12.53%  (p=0.008 n=5+5)
UnmarshalDataset/citm_catalog-2        169k ± 0%      158k ± 0%   -6.20%  (p=0.029 n=4+4)
UnmarshalDataset/twitter-2            55.8k ± 0%     54.7k ± 0%   -1.88%  (p=0.029 n=4+4)
UnmarshalDataset/code-2               1.06M ± 0%     1.05M ± 0%   -1.14%  (p=0.008 n=5+5)
UnmarshalDataset/example-2            1.31k ± 0%     1.28k ± 0%   -2.21%  (p=0.008 n=5+5)
Unmarshal/SimpleDocument/struct-2      8.00 ± 0%      8.00 ± 0%     ~     (all equal)
Unmarshal/SimpleDocument/map-2         13.0 ± 0%      13.0 ± 0%     ~     (all equal)
Unmarshal/ReferenceFile/struct-2        125 ± 0%       123 ± 0%   -1.60%  (p=0.008 n=5+5)
Unmarshal/ReferenceFile/map-2           600 ± 0%       590 ± 0%   -1.67%  (p=0.008 n=5+5)
Unmarshal/HugoFrontMatter-2             132 ± 0%       130 ± 0%   -1.52%  (p=0.008 n=5+5)
```
2021-11-13 19:20:20 -05:00
Thomas Pelletier 12244064bb Use global cache to unmarshal all slice types 2021-11-13 19:20:20 -05:00
Thomas Pelletier 6430ee0bfa Generic slice unmarshal fn 2021-11-13 19:20:20 -05:00
Thomas Pelletier cf530eba46 Specialize array unmarshal into []interface{} 2021-11-13 19:20:19 -05:00
Thomas Pelletier 64fe47161f API: Encoder and Decoder options are chainable (#670)
Fixes #583
2021-11-13 19:04:53 -05:00
Thomas Pelletier 4dff8eaa4d Decoder: prevent duplicates of inline tables (#667)
* seen: prevent duplicates of inline tables

* Provide clearer error message for redefined keys

For example:

``
toml: key b is already defined
```
2021-11-10 10:04:43 -05:00
Cameron Moore 2dbd29a565 parser: Fix missing check for upper exponent (#665) 2021-11-09 21:15:23 -05:00
Thomas Pelletier f27a07d31a seen: verify arrays (#663)
Fixes #662
2021-11-09 20:26:30 -05:00
Thomas Pelletier 644515958c Update TOML test suite (#661)
Ref #658
2021-11-08 22:35:35 -05:00
Thomas Pelletier 8683be35f6 seen: check inline tables (#660)
Fixes #658
2021-11-08 21:53:02 -05:00
Thomas Pelletier dc1740d473 Decode: code cleanup for struct cache (#659) 2021-11-07 18:35:30 -05:00
Thomas Pelletier 11f789ef11 Decode: prevent comments that look like dates to be accepted (#657)
* parser: fix date detection

When the parser has to decide between parsing and integer or a date, it should
check that all characters are actually acceptable (digits, or date/time
elements).

Fixes #655
2021-11-04 22:06:12 -04:00
Thomas Pelletier 74d21b367f scanner: handle carriage return in comments (#656)
Fixes #653
2021-11-04 21:40:16 -04:00
Thomas Pelletier 6617e7e73d utf8: use lookup table to validate ASCII (#654) 2021-11-04 16:05:36 -04:00
Thomas Pelletier 3dbca20bc9 Decoder: flag invalid carriage returns in strings (#652)
Fixes #651
2021-11-02 10:02:25 -04:00
Thomas Pelletier 85c0658984 Decode: add missing checks for LocalTime (#650) 2021-10-29 22:13:08 -04:00
Thomas Pelletier 772d169b52 testsuite: return error when can't encode tag (#648) 2021-10-29 21:51:50 -04:00
Cameron Moore b4ec220f7e Update tomltestgen and regenerate tests (#645)
Remove testsuite build tag from generated tests file

Co-authored-by: Thomas Pelletier <thomas@pelletier.codes>
2021-10-28 20:46:08 -04:00
Thomas Pelletier 3694ae88f6 decode: error on _ before exponent in floats (#647)
Fixes #646
2021-10-28 20:41:10 -04:00
Thomas Pelletier 19751e8a51 Missing performance section 2021-10-28 19:10:43 -04:00
Thomas Pelletier 925f214125 Add GitHub release configuration (#644) 2021-10-28 19:06:14 -04:00
Thomas Pelletier 39f893ad99 Multiline strings fixes (#643)
* scanner: allow multiline strings to end with "" or ''

* parser: trim all whitespaces after \ in multiline
2021-10-28 18:26:34 -04:00
Thomas Pelletier c871a61015 unmarshal: use UnmarshalText for any type (#642)
Not only structs can implement TextUnmarshaler.

Fixes #564
2021-10-28 17:02:47 -04:00
Thomas Pelletier d0d001625c unmarshal: don't panic when storing table in slice (#641)
New error message:

```
toml: cannot store a table in a slice
1| [things]
 |  ~~~~~~ cannot store a table in a slice
2| foo = "bar"
```

Fixes #623
2021-10-25 16:47:10 -04:00
Thomas Pelletier 64941b99e2 unmarshal: empty document results in map (#640)
Fixes #602
2021-10-25 15:55:54 -04:00
Thomas Pelletier ed02a1f192 seen: check for explicit tables on dotted keys (#639)
The TOML spec is being clarified to say that dotted keys "define" their
intermediate tables. Therefore the seen tracker needs to verify that none of
them reference an explicit table.

Also added a missing seen expression check for key-values parsed as part of a
table section.

See https://github.com/toml-lang/toml/issues/846
2021-10-22 23:25:28 -04:00
Thomas Pelletier 4d7c9ddac7 Floats and integers parsing fixes (#638)
* parser: fix scan of float with exp but no decimal
* decoder: validate leading zeros for decimals
2021-10-22 22:25:56 -04:00
Thomas Pelletier feb1830dcc tomltest: enable TestTOMLTest_Valid_Comment_Tricky 2021-10-21 22:30:58 -04:00
Thomas Pelletier 1c33d6ce20 tomltest: custom comparison functions (#637) 2021-10-21 22:29:04 -04:00
Thomas Pelletier 3000471a12 parser: improve floats validation (#636) 2021-10-20 08:49:28 -04:00
Thomas Pelletier 1f33a6a476 tomltest: enable Valid_Datetime_LocalTime 2021-10-19 21:20:45 -04:00
Thomas Pelletier 2700aad5d2 tomltest: run UTF8 tests (#634) 2021-10-19 16:00:56 -04:00
Thomas Pelletier 7ccaa2744e tomltest: unmarshal JSONs for tests (#633)
Comparing the output and the expected results byte-wise means we get false
negative when order doesn't matter (for example the ValidTableKeyword test).
2021-10-19 15:29:49 -04:00
Johanan Idicula df4bb061f8 time: follow RFC3339 spec for datetime (#632) 2021-10-18 09:56:07 -04:00
Thomas Pelletier 9e81ce1c33 Create SECURITY.md 2021-10-17 20:57:34 -04:00
Cameron Moore a23850f29b decode: preserve nanosecond precision when decoding time (#626)
Co-authored-by: Thomas Pelletier <thomas@pelletier.codes>
2021-10-17 20:43:29 -04:00
Johanan Idicula 76f53c857b unmarshal: validate date (#622) 2021-10-17 20:18:20 -04:00
Thomas Pelletier 85f5d567e4 parser: validate invalid ASCII control characters 2021-10-16 07:41:12 -04:00
Thomas Pelletier bd5cba0b0b Update benchmarks readme (#630)
* Fix ci.sh for new benchmarks

Nice + taskset are more stable on my machine. We want to excude non high-level
benchmarks. BurntSushi/toml now supports canada.toml.

* Update latest benchmarks in README
2021-10-15 19:53:40 -04:00
Thomas Pelletier cd54472d03 Validate UTF-8 (#629) 2021-10-15 19:13:21 -04:00
Thomas Pelletier cc0d1a90ff testgen: skip currently failing tests (#627) 2021-10-14 11:14:44 -04:00
Sterling Hanenkamp 4984dcb5e9 encode: ensure floats have decimal point (#615)
Fixes #571

Co-authored-by: Sterling Hanenkamp <sterling@ziprecruiter.com>
2021-10-14 08:34:54 -04:00
jidicula 86632bc190 parser: fail when missing array separator (#616)
Co-authored-by: Thomas Pelletier <thomas@pelletier.codes>
2021-10-14 08:26:29 -04:00
Cameron Moore d25eec183f gotoml-test-decoder: add toml-test decoder command (#619) 2021-10-14 08:14:34 -04:00
Riya John e96746311c decoder: fix panic date time should have a timezone (#614)
Fixes #596

Co-authored-by: Thomas Pelletier <thomas@pelletier.codes>
2021-10-06 21:24:25 -04:00
Cameron Moore 62acca2b68 tomltestgen: add toml-test unit test generation command (#610)
Tests are hidden behind a "testsuite" build tag for now since many tests
are failing.  Use `go test -tags testsuite` to activate.

Use `go generate` to regenerate toml_testgen_test.go.

Co-authored-by: Thomas Pelletier <thomas@pelletier.codes>
2021-10-03 22:15:30 -04:00
Cameron Moore 476492a85c unmarshal: support lowercase 'T' and 'Z' in date-time parsing (#601)
RFC3399 allows for lowercase 't' and 'z' in date-time values.

Fixes #600
2021-09-25 10:02:23 -07:00
Thomas Pelletier ee9b902222 unmarshal: convert ints if target type is compatible (#594)
This is required to support custom types.

Fixes #590
2021-09-09 21:25:14 -04:00
Thomas Pelletier fa56f48daf parser: don't overflow when parsing bad times (#593)
Fixes #585
2021-09-09 11:59:37 -04:00
Thomas Pelletier f34c9c332f scanner: fix error reporting for last comments (#591)
When an invalid TOML expression ends with a comment before the end of
file, the decode error would take a nil from scanComment, which is not
part of the document.

Fixes #588
2021-09-08 21:54:30 -04:00
Thomas Pelletier a0d685d482 unmarshal: don't crash on unterminated inline table (#587)
Fixes #586
2021-09-07 20:08:59 -04:00
Thomas Pelletier 4a5ae9e81e errors: fix context generation with only one line 2021-09-07 10:36:22 -04:00
Thomas Pelletier 7e2fa1bc80 unmarshal: fix non-terminated array error
Fixes #581
2021-09-07 10:36:22 -04:00
Thomas Pelletier 40cfb6f458 parser: don't crash on unterminated table key (#580)
* parser: don't crash on unterminated table key

Fixes #579

* parser: fix format of error returned by expect

EOF was missing the format string and %U is not very human friendly.
2021-09-06 12:18:45 -04:00
Thomas Pelletier 1230ca485e unmarshal: make copy of non addressable values (#576)
When unmarshaling into a nested struct in a map, the value is not
addressable. In that case, make a copy of it and modify it instead.

Fixes #575
2021-08-31 20:22:38 -04:00
Thomas Pelletier 69ab7e10d1 Go 1.17 release (#574)
Minimum supported version: Go 1.16.
2021-08-17 09:43:52 -04:00
Thomas Pelletier fa07960695 Add installation instructions (#572) 2021-07-27 18:12:44 -04:00
kkHAIKE 8be357dfa1 Add LocalTime to interface{} decode support (#567)
Co-authored-by: Thomas Pelletier <thomas@pelletier.codes>
2021-07-21 17:50:12 +02:00
kkHAIKE a93b34d984 Unicode parsing optimization (#568)
Inline call to hexToRune and uses specialized parsing, as found in encoding/json.

Co-authored-by: Thomas Pelletier <thomas@pelletier.codes>
2021-07-21 10:50:03 +02:00
Matthieu MOREL 9c24fbeaad Set up Dependabot for GitHub actions and docker (#570) 2021-07-20 16:54:26 +02:00
Thomas Pelletier f6b38c33b7 Provide own implementation of Local* (#558)
* Reduces the public API.
* Reuses optimized parsing functions.
* Removes reliance on Google code under Apache license.
2021-06-08 20:27:05 -04:00
Thomas Pelletier 773f10110c Unmarshal recursive structs (#557)
Co-authored-by: Nabetani <takenori@nabetani.sakura.ne.jp>
2021-06-08 14:22:39 -04:00
Thomas Pelletier 618f0181ac AST Tweaks (#551)
* Use pointers instead of copying around ast.Node

Node is a 56B struct that is constantly in the hot path. Passing nodes
around by copy had a cost that started to add up. This change replaces
them by pointers. Using unsafe pointer arithmetic and converting
sibling/child indexes to relative offsets, it removes the need to carry
around a pointer to the root of the tree. This saves 8B per Node. This
space will be used to store an extra []byte slice to provide contextual
error handling on all nodes, including the ones whose data is different
than the raw input (for example: strings with escaped characters), while
staying under the size of a cache line.

* Remove conditional

* Add Raw to track range in data for parsed values

* Simplify reference tracking
2021-06-03 21:48:51 -04:00
Thomas Pelletier f3bb20ea79 Benchmark marshal (#550) 2021-06-02 09:29:19 -04:00
Thomas Pelletier b0d6c62255 Don't use bytes.Buffer when not necessary (#549)
When parsing strings, they can be referenced directly from the document
when they don't contain escaped characters. This avoids paying to cost
of allocating (and sometimes growing) the bytes buffer unecessarily.
2021-06-01 09:51:59 -04:00
Thomas Pelletier b202375414 Add benchmarks results to readme (#548) 2021-06-01 09:10:17 -04:00
Thomas Pelletier 250e073408 Stack-based unmarshaler (#546)
* Benchmark script

* Rewrite unmarshaler using the stack

Instead of tracking the build chain using `target`s, use the stack
instead.

Working and most benchmarks look good, but regression on structs unmarshalling.

~60% slower on ReferenceFile/struct.

* Shortcut to check if last node of iterator

* Remove unecessary pointer allocation

* Skip over unused keys without marking them as seen

* Add some tests

* Fix mktemp on macos
2021-05-31 12:14:13 -04:00
Thomas Pelletier 11f022ab09 Fix parser skipping over the whole file (#547)
* Add test for ReferenceFile/struct
* Stop skipping after table
* Add .gitattributes to force LF encoding
* Fix the reference file
2021-05-30 18:53:07 -04:00
Thomas Pelletier 840df4a229 seen tracker: use array for storage (#545)
name                              old time/op    new time/op    delta
UnmarshalDataset/config-32          81.2ms ± 3%    77.8ms ± 3%   -4.25%  (p=0.000 n=20+19)
UnmarshalDataset/canada-32           104ms ± 5%     105ms ± 4%     ~     (p=0.102 n=20+20)
UnmarshalDataset/citm_catalog-32    57.5ms ± 5%    59.0ms ± 5%   +2.54%  (p=0.033 n=20+20)
UnmarshalDataset/twitter-32         25.7ms ± 7%    28.1ms ± 5%   +9.33%  (p=0.000 n=20+20)
UnmarshalDataset/code-32             305ms ± 6%     292ms ± 5%   -4.29%  (p=0.000 n=20+19)
UnmarshalDataset/example-32          519µs ± 6%     522µs ± 5%     ~     (p=0.659 n=20+20)
UnmarshalSimple/struct-32           1.44µs ± 1%    1.17µs ± 6%  -18.78%  (p=0.000 n=14+20)
UnmarshalSimple/map-32              2.30µs ± 4%    1.99µs ± 4%  -13.65%  (p=0.000 n=20+19)
ReferenceFile/struct-32             44.1µs ± 4%    38.1µs ± 5%  -13.61%  (p=0.000 n=18+20)
ReferenceFile/map-32                 197µs ± 7%     189µs ± 5%   -3.91%  (p=0.000 n=20+20)
HugoFrontMatter-32                  45.9µs ± 6%    39.3µs ± 6%  -14.46%  (p=0.000 n=19+20)

name                              old speed      new speed      delta
UnmarshalDataset/config-32        12.9MB/s ± 3%  13.5MB/s ± 3%   +4.42%  (p=0.000 n=20+19)
UnmarshalDataset/canada-32        21.1MB/s ± 5%  20.9MB/s ± 4%     ~     (p=0.101 n=20+20)
UnmarshalDataset/citm_catalog-32  9.72MB/s ± 6%  9.47MB/s ± 5%   -2.53%  (p=0.031 n=20+20)
UnmarshalDataset/twitter-32       17.2MB/s ± 7%  15.8MB/s ± 5%   -8.57%  (p=0.000 n=20+20)
UnmarshalDataset/code-32          8.81MB/s ± 7%  9.20MB/s ± 5%   +4.47%  (p=0.000 n=20+19)
UnmarshalDataset/example-32       15.6MB/s ± 6%  15.5MB/s ± 5%     ~     (p=0.644 n=20+20)
UnmarshalSimple/struct-32         7.61MB/s ± 1%  9.39MB/s ± 7%  +23.33%  (p=0.000 n=15+20)
UnmarshalSimple/map-32            4.78MB/s ± 4%  5.54MB/s ± 5%  +15.85%  (p=0.000 n=20+19)
ReferenceFile/struct-32            119MB/s ± 4%   138MB/s ± 5%  +15.79%  (p=0.000 n=18+20)
ReferenceFile/map-32              26.6MB/s ± 7%  27.7MB/s ± 5%   +4.06%  (p=0.000 n=20+20)
HugoFrontMatter-32                11.9MB/s ± 6%  13.9MB/s ± 6%  +16.91%  (p=0.000 n=19+20)

name                              old alloc/op   new alloc/op   delta
UnmarshalDataset/config-32          16.9MB ± 0%    13.9MB ± 0%  -17.48%  (p=0.000 n=18+18)
UnmarshalDataset/canada-32          74.3MB ± 0%    74.3MB ± 0%   -0.00%  (p=0.000 n=20+20)
UnmarshalDataset/citm_catalog-32    37.3MB ± 0%    37.3MB ± 0%   +0.11%  (p=0.000 n=20+20)
UnmarshalDataset/twitter-32         15.6MB ± 0%    15.6MB ± 0%     ~     (p=0.211 n=19+20)
UnmarshalDataset/code-32            59.5MB ± 0%    52.4MB ± 0%  -11.96%  (p=0.000 n=19+20)
UnmarshalDataset/example-32          238kB ± 0%     239kB ± 0%   +0.02%  (p=0.000 n=18+20)
UnmarshalSimple/struct-32             981B ± 0%      709B ± 0%  -27.73%  (p=0.000 n=20+20)
UnmarshalSimple/map-32              1.45kB ± 0%    1.17kB ± 0%  -18.82%  (p=0.000 n=20+20)
ReferenceFile/struct-32             11.8kB ± 0%     9.7kB ± 0%  -17.64%  (p=0.000 n=20+20)
ReferenceFile/map-32                51.5kB ± 0%    52.2kB ± 0%   +1.30%  (p=0.000 n=20+17)
HugoFrontMatter-32                  12.1kB ± 0%    11.1kB ± 0%   -7.97%  (p=0.000 n=20+19)

name                              old allocs/op  new allocs/op  delta
UnmarshalDataset/config-32            645k ± 0%      557k ± 0%  -13.76%  (p=0.000 n=20+16)
UnmarshalDataset/canada-32            896k ± 0%      896k ± 0%   -0.00%  (p=0.000 n=20+20)
UnmarshalDataset/citm_catalog-32      380k ± 0%      377k ± 0%   -0.75%  (p=0.000 n=19+20)
UnmarshalDataset/twitter-32           158k ± 0%      158k ± 0%   -0.01%  (p=0.000 n=18+18)
UnmarshalDataset/code-32             2.92M ± 0%     2.57M ± 0%  -11.87%  (p=0.000 n=20+20)
UnmarshalDataset/example-32          3.66k ± 0%     3.64k ± 0%   -0.63%  (p=0.000 n=20+20)
UnmarshalSimple/struct-32             13.0 ± 0%      10.0 ± 0%  -23.08%  (p=0.000 n=20+20)
UnmarshalSimple/map-32                22.0 ± 0%      19.0 ± 0%  -13.64%  (p=0.000 n=20+20)
ReferenceFile/struct-32                253 ± 0%       155 ± 0%  -38.74%  (p=0.000 n=20+20)
ReferenceFile/map-32                 1.67k ± 0%     1.44k ± 0%  -14.23%  (p=0.000 n=20+20)
HugoFrontMatter-32                     357 ± 0%       313 ± 0%  -12.32%  (p=0.000 n=20+20)
2021-05-26 18:47:00 -04:00
Thomas Pelletier c2d1fd86e5 Fix timezone detection when time has fractional component (#544) 2021-05-21 09:37:43 -04:00
Thomas Pelletier 238a6fef7d Add links to proper feedback channels 2021-05-15 08:55:07 -04:00
Thomas Pelletier 67852cf007 Clarify default struct tag being unsupported (#543)
Follow up to https://github.com/pelletier/go-toml/issues/542
2021-05-15 08:49:15 -04:00
Thomas Pelletier d276c42adc Run coverage test on branches only 2021-05-10 20:22:12 -04:00
53 changed files with 6915 additions and 3519 deletions
+3
View File
@@ -0,0 +1,3 @@
* text=auto
benchmark/benchmark.toml text eol=lf
+15 -4
View File
@@ -1,6 +1,17 @@
version: 2
updates:
- package-ecosystem: "gomod"
directory: "/" # Location of package manifests
schedule:
interval: "daily"
- package-ecosystem: gomod
directory: /
schedule:
interval: daily
open-pull-requests-limit: 10
- package-ecosystem: github-actions
directory: /
schedule:
interval: daily
open-pull-requests-limit: 10
- package-ecosystem: docker
directory: /
schedule:
interval: daily
open-pull-requests-limit: 10
+20
View File
@@ -0,0 +1,20 @@
changelog:
exclude:
labels:
- build
categories:
- title: What's new
labels:
- feature
- title: Performance
labels:
- performance
- title: Fixed bugs
labels:
- bug
- title: Documentation
labels:
- doc
- title: Other changes
labels:
- "*"
+1 -4
View File
@@ -1,15 +1,12 @@
name: coverage
on:
push:
branches:
- v2
pull_request:
branches:
- v2
jobs:
report:
runs-on: 'ubuntu-latest'
runs-on: "ubuntu-latest"
name: report
steps:
- uses: actions/checkout@master
+1 -1
View File
@@ -12,7 +12,7 @@ jobs:
strategy:
matrix:
os: [ 'ubuntu-latest', 'windows-latest', 'macos-latest']
go: [ '1.15', '1.16' ]
go: [ '1.16', '1.17' ]
runs-on: ${{ matrix.os }}
name: ${{ matrix.go }}/${{ matrix.os }}
steps:
+1 -1
View File
@@ -60,7 +60,7 @@ enable = [
# "nlreturn",
"noctx",
"nolintlint",
"paralleltest",
#"paralleltest",
"prealloc",
"predeclared",
"revive",
+85 -10
View File
@@ -4,7 +4,6 @@ Go library for the [TOML](https://toml.io/en/) format.
This library supports [TOML v1.0.0](https://toml.io/en/v1.0.0).
## Development status
This is the upcoming major version of go-toml. It is currently in active
@@ -14,8 +13,11 @@ with v1, and fixes a lot known bugs and performance issues along the way.
If you do not need the advanced document editing features of v1, you are
encouraged to try out this version.
👉 [Roadmap for v2](https://github.com/pelletier/go-toml/discussions/506).
[👉 Roadmap for v2](https://github.com/pelletier/go-toml/discussions/506)
[🐞 Bug Reports](https://github.com/pelletier/go-toml/issues)
[💬 Anything else](https://github.com/pelletier/go-toml/discussions)
## Documentation
@@ -23,13 +25,14 @@ Full API, examples, and implementation notes are available in the Go documentati
[![Go Reference](https://pkg.go.dev/badge/github.com/pelletier/go-toml/v2.svg)](https://pkg.go.dev/github.com/pelletier/go-toml/v2)
## Import
```go
import "github.com/pelletier/go-toml/v2"
```
See [Modules](#Modules).
## Features
### Stdlib behavior
@@ -40,7 +43,7 @@ standard library's `encoding/json`.
### Performance
While go-toml favors usability, it is written with performance in mind. Most
operations should not be shockingly slow.
operations should not be shockingly slow. See [benchmarks](#benchmarks).
### Strict mode
@@ -146,6 +149,63 @@ fmt.Println(string(b))
[marshal]: https://pkg.go.dev/github.com/pelletier/go-toml/v2#Marshal
## Benchmarks
Execution time speedup compared to other Go TOML libraries:
<table>
<thead>
<tr><th>Benchmark</th><th>go-toml v1</th><th>BurntSushi/toml</th></tr>
</thead>
<tbody>
<tr><td>Marshal/HugoFrontMatter-2</td><td>1.9x</td><td>1.9x</td></tr>
<tr><td>Marshal/ReferenceFile/map-2</td><td>1.7x</td><td>1.9x</td></tr>
<tr><td>Marshal/ReferenceFile/struct-2</td><td>2.4x</td><td>2.6x</td></tr>
<tr><td>Unmarshal/HugoFrontMatter-2</td><td>2.9x</td><td>2.5x</td></tr>
<tr><td>Unmarshal/ReferenceFile/map-2</td><td>2.7x</td><td>2.6x</td></tr>
<tr><td>Unmarshal/ReferenceFile/struct-2</td><td>4.8x</td><td>5.1x</td></tr>
</tbody>
</table>
<details><summary>See more</summary>
<p>The table above has the results of the most common use-cases. The table below
contains the results of all benchmarks, including unrealistic ones. It is
provided for completeness.</p>
<table>
<thead>
<tr><th>Benchmark</th><th>go-toml v1</th><th>BurntSushi/toml</th></tr>
</thead>
<tbody>
<tr><td>Marshal/SimpleDocument/map-2</td><td>1.7x</td><td>2.1x</td></tr>
<tr><td>Marshal/SimpleDocument/struct-2</td><td>2.5x</td><td>2.8x</td></tr>
<tr><td>Unmarshal/SimpleDocument/map-2</td><td>4.1x</td><td>3.1x</td></tr>
<tr><td>Unmarshal/SimpleDocument/struct-2</td><td>6.4x</td><td>4.3x</td></tr>
<tr><td>UnmarshalDataset/example-2</td><td>3.4x</td><td>3.2x</td></tr>
<tr><td>UnmarshalDataset/code-2</td><td>2.2x</td><td>2.5x</td></tr>
<tr><td>UnmarshalDataset/twitter-2</td><td>2.8x</td><td>2.7x</td></tr>
<tr><td>UnmarshalDataset/citm_catalog-2</td><td>2.2x</td><td>2.0x</td></tr>
<tr><td>UnmarshalDataset/canada-2</td><td>1.8x</td><td>1.4x</td></tr>
<tr><td>UnmarshalDataset/config-2</td><td>4.4x</td><td>2.9x</td></tr>
<tr><td>[Geo mean]</td><td>2.8x</td><td>2.6x</td></tr>
</tbody>
</table>
<p>This table can be generated with <code>./ci.sh benchmark -a -html</code>.</p>
</details>
## Modules
go-toml uses Go's standard modules system.
Installation instructions:
- Go ≥ 1.16: Nothing to do. Use the import in your code. The `go` command deals
with it automatically.
- Go ≥ 1.13: `GO111MODULE=on go get github.com/pelletier/go-toml/v2`.
In case of trouble: [Go Modules FAQ][mod-faq].
[mod-faq]: https://github.com/golang/go/wiki/Modules#why-does-installing-a-tool-via-go-get-fail-with-error-cannot-find-main-module
## Migrating from v1
This section describes the differences between v1 and v2, with some pointers on
@@ -187,7 +247,7 @@ d := doc{
}
data := `
[A]
[A]
B = "After"
`
@@ -248,6 +308,22 @@ This method was not widely used, poorly defined, and added a lot of complexity.
A similar effect can be achieved by implementing the `encoding.TextUnmarshaler`
interface and use strings.
#### Support for `default` struct tag has been dropped
This feature adds complexity and a poorly defined API for an effect that can be
accomplished outside of the library.
It does not seem like other format parsers in Go support that feature (the
project referenced in the original ticket #202 has not been updated since 2017).
Given that go-toml v2 should not touch values not in the document, the same
effect can be achieved by pre-filling the struct with defaults (libraries like
[go-defaults][go-defaults] can help). Also, string representation is not well
defined for all types: it creates issues like #278.
The recommended replacement is pre-filling the struct before unmarshaling.
[go-defaults]: https://github.com/mcuadros/go-defaults
### Encoding / Marshal
#### Default struct fields order
@@ -290,7 +366,6 @@ manually sort the fields alphabetically in the struct definition.
V1 automatically indents content of tables by default. V2 does not. However the
same behavior can be obtained using [`Encoder.SetIndentTables`][sit]. For example:
```go
data := map[string]interface{}{
"table": map[string]string{
@@ -312,15 +387,15 @@ fmt.Println("v2 Encoder:\n" + string(buf.Bytes()))
// Output:
// v1:
//
//
// [table]
// key = "value"
//
//
// v2:
// [table]
// key = 'value'
//
//
//
//
// v2 Encoder:
// [table]
// key = 'value'
+19
View File
@@ -0,0 +1,19 @@
# Security Policy
## Supported Versions
Use this section to tell people about which versions of your project are
currently being supported with security updates.
| Version | Supported |
| ---------- | ------------------ |
| Latest 2.x | :white_check_mark: |
| All 1.x | :x: |
| All 0.x | :x: |
## Reporting a Vulnerability
Email a vulnerability report to `security@pelletier.codes`. Make sure to include
as many details as possible to reproduce the vulnerability. This is a
side-project: I will try to get back to you as quickly as possible, time
permitting in my personal life. Providing a working patch helps very much!
+16 -17
View File
@@ -31,13 +31,14 @@ var bench_inputs = []struct {
func TestUnmarshalDatasetCode(t *testing.T) {
for _, tc := range bench_inputs {
buf := fixture(t, tc.name)
t.Run(tc.name, func(t *testing.T) {
buf := fixture(t, tc.name)
var v interface{}
check(t, toml.Unmarshal(buf, &v))
require.NoError(t, toml.Unmarshal(buf, &v))
b, err := json.Marshal(v)
check(t, err)
require.NoError(t, err)
require.Equal(t, len(b), tc.jsonLen)
})
}
@@ -45,14 +46,14 @@ func TestUnmarshalDatasetCode(t *testing.T) {
func BenchmarkUnmarshalDataset(b *testing.B) {
for _, tc := range bench_inputs {
buf := fixture(b, tc.name)
b.Run(tc.name, func(b *testing.B) {
buf := fixture(b, tc.name)
b.SetBytes(int64(len(buf)))
b.ReportAllocs()
b.ResetTimer()
for i := 0; i < b.N; i++ {
var v interface{}
check(b, toml.Unmarshal(buf, &v))
require.NoError(b, toml.Unmarshal(buf, &v))
}
})
}
@@ -60,22 +61,20 @@ func BenchmarkUnmarshalDataset(b *testing.B) {
// fixture returns the uncompressed contents of path.
func fixture(tb testing.TB, path string) []byte {
f, err := os.Open(filepath.Join("testdata", path+".toml.gz"))
check(tb, err)
tb.Helper()
file := path + ".toml.gz"
f, err := os.Open(filepath.Join("testdata", file))
if os.IsNotExist(err) {
tb.Skip("benchmark fixture not found:", file)
}
require.NoError(tb, err)
defer f.Close()
gz, err := gzip.NewReader(f)
check(tb, err)
require.NoError(tb, err)
buf, err := ioutil.ReadAll(gz)
check(tb, err)
require.NoError(tb, err)
return buf
}
func check(tb testing.TB, err error) {
if err != nil {
tb.Helper()
tb.Fatal(err)
}
}
+1 -1
View File
@@ -186,7 +186,7 @@ key3 = 1979-05-27T00:32:00.999999-07:00
key1 = [ 1, 2, 3 ]
key2 = [ "red", "yellow", "green" ]
key3 = [ [ 1, 2 ], [3, 4, 5] ]
#key4 = [ [ 1, 2 ], ["a", "b", "c"] ] # this is ok
key4 = [ [ 1, 2 ], ["a", "b", "c"] ] # this is ok
# Arrays can also be multiline. So in addition to ignoring whitespace, arrays
# also ignore newlines between the brackets. Terminating commas are ok before
+529 -30
View File
@@ -1,6 +1,7 @@
package benchmark_test
import (
"bytes"
"io/ioutil"
"testing"
"time"
@@ -9,17 +10,230 @@ import (
"github.com/stretchr/testify/require"
)
func BenchmarkUnmarshalSimple(b *testing.B) {
func TestUnmarshalSimple(t *testing.T) {
doc := []byte(`A = "hello"`)
d := struct {
A string
}{}
doc := []byte(`A = "hello"`)
for i := 0; i < b.N; i++ {
err := toml.Unmarshal(doc, &d)
err := toml.Unmarshal(doc, &d)
if err != nil {
panic(err)
}
}
func BenchmarkUnmarshal(b *testing.B) {
b.Run("SimpleDocument", func(b *testing.B) {
doc := []byte(`A = "hello"`)
b.Run("struct", func(b *testing.B) {
b.SetBytes(int64(len(doc)))
b.ReportAllocs()
b.ResetTimer()
for i := 0; i < b.N; i++ {
d := struct {
A string
}{}
err := toml.Unmarshal(doc, &d)
if err != nil {
panic(err)
}
}
})
b.Run("map", func(b *testing.B) {
b.SetBytes(int64(len(doc)))
b.ReportAllocs()
b.ResetTimer()
for i := 0; i < b.N; i++ {
d := map[string]interface{}{}
err := toml.Unmarshal(doc, &d)
if err != nil {
panic(err)
}
}
})
})
b.Run("ReferenceFile", func(b *testing.B) {
bytes, err := ioutil.ReadFile("benchmark.toml")
if err != nil {
b.Fatal(err)
}
b.Run("struct", func(b *testing.B) {
b.SetBytes(int64(len(bytes)))
b.ReportAllocs()
b.ResetTimer()
for i := 0; i < b.N; i++ {
d := benchmarkDoc{}
err := toml.Unmarshal(bytes, &d)
if err != nil {
panic(err)
}
}
})
b.Run("map", func(b *testing.B) {
b.SetBytes(int64(len(bytes)))
b.ReportAllocs()
b.ResetTimer()
for i := 0; i < b.N; i++ {
d := map[string]interface{}{}
err := toml.Unmarshal(bytes, &d)
if err != nil {
panic(err)
}
}
})
})
b.Run("HugoFrontMatter", func(b *testing.B) {
b.SetBytes(int64(len(hugoFrontMatterbytes)))
b.ReportAllocs()
b.ResetTimer()
for i := 0; i < b.N; i++ {
d := map[string]interface{}{}
err := toml.Unmarshal(hugoFrontMatterbytes, &d)
if err != nil {
panic(err)
}
}
})
}
func marshal(v interface{}) ([]byte, error) {
var b bytes.Buffer
enc := toml.NewEncoder(&b)
err := enc.Encode(v)
return b.Bytes(), err
}
func BenchmarkMarshal(b *testing.B) {
b.Run("SimpleDocument", func(b *testing.B) {
doc := []byte(`A = "hello"`)
b.Run("struct", func(b *testing.B) {
d := struct {
A string
}{}
err := toml.Unmarshal(doc, &d)
if err != nil {
panic(err)
}
b.ReportAllocs()
b.ResetTimer()
var out []byte
for i := 0; i < b.N; i++ {
out, err = marshal(d)
if err != nil {
panic(err)
}
}
b.SetBytes(int64(len(out)))
})
b.Run("map", func(b *testing.B) {
d := map[string]interface{}{}
err := toml.Unmarshal(doc, &d)
if err != nil {
panic(err)
}
b.ReportAllocs()
b.ResetTimer()
var out []byte
for i := 0; i < b.N; i++ {
out, err = marshal(d)
if err != nil {
panic(err)
}
}
b.SetBytes(int64(len(out)))
})
})
b.Run("ReferenceFile", func(b *testing.B) {
bytes, err := ioutil.ReadFile("benchmark.toml")
if err != nil {
b.Fatal(err)
}
b.Run("struct", func(b *testing.B) {
d := benchmarkDoc{}
err := toml.Unmarshal(bytes, &d)
if err != nil {
panic(err)
}
b.ReportAllocs()
b.ResetTimer()
var out []byte
for i := 0; i < b.N; i++ {
out, err = marshal(d)
if err != nil {
panic(err)
}
}
b.SetBytes(int64(len(out)))
})
b.Run("map", func(b *testing.B) {
d := map[string]interface{}{}
err := toml.Unmarshal(bytes, &d)
if err != nil {
panic(err)
}
b.ReportAllocs()
b.ResetTimer()
var out []byte
for i := 0; i < b.N; i++ {
out, err = marshal(d)
if err != nil {
panic(err)
}
}
b.SetBytes(int64(len(out)))
})
})
b.Run("HugoFrontMatter", func(b *testing.B) {
d := map[string]interface{}{}
err := toml.Unmarshal(hugoFrontMatterbytes, &d)
if err != nil {
panic(err)
}
}
b.ReportAllocs()
b.ResetTimer()
var out []byte
for i := 0; i < b.N; i++ {
out, err = marshal(d)
if err != nil {
panic(err)
}
}
b.SetBytes(int64(len(out)))
})
}
type benchmarkDoc struct {
@@ -35,7 +249,7 @@ type benchmarkDoc struct {
}
Point struct {
X int64
U int64
Y int64
}
}
}
@@ -107,7 +321,7 @@ type benchmarkDoc struct {
Key1 []int64
Key2 []string
Key3 [][]int64
// TODO: Key4 not supported by go-toml's Unmarshal
Key4 []interface{}
Key5 []int64
Key6 []int64
}
@@ -119,36 +333,321 @@ type benchmarkDoc struct {
Fruit []struct {
Name string
Physical struct {
Color string
Shape string
Variety []struct {
Name string
}
Color string
Shape string
}
Variety []struct {
Name string
}
}
}
func BenchmarkReferenceFile(b *testing.B) {
bytes, err := ioutil.ReadFile("benchmark.toml")
if err != nil {
b.Fatal(err)
}
b.SetBytes(int64(len(bytes)))
b.ReportAllocs()
b.ResetTimer()
for i := 0; i < b.N; i++ {
d := benchmarkDoc{}
err := toml.Unmarshal(bytes, &d)
if err != nil {
panic(err)
}
}
}
func TestReferenceFile(t *testing.T) {
func TestUnmarshalReferenceFile(t *testing.T) {
bytes, err := ioutil.ReadFile("benchmark.toml")
require.NoError(t, err)
d := benchmarkDoc{}
err = toml.Unmarshal(bytes, &d)
require.NoError(t, err)
expected := benchmarkDoc{
Table: struct {
Key string
Subtable struct{ Key string }
Inline struct {
Name struct {
First string
Last string
}
Point struct {
X int64
Y int64
}
}
}{
Key: "value",
Subtable: struct{ Key string }{
Key: "another value",
},
// note: x.y.z.w is purposefully missing
Inline: struct {
Name struct {
First string
Last string
}
Point struct {
X int64
Y int64
}
}{
Name: struct {
First string
Last string
}{
First: "Tom",
Last: "Preston-Werner",
},
Point: struct {
X int64
Y int64
}{
X: 1,
Y: 2,
},
},
},
String: struct {
Basic struct{ Basic string }
Multiline struct {
Key1 string
Key2 string
Key3 string
Continued struct {
Key1 string
Key2 string
Key3 string
}
}
Literal struct {
Winpath string
Winpath2 string
Quoted string
Regex string
Multiline struct {
Regex2 string
Lines string
}
}
}{
Basic: struct{ Basic string }{
Basic: "I'm a string. \"You can quote me\". Name\tJos\u00E9\nLocation\tSF.",
},
Multiline: struct {
Key1 string
Key2 string
Key3 string
Continued struct {
Key1 string
Key2 string
Key3 string
}
}{
Key1: "One\nTwo",
Key2: "One\nTwo",
Key3: "One\nTwo",
Continued: struct {
Key1 string
Key2 string
Key3 string
}{
Key1: `The quick brown fox jumps over the lazy dog.`,
Key2: `The quick brown fox jumps over the lazy dog.`,
Key3: `The quick brown fox jumps over the lazy dog.`,
},
},
Literal: struct {
Winpath string
Winpath2 string
Quoted string
Regex string
Multiline struct {
Regex2 string
Lines string
}
}{
Winpath: `C:\Users\nodejs\templates`,
Winpath2: `\\ServerX\admin$\system32\`,
Quoted: `Tom "Dubs" Preston-Werner`,
Regex: `<\i\c*\s*>`,
Multiline: struct {
Regex2 string
Lines string
}{
Regex2: `I [dw]on't need \d{2} apples`,
Lines: `The first newline is
trimmed in raw strings.
All other whitespace
is preserved.
`,
},
},
},
Integer: struct {
Key1 int64
Key2 int64
Key3 int64
Key4 int64
Underscores struct {
Key1 int64
Key2 int64
Key3 int64
}
}{
Key1: 99,
Key2: 42,
Key3: 0,
Key4: -17,
Underscores: struct {
Key1 int64
Key2 int64
Key3 int64
}{
Key1: 1000,
Key2: 5349221,
Key3: 12345,
},
},
Float: struct {
Fractional struct {
Key1 float64
Key2 float64
Key3 float64
}
Exponent struct {
Key1 float64
Key2 float64
Key3 float64
}
Both struct{ Key float64 }
Underscores struct {
Key1 float64
Key2 float64
}
}{
Fractional: struct {
Key1 float64
Key2 float64
Key3 float64
}{
Key1: 1.0,
Key2: 3.1415,
Key3: -0.01,
},
Exponent: struct {
Key1 float64
Key2 float64
Key3 float64
}{
Key1: 5e+22,
Key2: 1e6,
Key3: -2e-2,
},
Both: struct{ Key float64 }{
Key: 6.626e-34,
},
Underscores: struct {
Key1 float64
Key2 float64
}{
Key1: 9224617.445991228313,
Key2: 1e100,
},
},
Boolean: struct {
True bool
False bool
}{
True: true,
False: false,
},
Datetime: struct {
Key1 time.Time
Key2 time.Time
Key3 time.Time
}{
Key1: time.Date(1979, 5, 27, 7, 32, 0, 0, time.UTC),
Key2: time.Date(1979, 5, 27, 0, 32, 0, 0, time.FixedZone("", -7*3600)),
Key3: time.Date(1979, 5, 27, 0, 32, 0, 999999000, time.FixedZone("", -7*3600)),
},
Array: struct {
Key1 []int64
Key2 []string
Key3 [][]int64
Key4 []interface{}
Key5 []int64
Key6 []int64
}{
Key1: []int64{1, 2, 3},
Key2: []string{"red", "yellow", "green"},
Key3: [][]int64{{1, 2}, {3, 4, 5}},
Key4: []interface{}{
[]interface{}{int64(1), int64(2)},
[]interface{}{"a", "b", "c"},
},
Key5: []int64{1, 2, 3},
Key6: []int64{1, 2},
},
Products: []struct {
Name string
Sku int64
Color string
}{
{
Name: "Hammer",
Sku: 738594937,
},
{},
{
Name: "Nail",
Sku: 284758393,
Color: "gray",
},
},
Fruit: []struct {
Name string
Physical struct {
Color string
Shape string
}
Variety []struct{ Name string }
}{
{
Name: "apple",
Physical: struct {
Color string
Shape string
}{
Color: "red",
Shape: "round",
},
Variety: []struct{ Name string }{
{Name: "red delicious"},
{Name: "granny smith"},
},
},
{
Name: "banana",
Variety: []struct{ Name string }{
{Name: "plantain"},
},
},
},
}
require.Equal(t, expected, d)
}
var hugoFrontMatterbytes = []byte(`
categories = ["Development", "VIM"]
date = "2012-04-06"
description = "spf13-vim is a cross platform distribution of vim plugins and resources for Vim."
slug = "spf13-vim-3-0-release-and-new-website"
tags = [".vimrc", "plugins", "spf13-vim", "vim"]
title = "spf13-vim 3.0 release and new website"
include_toc = true
show_comments = false
[[cascade]]
background = "yosemite.jpg"
[cascade._target]
kind = "page"
lang = "en"
path = "/blog/**"
[[cascade]]
background = "goldenbridge.jpg"
[cascade._target]
kind = "section"
`)
+71
View File
@@ -0,0 +1,71 @@
package toml
import (
"bytes"
"testing"
)
var valid10Ascii = []byte("1234567890")
var valid10Utf8 = []byte("日本語a")
var valid1kUtf8 = bytes.Repeat([]byte("0123456789日本語日本語日本語日abcdefghijklmnopqrstuvwx"), 16)
var valid1MUtf8 = bytes.Repeat(valid1kUtf8, 1024)
var valid1kAscii = bytes.Repeat([]byte("012345678998jhjklasDJKLAAdjdfjsdklfjdslkabcdefghijklmnopqrstuvwx"), 16)
var valid1MAscii = bytes.Repeat(valid1kAscii, 1024)
func BenchmarkScanComments(b *testing.B) {
wrap := func(x []byte) []byte {
return []byte("# " + string(x) + "\n")
}
inputs := map[string][]byte{
"10Valid": wrap(valid10Ascii),
"1kValid": wrap(valid1kAscii),
"1MValid": wrap(valid1MAscii),
"10ValidUtf8": wrap(valid10Utf8),
"1kValidUtf8": wrap(valid1kUtf8),
"1MValidUtf8": wrap(valid1MUtf8),
}
for name, input := range inputs {
b.Run(name, func(b *testing.B) {
b.SetBytes(int64(len(input)))
b.ReportAllocs()
b.ResetTimer()
for i := 0; i < b.N; i++ {
scanComment(input)
}
})
}
}
func BenchmarkParseLiteralStringValid(b *testing.B) {
wrap := func(x []byte) []byte {
return []byte("'" + string(x) + "'")
}
inputs := map[string][]byte{
"10Valid": wrap(valid10Ascii),
"1kValid": wrap(valid1kAscii),
"1MValid": wrap(valid1MAscii),
"10ValidUtf8": wrap(valid10Utf8),
"1kValidUtf8": wrap(valid1kUtf8),
"1MValidUtf8": wrap(valid1MUtf8),
}
for name, input := range inputs {
b.Run(name, func(b *testing.B) {
p := parser{}
b.SetBytes(int64(len(input)))
b.ReportAllocs()
b.ResetTimer()
for i := 0; i < b.N; i++ {
_, _, _, err := p.parseLiteralString(input)
if err != nil {
panic(err)
}
}
})
}
}
+112 -3
View File
@@ -39,6 +39,12 @@ benchmark [OPTIONS...] [BRANCH]
-d Compare benchmarks of HEAD with BRANCH using benchstats. In
this form the BRANCH argument is required.
-a Compare benchmarks of HEAD against go-toml v1 and
BurntSushi/toml.
-html When used with -a, emits the output as HTML, ready to be
embedded in the README.
coverage [OPTIONS...] [BRANCH]
Generates code coverage.
@@ -118,6 +124,7 @@ coverage() {
bench() {
branch="${1}"
out="${2}"
replace="${3}"
dir="$(mktemp -d)"
stderr "Executing benchmark for ${branch} at ${dir}"
@@ -129,7 +136,14 @@ bench() {
fi
pushd "$dir"
go test -bench=. -count=10 ./... | tee "${out}"
if [ "${replace}" != "" ]; then
find ./benchmark/ -iname '*.go' -exec sed -i -E "s|github.com/pelletier/go-toml/v2|${replace}|g" {} \;
go get "${replace}"
fi
export GOMAXPROCS=2
nice -n -19 taskset --cpu-list 0,1 go test '-bench=^Benchmark(Un)?[mM]arshal' -count=5 -run=Nothing ./... | tee "${out}"
popd
if [ "${branch}" != "HEAD" ]; then
@@ -137,19 +151,114 @@ bench() {
fi
}
fmktemp() {
if mktemp --version|grep GNU >/dev/null; then
mktemp --suffix=-$1;
else
mktemp -t $1;
fi
}
benchstathtml() {
python3 - $1 <<'EOF'
import sys
lines = []
stop = False
with open(sys.argv[1]) as f:
for line in f.readlines():
line = line.strip()
if line == "":
stop = True
if not stop:
lines.append(line.split(','))
results = []
for line in reversed(lines[1:]):
v2 = float(line[1])
results.append([
line[0].replace("-32", ""),
"%.1fx" % (float(line[3])/v2), # v1
"%.1fx" % (float(line[5])/v2), # bs
])
# move geomean to the end
results.append(results[0])
del results[0]
def printtable(data):
print("""
<table>
<thead>
<tr><th>Benchmark</th><th>go-toml v1</th><th>BurntSushi/toml</th></tr>
</thead>
<tbody>""")
for r in data:
print(" <tr><td>{}</td><td>{}</td><td>{}</td></tr>".format(*r))
print(""" </tbody>
</table>""")
def match(x):
return "ReferenceFile" in x[0] or "HugoFrontMatter" in x[0]
above = [x for x in results if match(x)]
below = [x for x in results if not match(x)]
printtable(above)
print("<details><summary>See more</summary>")
print("""<p>The table above has the results of the most common use-cases. The table below
contains the results of all benchmarks, including unrealistic ones. It is
provided for completeness.</p>""")
printtable(below)
print('<p>This table can be generated with <code>./ci.sh benchmark -a -html</code>.</p>')
print("</details>")
EOF
}
benchmark() {
case "$1" in
-d)
shift
target="${1?Need to provide a target branch argument}"
old=`mktemp`
old=`fmktemp ${target}`
bench "${target}" "${old}"
new=`mktemp`
new=`fmktemp HEAD`
bench HEAD "${new}"
benchstat "${old}" "${new}"
return 0
;;
-a)
shift
v2stats=`fmktemp go-toml-v2`
bench HEAD "${v2stats}" "github.com/pelletier/go-toml/v2"
v1stats=`fmktemp go-toml-v1`
bench HEAD "${v1stats}" "github.com/pelletier/go-toml"
bsstats=`fmktemp bs-toml`
bench HEAD "${bsstats}" "github.com/BurntSushi/toml"
cp "${v2stats}" go-toml-v2.txt
cp "${v1stats}" go-toml-v1.txt
cp "${bsstats}" bs-toml.txt
if [ "$1" = "-html" ]; then
tmpcsv=`fmktemp csv`
benchstat -csv -geomean go-toml-v2.txt go-toml-v1.txt bs-toml.txt > $tmpcsv
benchstathtml $tmpcsv
else
benchstat -geomean go-toml-v2.txt go-toml-v1.txt bs-toml.txt
fi
rm -f go-toml-v2.txt go-toml-v1.txt bs-toml.txt
return $?
esac
bench "${1-HEAD}" `mktemp`
+30
View File
@@ -0,0 +1,30 @@
package main
import (
"flag"
"log"
"os"
"path"
"github.com/pelletier/go-toml/v2/testsuite"
)
func main() {
log.SetFlags(0)
flag.Usage = usage
flag.Parse()
if flag.NArg() != 0 {
flag.Usage()
}
err := testsuite.DecodeStdin()
if err != nil {
log.Fatal(err)
}
}
func usage() {
log.Printf("Usage: %s < toml-file\n", path.Base(os.Args[0]))
flag.PrintDefaults()
os.Exit(1)
}
+223
View File
@@ -0,0 +1,223 @@
// tomltestgen retrieves a given version of the language-agnostic TOML test suite in
// https://github.com/BurntSushi/toml-test and generates go-toml unit tests.
//
// Within the go-toml package, run `go generate`. Otherwise, use:
//
// go run github.com/pelletier/go-toml/cmd/tomltestgen -o toml_testgen_test.go
package main
import (
"archive/zip"
"bytes"
"flag"
"fmt"
"go/format"
"io"
"io/ioutil"
"log"
"net/http"
"os"
"regexp"
"strconv"
"strings"
"text/template"
"time"
)
type invalid struct {
Name string
Input string
}
type valid struct {
Name string
Input string
JsonRef string
}
type testsCollection struct {
Ref string
Timestamp string
Invalid []invalid
Valid []valid
Count int
}
const srcTemplate = "// Generated by tomltestgen for toml-test ref {{.Ref}} on {{.Timestamp}}\n" +
"package toml_test\n" +
" import (\n" +
" \"testing\"\n" +
")\n" +
"{{range .Invalid}}\n" +
"func TestTOMLTest_Invalid_{{.Name}}(t *testing.T) {\n" +
" input := {{.Input|gostr}}\n" +
" testgenInvalid(t, input)\n" +
"}\n" +
"{{end}}\n" +
"\n" +
"{{range .Valid}}\n" +
"func TestTOMLTest_Valid_{{.Name}}(t *testing.T) {\n" +
" input := {{.Input|gostr}}\n" +
" jsonRef := {{.JsonRef|gostr}}\n" +
" testgenValid(t, input, jsonRef)\n" +
"}\n" +
"{{end}}\n"
func downloadTmpFile(url string) string {
log.Println("starting to download file from", url)
resp, err := http.Get(url)
if err != nil {
panic(err)
}
defer resp.Body.Close()
tmpfile, err := ioutil.TempFile("", "toml-test-*.zip")
if err != nil {
panic(err)
}
defer tmpfile.Close()
copiedLen, err := io.Copy(tmpfile, resp.Body)
if err != nil {
panic(err)
}
if resp.ContentLength > 0 && copiedLen != resp.ContentLength {
panic(fmt.Errorf("copied %d bytes, request body had %d", copiedLen, resp.ContentLength))
}
return tmpfile.Name()
}
func kebabToCamel(kebab string) string {
camel := ""
nextUpper := true
for _, c := range kebab {
if nextUpper {
camel += strings.ToUpper(string(c))
nextUpper = false
} else if c == '-' {
nextUpper = true
} else if c == '/' {
nextUpper = true
camel += "_"
} else {
camel += string(c)
}
}
return camel
}
func readFileFromZip(f *zip.File) string {
reader, err := f.Open()
if err != nil {
panic(err)
}
defer reader.Close()
bytes, err := ioutil.ReadAll(reader)
if err != nil {
panic(err)
}
return string(bytes)
}
func templateGoStr(input string) string {
return strconv.Quote(input)
}
var (
ref = flag.String("r", "master", "git reference")
out = flag.String("o", "", "output file")
)
func usage() {
_, _ = fmt.Fprintf(os.Stderr, "usage: tomltestgen [flags]\n")
flag.PrintDefaults()
}
func main() {
flag.Usage = usage
flag.Parse()
url := "https://codeload.github.com/BurntSushi/toml-test/zip/" + *ref
resultFile := downloadTmpFile(url)
defer os.Remove(resultFile)
log.Println("file written to", resultFile)
zipReader, err := zip.OpenReader(resultFile)
if err != nil {
panic(err)
}
defer zipReader.Close()
collection := testsCollection{
Ref: *ref,
Timestamp: time.Now().Format(time.RFC3339),
}
zipFilesMap := map[string]*zip.File{}
for _, f := range zipReader.File {
zipFilesMap[f.Name] = f
}
testFileRegexp := regexp.MustCompile(`([^/]+/tests/(valid|invalid)/(.+))\.(toml)`)
for _, f := range zipReader.File {
groups := testFileRegexp.FindStringSubmatch(f.Name)
if len(groups) > 0 {
name := kebabToCamel(groups[3])
testType := groups[2]
log.Printf("> [%s] %s\n", testType, name)
tomlContent := readFileFromZip(f)
switch testType {
case "invalid":
collection.Invalid = append(collection.Invalid, invalid{
Name: name,
Input: tomlContent,
})
collection.Count++
case "valid":
baseFilePath := groups[1]
jsonFilePath := baseFilePath + ".json"
jsonContent := readFileFromZip(zipFilesMap[jsonFilePath])
collection.Valid = append(collection.Valid, valid{
Name: name,
Input: tomlContent,
JsonRef: jsonContent,
})
collection.Count++
default:
panic(fmt.Sprintf("unknown test type: %s", testType))
}
}
}
log.Printf("Collected %d tests from toml-test\n", collection.Count)
funcMap := template.FuncMap{
"gostr": templateGoStr,
}
t := template.Must(template.New("src").Funcs(funcMap).Parse(srcTemplate))
buf := new(bytes.Buffer)
err = t.Execute(buf, collection)
if err != nil {
panic(err)
}
outputBytes, err := format.Source(buf.Bytes())
if err != nil {
panic(err)
}
if *out == "" {
fmt.Println(string(outputBytes))
return
}
err = os.WriteFile(*out, outputBytes, 0644)
if err != nil {
panic(err)
}
}
+209 -27
View File
@@ -35,26 +35,42 @@ func parseLocalDate(b []byte) (LocalDate, error) {
return date, newDecodeError(b, "dates are expected to have the format YYYY-MM-DD")
}
date.Year = parseDecimalDigits(b[0:4])
var err error
v := parseDecimalDigits(b[5:7])
date.Year, err = parseDecimalDigits(b[0:4])
if err != nil {
return LocalDate{}, err
}
date.Month = time.Month(v)
date.Month, err = parseDecimalDigits(b[5:7])
if err != nil {
return LocalDate{}, err
}
date.Day = parseDecimalDigits(b[8:10])
date.Day, err = parseDecimalDigits(b[8:10])
if err != nil {
return LocalDate{}, err
}
if !isValidDate(date.Year, date.Month, date.Day) {
return LocalDate{}, newDecodeError(b, "impossible date")
}
return date, nil
}
func parseDecimalDigits(b []byte) int {
func parseDecimalDigits(b []byte) (int, error) {
v := 0
for _, c := range b {
for i, c := range b {
if c < '0' || c > '9' {
return 0, newDecodeError(b[i:i+1], "expected digit (0-9)")
}
v *= 10
v += int(c - '0')
}
return v
return v, nil
}
func parseDateTime(b []byte) (time.Time, error) {
@@ -75,7 +91,7 @@ func parseDateTime(b []byte) (time.Time, error) {
panic("date time should have a timezone")
}
if b[0] == 'Z' {
if b[0] == 'Z' || b[0] == 'z' {
b = b[1:]
zone = time.UTC
} else {
@@ -100,13 +116,13 @@ func parseDateTime(b []byte) (time.Time, error) {
}
t := time.Date(
dt.Date.Year,
dt.Date.Month,
dt.Date.Day,
dt.Time.Hour,
dt.Time.Minute,
dt.Time.Second,
dt.Time.Nanosecond,
dt.Year,
time.Month(dt.Month),
dt.Day,
dt.Hour,
dt.Minute,
dt.Second,
dt.Nanosecond,
zone)
return t, nil
@@ -124,10 +140,10 @@ func parseLocalDateTime(b []byte) (LocalDateTime, []byte, error) {
if err != nil {
return dt, nil, err
}
dt.Date = date
dt.LocalDate = date
sep := b[10]
if sep != 'T' && sep != ' ' {
if sep != 'T' && sep != ' ' && sep != 't' {
return dt, nil, newDecodeError(b[10:11], "datetime separator is expected to be T or a space")
}
@@ -135,7 +151,7 @@ func parseLocalDateTime(b []byte) (LocalDateTime, []byte, error) {
if err != nil {
return dt, nil, err
}
dt.Time = t
dt.LocalTime = t
return dt, rest, nil
}
@@ -149,22 +165,45 @@ func parseLocalTime(b []byte) (LocalTime, []byte, error) {
t LocalTime
)
// check if b matches to have expected format HH:MM:SS[.NNNNNN]
const localTimeByteLen = 8
if len(b) < localTimeByteLen {
return t, nil, newDecodeError(b, "times are expected to have the format HH:MM:SS[.NNNNNN]")
}
t.Hour = parseDecimalDigits(b[0:2])
var err error
t.Hour, err = parseDecimalDigits(b[0:2])
if err != nil {
return t, nil, err
}
if t.Hour > 23 {
return t, nil, newDecodeError(b[0:2], "hour cannot be greater 23")
}
if b[2] != ':' {
return t, nil, newDecodeError(b[2:3], "expecting colon between hours and minutes")
}
t.Minute = parseDecimalDigits(b[3:5])
t.Minute, err = parseDecimalDigits(b[3:5])
if err != nil {
return t, nil, err
}
if t.Minute > 59 {
return t, nil, newDecodeError(b[3:5], "minutes cannot be greater 59")
}
if b[5] != ':' {
return t, nil, newDecodeError(b[5:6], "expecting colon between minutes and seconds")
}
t.Second = parseDecimalDigits(b[6:8])
t.Second, err = parseDecimalDigits(b[6:8])
if err != nil {
return t, nil, err
}
if t.Second > 59 {
return t, nil, newDecodeError(b[3:5], "seconds cannot be greater 59")
}
const minLengthWithFrac = 9
if len(b) >= minLengthWithFrac && b[minLengthWithFrac-1] == '.' {
@@ -172,6 +211,14 @@ func parseLocalTime(b []byte) (LocalTime, []byte, error) {
digits := 0
for i, c := range b[minLengthWithFrac:] {
if !isDigit(c) {
if i == 0 {
return t, nil, newDecodeError(b[i:i+1], "need at least one digit after fraction point")
}
break
}
const maxFracPrecision = 9
if i >= maxFracPrecision {
return t, nil, newDecodeError(b[i:i+1], "maximum precision for date time is nanosecond")
@@ -182,7 +229,12 @@ func parseLocalTime(b []byte) (LocalTime, []byte, error) {
digits++
}
if digits == 0 {
return t, nil, newDecodeError(b[minLengthWithFrac-1:minLengthWithFrac], "nanoseconds need at least one digit")
}
t.Nanosecond = frac * nspow[digits]
t.Precision = digits
return t, b[9+digits:], nil
}
@@ -196,7 +248,7 @@ func parseFloat(b []byte) (float64, error) {
return math.NaN(), nil
}
cleaned, err := checkAndRemoveUnderscores(b)
cleaned, err := checkAndRemoveUnderscoresFloats(b)
if err != nil {
return 0, err
}
@@ -209,6 +261,30 @@ func parseFloat(b []byte) (float64, error) {
return 0, newDecodeError(b, "float cannot end with a dot")
}
dotAlreadySeen := false
for i, c := range cleaned {
if c == '.' {
if dotAlreadySeen {
return 0, newDecodeError(b[i:i+1], "float can have at most one decimal point")
}
if !isDigit(cleaned[i-1]) {
return 0, newDecodeError(b[i-1:i+1], "float decimal point must be preceded by a digit")
}
if !isDigit(cleaned[i+1]) {
return 0, newDecodeError(b[i:i+2], "float decimal point must be followed by a digit")
}
dotAlreadySeen = true
}
}
start := 0
if b[0] == '+' || b[0] == '-' {
start = 1
}
if b[start] == '0' && isDigit(b[start+1]) {
return 0, newDecodeError(b, "float integer part cannot have leading zeroes")
}
f, err := strconv.ParseFloat(string(cleaned), 64)
if err != nil {
return 0, newDecodeError(b, "unable to parse float: %w", err)
@@ -218,7 +294,7 @@ func parseFloat(b []byte) (float64, error) {
}
func parseIntHex(b []byte) (int64, error) {
cleaned, err := checkAndRemoveUnderscores(b[2:])
cleaned, err := checkAndRemoveUnderscoresIntegers(b[2:])
if err != nil {
return 0, err
}
@@ -232,7 +308,7 @@ func parseIntHex(b []byte) (int64, error) {
}
func parseIntOct(b []byte) (int64, error) {
cleaned, err := checkAndRemoveUnderscores(b[2:])
cleaned, err := checkAndRemoveUnderscoresIntegers(b[2:])
if err != nil {
return 0, err
}
@@ -246,7 +322,7 @@ func parseIntOct(b []byte) (int64, error) {
}
func parseIntBin(b []byte) (int64, error) {
cleaned, err := checkAndRemoveUnderscores(b[2:])
cleaned, err := checkAndRemoveUnderscoresIntegers(b[2:])
if err != nil {
return 0, err
}
@@ -259,12 +335,26 @@ func parseIntBin(b []byte) (int64, error) {
return i, nil
}
func isSign(b byte) bool {
return b == '+' || b == '-'
}
func parseIntDec(b []byte) (int64, error) {
cleaned, err := checkAndRemoveUnderscores(b)
cleaned, err := checkAndRemoveUnderscoresIntegers(b)
if err != nil {
return 0, err
}
startIdx := 0
if isSign(cleaned[0]) {
startIdx++
}
if len(cleaned) > startIdx+1 && cleaned[startIdx] == '0' {
return 0, newDecodeError(b, "leading zero not allowed on decimal number")
}
i, err := strconv.ParseInt(string(cleaned), 10, 64)
if err != nil {
return 0, newDecodeError(b, "couldn't parse decimal number: %w", err)
@@ -273,7 +363,7 @@ func parseIntDec(b []byte) (int64, error) {
return i, nil
}
func checkAndRemoveUnderscores(b []byte) ([]byte, error) {
func checkAndRemoveUnderscoresIntegers(b []byte) ([]byte, error) {
if b[0] == '_' {
return nil, newDecodeError(b[0:1], "number cannot start with underscore")
}
@@ -312,3 +402,95 @@ func checkAndRemoveUnderscores(b []byte) ([]byte, error) {
return cleaned, nil
}
func checkAndRemoveUnderscoresFloats(b []byte) ([]byte, error) {
if b[0] == '_' {
return nil, newDecodeError(b[0:1], "number cannot start with underscore")
}
if b[len(b)-1] == '_' {
return nil, newDecodeError(b[len(b)-1:], "number cannot end with underscore")
}
// fast path
i := 0
for ; i < len(b); i++ {
if b[i] == '_' {
break
}
}
if i == len(b) {
return b, nil
}
before := false
cleaned := make([]byte, 0, len(b))
for i := 0; i < len(b); i++ {
c := b[i]
switch c {
case '_':
if !before {
return nil, newDecodeError(b[i-1:i+1], "number must have at least one digit between underscores")
}
if i < len(b)-1 && (b[i+1] == 'e' || b[i+1] == 'E') {
return nil, newDecodeError(b[i+1:i+2], "cannot have underscore before exponent")
}
before = false
case 'e', 'E':
if i < len(b)-1 && b[i+1] == '_' {
return nil, newDecodeError(b[i+1:i+2], "cannot have underscore after exponent")
}
cleaned = append(cleaned, c)
case '.':
if i < len(b)-1 && b[i+1] == '_' {
return nil, newDecodeError(b[i+1:i+2], "cannot have underscore after decimal point")
}
if i > 0 && b[i-1] == '_' {
return nil, newDecodeError(b[i-1:i], "cannot have underscore before decimal point")
}
cleaned = append(cleaned, c)
default:
before = true
cleaned = append(cleaned, c)
}
}
return cleaned, nil
}
// isValidDate checks if a provided date is a date that exists.
func isValidDate(year int, month int, day int) bool {
return day <= daysIn(month, year)
}
// daysBefore[m] counts the number of days in a non-leap year
// before month m begins. There is an entry for m=12, counting
// the number of days before January of next year (365).
var daysBefore = [...]int32{
0,
31,
31 + 28,
31 + 28 + 31,
31 + 28 + 31 + 30,
31 + 28 + 31 + 30 + 31,
31 + 28 + 31 + 30 + 31 + 30,
31 + 28 + 31 + 30 + 31 + 30 + 31,
31 + 28 + 31 + 30 + 31 + 30 + 31 + 31,
31 + 28 + 31 + 30 + 31 + 30 + 31 + 31 + 30,
31 + 28 + 31 + 30 + 31 + 30 + 31 + 31 + 30 + 31,
31 + 28 + 31 + 30 + 31 + 30 + 31 + 31 + 30 + 31 + 30,
31 + 28 + 31 + 30 + 31 + 30 + 31 + 31 + 30 + 31 + 30 + 31,
}
func daysIn(m int, year int) int {
if m == 2 && isLeap(year) {
return 29
}
return int(daysBefore[m] - daysBefore[m-1])
}
func isLeap(year int) bool {
return year%4 == 0 && (year%100 != 0 || year%400 == 0)
}
+12 -3
View File
@@ -5,7 +5,7 @@ import (
"strconv"
"strings"
"github.com/pelletier/go-toml/v2/internal/unsafe"
"github.com/pelletier/go-toml/v2/internal/danger"
)
// DecodeError represents an error encountered during the parsing or decoding
@@ -105,7 +105,7 @@ func (e *DecodeError) Key() Key {
// highlight can be freely deallocated.
//nolint:funlen
func wrapDecodeError(document []byte, de *decodeError) *DecodeError {
offset := unsafe.SubsliceOffset(document, de.highlight)
offset := danger.SubsliceOffset(document, de.highlight)
errMessage := de.Error()
errLine, errColumn := positionAtEnd(document[:offset])
@@ -116,6 +116,7 @@ func wrapDecodeError(document []byte, de *decodeError) *DecodeError {
maxLine := errLine + len(after) - 1
lineColumnWidth := len(strconv.Itoa(maxLine))
// Write the lines of context strictly before the error.
for i := len(before) - 1; i > 0; i-- {
line := errLine - i
buf.WriteString(formatLineNumber(line, lineColumnWidth))
@@ -129,6 +130,8 @@ func wrapDecodeError(document []byte, de *decodeError) *DecodeError {
buf.WriteRune('\n')
}
// Write the document line that contains the error.
buf.WriteString(formatLineNumber(errLine, lineColumnWidth))
buf.WriteString("| ")
@@ -143,6 +146,10 @@ func wrapDecodeError(document []byte, de *decodeError) *DecodeError {
}
buf.WriteRune('\n')
// Write the line with the error message itself (so it does not have a line
// number).
buf.WriteString(strings.Repeat(" ", lineColumnWidth))
buf.WriteString("| ")
@@ -157,6 +164,8 @@ func wrapDecodeError(document []byte, de *decodeError) *DecodeError {
buf.WriteString(errMessage)
}
// Write the lines of context strictly after the error.
for i := 1; i < len(after); i++ {
buf.WriteRune('\n')
line := errLine + i
@@ -230,7 +239,7 @@ forward:
rest = rest[o+1:]
o = 0
case o == len(rest)-1 && o > 0:
case o == len(rest)-1:
// add last line only if it's non-empty
afterLines = append(afterLines, rest)
+8 -3
View File
@@ -12,7 +12,6 @@ import (
//nolint:funlen
func TestDecodeError(t *testing.T) {
t.Parallel()
examples := []struct {
desc string
@@ -149,12 +148,19 @@ line 5`,
6|
7| line 4`,
},
{
desc: "handle remainder of the error line when there is only one line",
doc: [3]string{`P=`, `[`, `#`},
msg: "array is incomplete",
expected: `1| P=[#
| ~ array is incomplete`,
},
}
for _, e := range examples {
e := e
t.Run(e.desc, func(t *testing.T) {
t.Parallel()
b := bytes.Buffer{}
b.Write([]byte(e.doc[0]))
start := b.Len()
@@ -182,7 +188,6 @@ line 5`,
}
func TestDecodeError_Accessors(t *testing.T) {
t.Parallel()
e := DecodeError{
message: "foo",
+100
View File
@@ -0,0 +1,100 @@
package toml_test
import (
"testing"
"github.com/pelletier/go-toml/v2"
"github.com/stretchr/testify/require"
)
func TestFastSimple(t *testing.T) {
m := map[string]int64{}
err := toml.Unmarshal([]byte(`a = 42`), &m)
require.NoError(t, err)
require.Equal(t, map[string]int64{"a": 42}, m)
}
func TestFastSimpleString(t *testing.T) {
m := map[string]string{}
err := toml.Unmarshal([]byte(`a = "hello"`), &m)
require.NoError(t, err)
require.Equal(t, map[string]string{"a": "hello"}, m)
}
func TestFastSimpleInterface(t *testing.T) {
m := map[string]interface{}{}
err := toml.Unmarshal([]byte(`
a = "hello"
b = 42`), &m)
require.NoError(t, err)
require.Equal(t, map[string]interface{}{
"a": "hello",
"b": int64(42),
}, m)
}
func TestFastMultipartKeyInterface(t *testing.T) {
m := map[string]interface{}{}
err := toml.Unmarshal([]byte(`
a.interim = "test"
a.b.c = "hello"
b = 42`), &m)
require.NoError(t, err)
require.Equal(t, map[string]interface{}{
"a": map[string]interface{}{
"interim": "test",
"b": map[string]interface{}{
"c": "hello",
},
},
"b": int64(42),
}, m)
}
func TestFastExistingMap(t *testing.T) {
m := map[string]interface{}{
"ints": map[string]int{},
}
err := toml.Unmarshal([]byte(`
ints.one = 1
ints.two = 2
strings.yo = "hello"`), &m)
require.NoError(t, err)
require.Equal(t, map[string]interface{}{
"ints": map[string]interface{}{
"one": int64(1),
"two": int64(2),
},
"strings": map[string]interface{}{
"yo": "hello",
},
}, m)
}
func TestFastArrayTable(t *testing.T) {
b := []byte(`
[root]
[[root.nested]]
name = 'Bob'
[[root.nested]]
name = 'Alice'
`)
m := map[string]interface{}{}
err := toml.Unmarshal(b, &m)
require.NoError(t, err)
require.Equal(t, map[string]interface{}{
"root": map[string]interface{}{
"nested": []interface{}{
map[string]interface{}{
"name": "Bob",
},
map[string]interface{}{
"name": "Alice",
},
},
},
}, m)
}
+3 -2
View File
@@ -1,5 +1,6 @@
module github.com/pelletier/go-toml/v2
go 1.15
go 1.16
require github.com/stretchr/testify v1.7.0
// latest (v1.7.0) doesn't have the fix for time.Time
require github.com/stretchr/testify v1.7.1-0.20210427113832-6241f9ab9942
+2 -2
View File
@@ -3,8 +3,8 @@ github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSs
github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
github.com/stretchr/testify v1.7.0 h1:nwc3DEeHmmLAfoZucVR881uASk0Mfjw8xYJ99tb5CcY=
github.com/stretchr/testify v1.7.0/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
github.com/stretchr/testify v1.7.1-0.20210427113832-6241f9ab9942 h1:t0lM6y/M5IiUZyvbBTcngso8SZEZICH7is9B6g/obVU=
github.com/stretchr/testify v1.7.1-0.20210427113832-6241f9ab9942/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405 h1:yhCVgyC4o1eVCa2tZl7eS0r+SDo693bJlVdllGtEeKM=
gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c h1:dUUwHk2QECo/6vqA44rthZ8ie2QXMNeKRTHCNY2nXvo=
+41 -34
View File
@@ -2,6 +2,9 @@ package ast
import (
"fmt"
"unsafe"
"github.com/pelletier/go-toml/v2/internal/danger"
)
// Iterator starts uninitialized, you need to call Next() first.
@@ -14,7 +17,7 @@ import (
// }
type Iterator struct {
started bool
node Node
node *Node
}
// Next moves the iterator forward and returns true if points to a node, false
@@ -28,8 +31,14 @@ func (c *Iterator) Next() bool {
return c.node.Valid()
}
// IsLast returns true if the current node of the iterator is the last one.
// Subsequent call to Next() will return false.
func (c *Iterator) IsLast() bool {
return c.node.next == 0
}
// Node returns a copy of the node pointed at by the iterator.
func (c *Iterator) Node() Node {
func (c *Iterator) Node() *Node {
return c.node
}
@@ -44,14 +53,13 @@ type Root struct {
func (r *Root) Iterator() Iterator {
it := Iterator{}
if len(r.nodes) > 0 {
it.node = r.nodes[0]
it.node = &r.nodes[0]
}
return it
}
func (r *Root) at(idx int) Node {
// TODO: unsafe to point to the node directly
return r.nodes[idx]
func (r *Root) at(idx Reference) *Node {
return &r.nodes[idx]
}
// Arrays have one child per element in the array.
@@ -63,42 +71,48 @@ func (r *Root) at(idx int) Node {
// children []Node
type Node struct {
Kind Kind
Data []byte // Raw bytes from the input
Raw Range // Raw bytes from the input.
Data []byte // Node value (could be either allocated or referencing the input).
// next idx (in the root array). 0 if last of the collection.
next int
// child idx (in the root array). 0 if no child.
child int
// pointer to the root array
root *Root
// References to other nodes, as offsets in the backing array from this
// node. References can go backward, so those can be negative.
next int // 0 if last element
child int // 0 if no child
}
type Range struct {
Offset uint32
Length uint32
}
// Next returns a copy of the next node, or an invalid Node if there is no
// next node.
func (n Node) Next() Node {
if n.next <= 0 {
return noNode
func (n *Node) Next() *Node {
if n.next == 0 {
return nil
}
return n.root.at(n.next)
ptr := unsafe.Pointer(n)
size := unsafe.Sizeof(Node{})
return (*Node)(danger.Stride(ptr, size, n.next))
}
// Child returns a copy of the first child node of this node. Other children
// can be accessed calling Next on the first child.
// Returns an invalid Node if there is none.
func (n Node) Child() Node {
if n.child <= 0 {
return noNode
func (n *Node) Child() *Node {
if n.child == 0 {
return nil
}
return n.root.at(n.child)
ptr := unsafe.Pointer(n)
size := unsafe.Sizeof(Node{})
return (*Node)(danger.Stride(ptr, size, n.child))
}
// Valid returns true if the node's kind is set (not to Invalid).
func (n Node) Valid() bool {
return n.Kind != Invalid
func (n *Node) Valid() bool {
return n != nil
}
var noNode = Node{}
// Key returns the child nodes making the Key on a supported node. Panics
// otherwise.
// They are guaranteed to be all be of the Kind Key. A simple key would return
@@ -121,18 +135,11 @@ func (n *Node) Key() Iterator {
// Value returns a pointer to the value node of a KeyValue.
// Guaranteed to be non-nil.
// Panics if not called on a KeyValue node, or if the Children are malformed.
func (n Node) Value() Node {
assertKind(KeyValue, n)
func (n *Node) Value() *Node {
return n.Child()
}
// Children returns an iterator over a node's children.
func (n Node) Children() Iterator {
func (n *Node) Children() Iterator {
return Iterator{node: n.Child()}
}
func assertKind(k Kind, n Node) {
if n.Kind != k {
panic(fmt.Errorf("method was expecting a %s, not a %s", k, n.Kind))
}
}
+11 -20
View File
@@ -1,12 +1,11 @@
package ast
type Reference struct {
idx int
set bool
}
type Reference int
const InvalidReference Reference = -1
func (r Reference) Valid() bool {
return r.set
return r != InvalidReference
}
type Builder struct {
@@ -18,8 +17,8 @@ func (b *Builder) Tree() *Root {
return &b.tree
}
func (b *Builder) NodeAt(ref Reference) Node {
return b.tree.at(ref.idx)
func (b *Builder) NodeAt(ref Reference) *Node {
return b.tree.at(ref)
}
func (b *Builder) Reset() {
@@ -28,33 +27,25 @@ func (b *Builder) Reset() {
}
func (b *Builder) Push(n Node) Reference {
n.root = &b.tree
b.lastIdx = len(b.tree.nodes)
b.tree.nodes = append(b.tree.nodes, n)
return Reference{
idx: b.lastIdx,
set: true,
}
return Reference(b.lastIdx)
}
func (b *Builder) PushAndChain(n Node) Reference {
n.root = &b.tree
newIdx := len(b.tree.nodes)
b.tree.nodes = append(b.tree.nodes, n)
if b.lastIdx >= 0 {
b.tree.nodes[b.lastIdx].next = newIdx
b.tree.nodes[b.lastIdx].next = newIdx - b.lastIdx
}
b.lastIdx = newIdx
return Reference{
idx: b.lastIdx,
set: true,
}
return Reference(b.lastIdx)
}
func (b *Builder) AttachChild(parent Reference, child Reference) {
b.tree.nodes[parent.idx].child = child.idx
b.tree.nodes[parent].child = int(child) - int(parent)
}
func (b *Builder) Chain(from Reference, to Reference) {
b.tree.nodes[from.idx].next = to.idx
b.tree.nodes[from].next = int(to) - int(from)
}
+3 -3
View File
@@ -25,9 +25,9 @@ const (
Float
Integer
LocalDate
LocalTime
LocalDateTime
DateTime
Time
)
func (k Kind) String() string {
@@ -58,12 +58,12 @@ func (k Kind) String() string {
return "Integer"
case LocalDate:
return "LocalDate"
case LocalTime:
return "LocalTime"
case LocalDateTime:
return "LocalDateTime"
case DateTime:
return "DateTime"
case Time:
return "Time"
}
panic(fmt.Errorf("Kind.String() not implemented for '%d'", k))
}
@@ -1,4 +1,4 @@
package unsafe
package danger
import (
"fmt"
@@ -57,3 +57,20 @@ func BytesRange(start []byte, end []byte) []byte {
return start[:l]
}
func Stride(ptr unsafe.Pointer, size uintptr, offset int) unsafe.Pointer {
// TODO: replace with unsafe.Add when Go 1.17 is released
// https://github.com/golang/go/issues/40481
return unsafe.Pointer(uintptr(ptr) + uintptr(int(size)*offset))
}
type Slice struct {
Data unsafe.Pointer
Len int
Cap int
}
type iface struct {
typ unsafe.Pointer
ptr unsafe.Pointer
}
@@ -1,15 +1,16 @@
package unsafe_test
package danger_test
import (
"testing"
"unsafe"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
"github.com/pelletier/go-toml/v2/internal/unsafe"
"github.com/pelletier/go-toml/v2/internal/danger"
)
func TestUnsafeSubsliceOffsetValid(t *testing.T) {
func TestSubsliceOffsetValid(t *testing.T) {
examples := []struct {
desc string
test func() ([]byte, []byte)
@@ -28,13 +29,13 @@ func TestUnsafeSubsliceOffsetValid(t *testing.T) {
for _, e := range examples {
t.Run(e.desc, func(t *testing.T) {
d, s := e.test()
offset := unsafe.SubsliceOffset(d, s)
offset := danger.SubsliceOffset(d, s)
assert.Equal(t, e.offset, offset)
})
}
}
func TestUnsafeSubsliceOffsetInvalid(t *testing.T) {
func TestSubsliceOffsetInvalid(t *testing.T) {
examples := []struct {
desc string
test func() ([]byte, []byte)
@@ -72,13 +73,22 @@ func TestUnsafeSubsliceOffsetInvalid(t *testing.T) {
t.Run(e.desc, func(t *testing.T) {
d, s := e.test()
require.Panics(t, func() {
unsafe.SubsliceOffset(d, s)
danger.SubsliceOffset(d, s)
})
})
}
}
func TestUnsafeBytesRange(t *testing.T) {
func TestStride(t *testing.T) {
a := []byte{1, 2, 3, 4}
x := &a[1]
n := (*byte)(danger.Stride(unsafe.Pointer(x), unsafe.Sizeof(byte(0)), 1))
require.Equal(t, &a[2], n)
n = (*byte)(danger.Stride(unsafe.Pointer(x), unsafe.Sizeof(byte(0)), -1))
require.Equal(t, &a[0], n)
}
func TestBytesRange(t *testing.T) {
type fn = func() ([]byte, []byte)
examples := []struct {
desc string
@@ -157,10 +167,10 @@ func TestUnsafeBytesRange(t *testing.T) {
start, end := e.test()
if e.expected == nil {
require.Panics(t, func() {
unsafe.BytesRange(start, end)
danger.BytesRange(start, end)
})
} else {
res := unsafe.BytesRange(start, end)
res := danger.BytesRange(start, end)
require.Equal(t, e.expected, res)
}
})
+20
View File
@@ -0,0 +1,20 @@
//go:build go1.18
// +build go1.18
package danger
import (
"reflect"
"unsafe"
)
func ExtendSlice(t reflect.Type, s *Slice, n int) Slice {
arrayType := reflect.ArrayOf(n, t.Elem())
arrayData := reflect.New(arrayType)
reflect.Copy(arrayData.Elem(), reflect.NewAt(t, unsafe.Pointer(s)).Elem())
return Slice{
Data: unsafe.Pointer(arrayData.Pointer()),
Len: s.Len,
Cap: n,
}
}
+30
View File
@@ -0,0 +1,30 @@
//go:build !go1.18
// +build !go1.18
package danger
import (
"reflect"
"unsafe"
)
//go:linkname unsafe_NewArray reflect.unsafe_NewArray
func unsafe_NewArray(rtype unsafe.Pointer, length int) unsafe.Pointer
//go:linkname typedslicecopy reflect.typedslicecopy
//go:noescape
func typedslicecopy(elemType unsafe.Pointer, dst, src Slice) int
func ExtendSlice(t reflect.Type, s *Slice, n int) Slice {
elemTypeRef := t.Elem()
elemTypePtr := ((*iface)(unsafe.Pointer(&elemTypeRef))).ptr
d := Slice{
Data: unsafe_NewArray(elemTypePtr, n),
Len: s.Len,
Cap: n,
}
typedslicecopy(elemTypePtr, d, *s)
return d
}
+23
View File
@@ -0,0 +1,23 @@
package danger
import (
"reflect"
"unsafe"
)
// typeID is used as key in encoder and decoder caches to enable using
// the optimize runtime.mapaccess2_fast64 function instead of the more
// expensive lookup if we were to use reflect.Type as map key.
//
// typeID holds the pointer to the reflect.Type value, which is unique
// in the program.
//
// https://github.com/segmentio/encoding/blob/master/json/codec.go#L59-L61
type TypeID unsafe.Pointer
func MakeTypeID(t reflect.Type) TypeID {
// reflect.Type has the fields:
// typ unsafe.Pointer
// ptr unsafe.Pointer
return TypeID((*[2]unsafe.Pointer)(unsafe.Pointer(&t))[1])
}
@@ -223,11 +223,13 @@ type testSubDoc struct {
unexported int `toml:"shouldntBeHere"`
}
var biteMe = "Bite me"
var float1 float32 = 12.3
var float2 float32 = 45.6
var float3 float32 = 78.9
var subdoc = testSubDoc{"Second", 0}
var (
biteMe = "Bite me"
float1 float32 = 12.3
float2 float32 = 45.6
float3 float32 = 78.9
subdoc = testSubDoc{"Second", 0}
)
var docData = testDoc{
Title: "TOML Marshal Testing",
@@ -382,7 +384,7 @@ var intErrTomls = []string{
}
func TestErrUnmarshal(t *testing.T) {
var errTomls = []string{
errTomls := []string{
"bool = truly\ndate = 1979-05-27T07:32:00Z\nfloat = 123.4\nint = 5000\nstring = \"Bite me\"",
"bool = true\ndate = 1979-05-27T07:3200Z\nfloat = 123.4\nint = 5000\nstring = \"Bite me\"",
"bool = true\ndate = 1979-05-27T07:32:00Z\nfloat = 123a4\nint = 5000\nstring = \"Bite me\"",
@@ -468,7 +470,7 @@ func TestEmptyUnmarshalOmit(t *testing.T) {
Map map[string]string `toml:"map,omitempty"`
}
var emptyTestData2 = emptyMarshalTestStruct2{
emptyTestData2 := emptyMarshalTestStruct2{
Title: "Placeholder",
Bool: false,
Int: 0,
@@ -496,21 +498,23 @@ type pointerMarshalTestStruct struct {
DblPtr *[]*[]*string
}
var pointerStr = "Hello"
var pointerList = []string{"Hello back"}
var pointerListPtr = []*string{&pointerStr}
var pointerMap = map[string]string{"response": "Goodbye"}
var pointerMapPtr = map[string]*string{"alternate": &pointerStr}
var pointerTestData = pointerMarshalTestStruct{
Str: &pointerStr,
List: &pointerList,
ListPtr: &pointerListPtr,
Map: &pointerMap,
MapPtr: &pointerMapPtr,
EmptyStr: nil,
EmptyList: nil,
EmptyMap: nil,
}
var (
pointerStr = "Hello"
pointerList = []string{"Hello back"}
pointerListPtr = []*string{&pointerStr}
pointerMap = map[string]string{"response": "Goodbye"}
pointerMapPtr = map[string]*string{"alternate": &pointerStr}
pointerTestData = pointerMarshalTestStruct{
Str: &pointerStr,
List: &pointerList,
ListPtr: &pointerListPtr,
Map: &pointerMap,
MapPtr: &pointerMapPtr,
EmptyStr: nil,
EmptyList: nil,
EmptyMap: nil,
}
)
var pointerTestToml = []byte(`List = ["Hello back"]
ListPtr = ["Hello"]
@@ -538,15 +542,17 @@ func TestUnmarshalTypeMismatch(t *testing.T) {
type nestedMarshalTestStruct struct {
String [][]string
//Struct [][]basicMarshalTestSubStruct
// Struct [][]basicMarshalTestSubStruct
StringPtr *[]*[]*string
// StructPtr *[]*[]*basicMarshalTestSubStruct
}
var str1 = "Three"
var str2 = "Four"
var strPtr = []*string{&str1, &str2}
var strPtr2 = []*[]*string{&strPtr}
var (
str1 = "Three"
str2 = "Four"
strPtr = []*string{&str1, &str2}
strPtr2 = []*[]*string{&strPtr}
)
var nestedTestData = nestedMarshalTestStruct{
String: [][]string{{"Five", "Six"}, {"One", "Two"}},
@@ -597,6 +603,7 @@ var nestedCustomMarshalerData = customMarshalerParent{
var nestedCustomMarshalerToml = []byte(`friends = ["Sally Fields"]
me = "Maiku Suteda"
`)
var nestedCustomMarshalerTomlForUnmarshal = []byte(`[friends]
FirstName = "Sally"
LastName = "Fields"`)
@@ -613,11 +620,11 @@ func (x *IntOrString) MarshalTOML() ([]byte, error) {
}
func TestUnmarshalTextMarshaler(t *testing.T) {
var nested = struct {
nested := struct {
Friends textMarshaler `toml:"friends"`
}{}
var expected = struct {
expected := struct {
Friends textMarshaler `toml:"friends"`
}{
Friends: textMarshaler{FirstName: "Sally", LastName: "Fields"},
@@ -1360,7 +1367,6 @@ func TestUnmarshalPreservesUnexportedFields(t *testing.T) {
t.Run("unexported field should not be set from toml", func(t *testing.T) {
var actual unexportedFieldPreservationTest
err := toml.Unmarshal([]byte(doc), &actual)
if err != nil {
t.Fatal("did not expect an error")
}
@@ -1394,7 +1400,6 @@ func TestUnmarshalPreservesUnexportedFields(t *testing.T) {
Nested3: &unexportedFieldPreservationTestNested{"baz", "bax"},
}
err := toml.Unmarshal([]byte(doc), &actual)
if err != nil {
t.Fatal("did not expect an error")
}
@@ -1431,7 +1436,6 @@ func TestUnmarshalLocalDate(t *testing.T) {
var obj dateStruct
err := toml.Unmarshal([]byte(doc), &obj)
if err != nil {
t.Fatal(err)
}
@@ -1457,7 +1461,6 @@ func TestUnmarshalLocalDate(t *testing.T) {
var obj dateStruct
err := toml.Unmarshal([]byte(doc), &obj)
if err != nil {
t.Fatal(err)
}
@@ -1484,32 +1487,34 @@ func TestUnmarshalLocalDateTime(t *testing.T) {
name: "normal",
in: "1979-05-27T07:32:00",
out: toml.LocalDateTime{
Date: toml.LocalDate{
LocalDate: toml.LocalDate{
Year: 1979,
Month: 5,
Day: 27,
},
Time: toml.LocalTime{
LocalTime: toml.LocalTime{
Hour: 7,
Minute: 32,
Second: 0,
Nanosecond: 0,
},
}},
},
},
{
name: "with nanoseconds",
in: "1979-05-27T00:32:00.999999",
out: toml.LocalDateTime{
Date: toml.LocalDate{
LocalDate: toml.LocalDate{
Year: 1979,
Month: 5,
Day: 27,
},
Time: toml.LocalTime{
LocalTime: toml.LocalTime{
Hour: 0,
Minute: 32,
Second: 0,
Nanosecond: 999999000,
Precision: 6,
},
},
},
@@ -1526,7 +1531,6 @@ func TestUnmarshalLocalDateTime(t *testing.T) {
var obj dateStruct
err := toml.Unmarshal([]byte(doc), &obj)
if err != nil {
t.Fatal(err)
}
@@ -1544,31 +1548,30 @@ func TestUnmarshalLocalDateTime(t *testing.T) {
var obj dateStruct
err := toml.Unmarshal([]byte(doc), &obj)
if err != nil {
t.Fatal(err)
}
if obj.Date.Year() != example.out.Date.Year {
t.Errorf("expected year %d, got %d", example.out.Date.Year, obj.Date.Year())
if obj.Date.Year() != example.out.Year {
t.Errorf("expected year %d, got %d", example.out.Year, obj.Date.Year())
}
if obj.Date.Month() != example.out.Date.Month {
t.Errorf("expected month %d, got %d", example.out.Date.Month, obj.Date.Month())
if obj.Date.Month() != time.Month(example.out.Month) {
t.Errorf("expected month %d, got %d", example.out.Month, obj.Date.Month())
}
if obj.Date.Day() != example.out.Date.Day {
t.Errorf("expected day %d, got %d", example.out.Date.Day, obj.Date.Day())
if obj.Date.Day() != example.out.Day {
t.Errorf("expected day %d, got %d", example.out.Day, obj.Date.Day())
}
if obj.Date.Hour() != example.out.Time.Hour {
t.Errorf("expected hour %d, got %d", example.out.Time.Hour, obj.Date.Hour())
if obj.Date.Hour() != example.out.Hour {
t.Errorf("expected hour %d, got %d", example.out.Hour, obj.Date.Hour())
}
if obj.Date.Minute() != example.out.Time.Minute {
t.Errorf("expected minute %d, got %d", example.out.Time.Minute, obj.Date.Minute())
if obj.Date.Minute() != example.out.Minute {
t.Errorf("expected minute %d, got %d", example.out.Minute, obj.Date.Minute())
}
if obj.Date.Second() != example.out.Time.Second {
t.Errorf("expected second %d, got %d", example.out.Time.Second, obj.Date.Second())
if obj.Date.Second() != example.out.Second {
t.Errorf("expected second %d, got %d", example.out.Second, obj.Date.Second())
}
if obj.Date.Nanosecond() != example.out.Time.Nanosecond {
t.Errorf("expected nanoseconds %d, got %d", example.out.Time.Nanosecond, obj.Date.Nanosecond())
if obj.Date.Nanosecond() != example.out.Nanosecond {
t.Errorf("expected nanoseconds %d, got %d", example.out.Nanosecond, obj.Date.Nanosecond())
}
})
}
@@ -1598,6 +1601,7 @@ func TestUnmarshalLocalTime(t *testing.T) {
Minute: 32,
Second: 0,
Nanosecond: 999999000,
Precision: 6,
},
},
}
@@ -1613,7 +1617,6 @@ func TestUnmarshalLocalTime(t *testing.T) {
var obj dateStruct
err := toml.Unmarshal([]byte(doc), &obj)
if err != nil {
t.Fatal(err)
}
@@ -2283,8 +2286,7 @@ func (d *durationString) UnmarshalTOML(v interface{}) error {
return nil
}
type config437Error struct {
}
type config437Error struct{}
func (e *config437Error) UnmarshalTOML(v interface{}) error {
return errors.New("expected")
+4 -4
View File
@@ -11,19 +11,19 @@ type KeyTracker struct {
}
// UpdateTable sets the state of the tracker with the AST table node.
func (t *KeyTracker) UpdateTable(node ast.Node) {
func (t *KeyTracker) UpdateTable(node *ast.Node) {
t.reset()
t.Push(node)
}
// UpdateArrayTable sets the state of the tracker with the AST array table node.
func (t *KeyTracker) UpdateArrayTable(node ast.Node) {
func (t *KeyTracker) UpdateArrayTable(node *ast.Node) {
t.reset()
t.Push(node)
}
// Push the given key on the stack.
func (t *KeyTracker) Push(node ast.Node) {
func (t *KeyTracker) Push(node *ast.Node) {
it := node.Key()
for it.Next() {
t.k = append(t.k, string(it.Node().Data))
@@ -31,7 +31,7 @@ func (t *KeyTracker) Push(node ast.Node) {
}
// Pop key from stack.
func (t *KeyTracker) Pop(node ast.Node) {
func (t *KeyTracker) Pop(node *ast.Node) {
it := node.Key()
for it.Next() {
t.k = t.k[:len(t.k)-1]
+201 -97
View File
@@ -1,6 +1,7 @@
package tracker
import (
"bytes"
"fmt"
"github.com/pelletier/go-toml/v2/internal/ast"
@@ -29,67 +30,93 @@ func (k keyKind) String() string {
panic("missing keyKind string mapping")
}
// SeenTracker tracks which keys have been seen with which TOML type to flag duplicates
// and mismatches according to the spec.
// SeenTracker tracks which keys have been seen with which TOML type to flag
// duplicates and mismatches according to the spec.
//
// Each node in the visited tree is represented by an entry. Each entry has an
// identifier, which is provided by a counter. Entries are stored in the array
// entries. As new nodes are discovered (referenced for the first time in the
// TOML document), entries are created and appended to the array. An entry
// points to its parent using its id.
//
// To find whether a given key (sequence of []byte) has already been visited,
// the entries are linearly searched, looking for one with the right name and
// parent id.
//
// Given that all keys appear in the document after their parent, it is
// guaranteed that all descendants of a node are stored after the node, this
// speeds up the search process.
//
// When encountering [[array tables]], the descendants of that node are removed
// to allow that branch of the tree to be "rediscovered". To maintain the
// invariant above, the deletion process needs to keep the order of entries.
// This results in more copies in that case.
type SeenTracker struct {
root *info
current *info
entries []entry
currentIdx int
nextID int
}
type info struct {
parent *info
type entry struct {
id int
parent int
name []byte
kind keyKind
children map[string]*info
explicit bool
}
func (i *info) Clear() {
i.children = nil
// Remove all descendants of node at position idx.
func (s *SeenTracker) clear(idx int) {
p := s.entries[idx].id
rest := clear(p, s.entries[idx+1:])
s.entries = s.entries[:idx+1+len(rest)]
}
func (i *info) Has(k string) (*info, bool) {
c, ok := i.children[k]
return c, ok
}
func (i *info) SetKind(kind keyKind) {
i.kind = kind
}
func (i *info) CreateTable(k string, explicit bool) *info {
return i.createChild(k, tableKind, explicit)
}
func (i *info) CreateArrayTable(k string, explicit bool) *info {
return i.createChild(k, arrayTableKind, explicit)
}
func (i *info) createChild(k string, kind keyKind, explicit bool) *info {
if i.children == nil {
i.children = make(map[string]*info, 1)
func clear(parentID int, entries []entry) []entry {
for i := 0; i < len(entries); {
if entries[i].parent == parentID {
id := entries[i].id
copy(entries[i:], entries[i+1:])
entries = entries[:len(entries)-1]
rest := clear(id, entries[i:])
entries = entries[:i+len(rest)]
} else {
i++
}
}
return entries
}
x := &info{
parent: i,
func (s *SeenTracker) create(parentIdx int, name []byte, kind keyKind, explicit bool) int {
parentID := s.id(parentIdx)
idx := len(s.entries)
s.entries = append(s.entries, entry{
id: s.nextID,
parent: parentID,
name: name,
kind: kind,
explicit: explicit,
}
i.children[k] = x
return x
})
s.nextID++
return idx
}
// CheckExpression takes a top-level node and checks that it does not contain keys
// that have been seen in previous calls, and validates that types are consistent.
func (s *SeenTracker) CheckExpression(node ast.Node) error {
if s.root == nil {
s.root = &info{
kind: tableKind,
}
s.current = s.root
// CheckExpression takes a top-level node and checks that it does not contain
// keys that have been seen in previous calls, and validates that types are
// consistent.
func (s *SeenTracker) CheckExpression(node *ast.Node) error {
if s.entries == nil {
// Skip ID = 0 to remove the confusion between nodes whose
// parent has id 0 and root nodes (parent id is 0 because it's
// the zero value).
s.nextID = 1
// Start unscoped, so idx is negative.
s.currentIdx = -1
}
switch node.Kind {
case ast.KeyValue:
return s.checkKeyValue(s.current, node)
return s.checkKeyValue(s.currentIdx, node)
case ast.Table:
return s.checkTable(node)
case ast.ArrayTable:
@@ -97,104 +124,181 @@ func (s *SeenTracker) CheckExpression(node ast.Node) error {
default:
panic(fmt.Errorf("this should not be a top level node type: %s", node.Kind))
}
}
func (s *SeenTracker) checkTable(node ast.Node) error {
s.current = s.root
func (s *SeenTracker) checkTable(node *ast.Node) error {
it := node.Key()
// handle the first parts of the key, excluding the last one
parentIdx := -1
// This code is duplicated in checkArrayTable. This is because factoring
// it in a function requires to copy the iterator, or allocate it to the
// heap, which is not cheap.
for it.Next() {
if !it.Node().Next().Valid() {
if it.IsLast() {
break
}
k := string(it.Node().Data)
child, found := s.current.Has(k)
if !found {
child = s.current.CreateTable(k, false)
k := it.Node().Data
idx := s.find(parentIdx, k)
if idx < 0 {
idx = s.create(parentIdx, k, tableKind, false)
}
s.current = child
parentIdx = idx
}
// handle the last part of the key
k := string(it.Node().Data)
k := it.Node().Data
idx := s.find(parentIdx, k)
i, found := s.current.Has(k)
if found {
if i.kind != tableKind {
return fmt.Errorf("toml: key %s should be a table, not a %s", k, i.kind)
if idx >= 0 {
kind := s.entries[idx].kind
if kind != tableKind {
return fmt.Errorf("toml: key %s should be a table, not a %s", string(k), kind)
}
if i.explicit {
return fmt.Errorf("toml: table %s already exists", k)
if s.entries[idx].explicit {
return fmt.Errorf("toml: table %s already exists", string(k))
}
i.explicit = true
s.current = i
s.entries[idx].explicit = true
} else {
s.current = s.current.CreateTable(k, true)
idx = s.create(parentIdx, k, tableKind, true)
}
s.currentIdx = idx
return nil
}
func (s *SeenTracker) checkArrayTable(node ast.Node) error {
s.current = s.root
func (s *SeenTracker) checkArrayTable(node *ast.Node) error {
it := node.Key()
// handle the first parts of the key, excluding the last one
parentIdx := -1
for it.Next() {
if !it.Node().Next().Valid() {
if it.IsLast() {
break
}
k := string(it.Node().Data)
child, found := s.current.Has(k)
if !found {
child = s.current.CreateTable(k, false)
k := it.Node().Data
idx := s.find(parentIdx, k)
if idx < 0 {
idx = s.create(parentIdx, k, tableKind, false)
}
s.current = child
parentIdx = idx
}
// handle the last part of the key
k := string(it.Node().Data)
k := it.Node().Data
idx := s.find(parentIdx, k)
info, found := s.current.Has(k)
if found {
if info.kind != arrayTableKind {
return fmt.Errorf("toml: key %s already exists as a %s, but should be an array table", info.kind, k)
if idx >= 0 {
kind := s.entries[idx].kind
if kind != arrayTableKind {
return fmt.Errorf("toml: key %s already exists as a %s, but should be an array table", kind, string(k))
}
info.Clear()
s.clear(idx)
} else {
info = s.current.CreateArrayTable(k, true)
idx = s.create(parentIdx, k, arrayTableKind, true)
}
s.current = info
s.currentIdx = idx
return nil
}
func (s *SeenTracker) checkKeyValue(context *info, node ast.Node) error {
func (s *SeenTracker) checkKeyValue(parentIdx int, node *ast.Node) error {
it := node.Key()
// handle the first parts of the key, excluding the last one
for it.Next() {
k := string(it.Node().Data)
child, found := context.Has(k)
if found {
if child.kind != tableKind {
return fmt.Errorf("toml: expected %s to be a table, not a %s", k, child.kind)
}
k := it.Node().Data
idx := s.find(parentIdx, k)
if idx < 0 {
idx = s.create(parentIdx, k, tableKind, false)
} else {
child = context.CreateTable(k, false)
entry := s.entries[idx]
if it.IsLast() {
return fmt.Errorf("toml: key %s is already defined", string(k))
} else if entry.kind != tableKind {
return fmt.Errorf("toml: expected %s to be a table, not a %s", string(k), entry.kind)
} else if entry.explicit {
return fmt.Errorf("toml: cannot redefine table %s that has already been explicitly defined", string(k))
}
}
context = child
parentIdx = idx
}
if node.Value().Kind == ast.InlineTable {
context.SetKind(tableKind)
} else {
context.SetKind(valueKind)
s.entries[parentIdx].kind = valueKind
value := node.Value()
switch value.Kind {
case ast.InlineTable:
return s.checkInlineTable(parentIdx, value)
case ast.Array:
return s.checkArray(parentIdx, value)
}
return nil
}
func (s *SeenTracker) checkArray(parentIdx int, node *ast.Node) error {
set := false
it := node.Children()
for it.Next() {
if set {
s.clear(parentIdx)
}
n := it.Node()
switch n.Kind {
case ast.InlineTable:
err := s.checkInlineTable(parentIdx, n)
if err != nil {
return err
}
set = true
case ast.Array:
err := s.checkArray(parentIdx, n)
if err != nil {
return err
}
set = true
}
}
return nil
}
func (s *SeenTracker) checkInlineTable(parentIdx int, node *ast.Node) error {
it := node.Children()
for it.Next() {
n := it.Node()
err := s.checkKeyValue(parentIdx, n)
if err != nil {
return err
}
}
return nil
}
func (s *SeenTracker) id(idx int) int {
if idx >= 0 {
return s.entries[idx].id
}
return 0
}
func (s *SeenTracker) find(parentIdx int, k []byte) int {
parentID := s.id(parentIdx)
for i := parentIdx + 1; i < len(s.entries); i++ {
if s.entries[i].parent == parentID && bytes.Equal(s.entries[i].name, k) {
return i
}
}
return -1
}
+79 -259
View File
@@ -1,300 +1,120 @@
// Implementation of TOML's local date/time.
// Copied over from https://github.com/googleapis/google-cloud-go/blob/master/civil/civil.go
// to avoid pulling all the Google dependencies.
//
// Copyright 2016 Google LLC
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
// Package civil implements types for civil time, a time-zone-independent
// representation of time that follows the rules of the proleptic
// Gregorian calendar with exactly 24-hour days, 60-minute hours, and 60-second
// minutes.
//
// Because they lack location information, these types do not represent unique
// moments or intervals of time. Use time.Time for that purpose.
package toml
import (
"fmt"
"strings"
"time"
)
// A LocalDate represents a date (year, month, day).
//
// This type does not include location information, and therefore does not
// describe a unique 24-hour timespan.
// LocalDate represents a calendar day in no specific timezone.
type LocalDate struct {
Year int // Year (e.g., 2014).
Month time.Month // Month of the year (January = 1, ...).
Day int // Day of the month, starting at 1.
Year int
Month int
Day int
}
// LocalDateOf returns the LocalDate in which a time occurs in that time's location.
func LocalDateOf(t time.Time) LocalDate {
var d LocalDate
d.Year, d.Month, d.Day = t.Date()
return d
// AsTime converts d into a specific time instance at midnight in zone.
func (d LocalDate) AsTime(zone *time.Location) time.Time {
return time.Date(d.Year, time.Month(d.Month), d.Day, 0, 0, 0, 0, zone)
}
// ParseLocalDate parses a string in RFC3339 full-date format and returns the date value it represents.
func ParseLocalDate(s string) (LocalDate, error) {
t, err := time.Parse("2006-01-02", s)
if err != nil {
return LocalDate{}, err
}
return LocalDateOf(t), nil
}
// String returns the date in RFC3339 full-date format.
// String returns RFC 3339 representation of d.
func (d LocalDate) String() string {
return fmt.Sprintf("%04d-%02d-%02d", d.Year, d.Month, d.Day)
}
// IsValid reports whether the date is valid.
func (d LocalDate) IsValid() bool {
return LocalDateOf(d.In(time.UTC)) == d
}
// In returns the time corresponding to time 00:00:00 of the date in the location.
//
// In is always consistent with time.LocalDate, even when time.LocalDate returns a time
// on a different day. For example, if loc is America/Indiana/Vincennes, then both
// time.LocalDate(1955, time.May, 1, 0, 0, 0, 0, loc)
// and
// civil.LocalDate{Year: 1955, Month: time.May, Day: 1}.In(loc)
// return 23:00:00 on April 30, 1955.
//
// In panics if loc is nil.
func (d LocalDate) In(loc *time.Location) time.Time {
return time.Date(d.Year, d.Month, d.Day, 0, 0, 0, 0, loc)
}
// AddDays returns the date that is n days in the future.
// n can also be negative to go into the past.
func (d LocalDate) AddDays(n int) LocalDate {
return LocalDateOf(d.In(time.UTC).AddDate(0, 0, n))
}
// DaysSince returns the signed number of days between the date and s, not including the end day.
// This is the inverse operation to AddDays.
func (d LocalDate) DaysSince(s LocalDate) (days int) {
// We convert to Unix time so we do not have to worry about leap seconds:
// Unix time increases by exactly 86400 seconds per day.
deltaUnix := d.In(time.UTC).Unix() - s.In(time.UTC).Unix()
const secondsInADay = 86400
return int(deltaUnix / secondsInADay)
}
// Before reports whether d1 occurs before future date.
func (d LocalDate) Before(future LocalDate) bool {
if d.Year != future.Year {
return d.Year < future.Year
}
if d.Month != future.Month {
return d.Month < future.Month
}
return d.Day < future.Day
}
// After reports whether d1 occurs after past date.
func (d LocalDate) After(past LocalDate) bool {
return past.Before(d)
}
// MarshalText implements the encoding.TextMarshaler interface.
// The output is the result of d.String().
// MarshalText returns RFC 3339 representation of d.
func (d LocalDate) MarshalText() ([]byte, error) {
return []byte(d.String()), nil
}
// UnmarshalText implements the encoding.TextUnmarshaler interface.
// The date is expected to be a string in a format accepted by ParseLocalDate.
func (d *LocalDate) UnmarshalText(data []byte) error {
var err error
*d, err = ParseLocalDate(string(data))
return err
// UnmarshalText parses b using RFC 3339 to fill d.
func (d *LocalDate) UnmarshalText(b []byte) error {
res, err := parseLocalDate(b)
if err != nil {
return err
}
*d = res
return nil
}
// A LocalTime represents a time with nanosecond precision.
//
// This type does not include location information, and therefore does not
// describe a unique moment in time.
//
// This type exists to represent the TIME type in storage-based APIs like BigQuery.
// Most operations on Times are unlikely to be meaningful. Prefer the LocalDateTime type.
// LocalTime represents a time of day of no specific day in no specific
// timezone.
type LocalTime struct {
Hour int // The hour of the day in 24-hour format; range [0-23]
Minute int // The minute of the hour; range [0-59]
Second int // The second of the minute; range [0-59]
Nanosecond int // The nanosecond of the second; range [0-999999999]
Hour int // Hour of the day: [0; 24[
Minute int // Minute of the hour: [0; 60[
Second int // Second of the minute: [0; 60[
Nanosecond int // Nanoseconds within the second: [0, 1000000000[
Precision int // Number of digits to display for Nanosecond.
}
// LocalTimeOf returns the LocalTime representing the time of day in which a time occurs
// in that time's location. It ignores the date.
func LocalTimeOf(t time.Time) LocalTime {
var tm LocalTime
tm.Hour, tm.Minute, tm.Second = t.Clock()
tm.Nanosecond = t.Nanosecond()
// String returns RFC 3339 representation of d.
// If d.Nanosecond and d.Precision are zero, the time won't have a nanosecond
// component. If d.Nanosecond > 0 but d.Precision = 0, then the minimum number
// of digits for nanoseconds is provided.
func (d LocalTime) String() string {
s := fmt.Sprintf("%02d:%02d:%02d", d.Hour, d.Minute, d.Second)
return tm
if d.Precision > 0 {
s += fmt.Sprintf(".%09d", d.Nanosecond)[:d.Precision+1]
} else if d.Nanosecond > 0 {
// Nanoseconds are specified, but precision is not provided. Use the
// minimum.
s += strings.Trim(fmt.Sprintf(".%09d", d.Nanosecond), "0")
}
return s
}
// ParseLocalTime parses a string and returns the time value it represents.
// ParseLocalTime accepts an extended form of the RFC3339 partial-time format. After
// the HH:MM:SS part of the string, an optional fractional part may appear,
// consisting of a decimal point followed by one to nine decimal digits.
// (RFC3339 admits only one digit after the decimal point).
func ParseLocalTime(s string) (LocalTime, error) {
t, err := time.Parse("15:04:05.999999999", s)
// MarshalText returns RFC 3339 representation of d.
func (d LocalTime) MarshalText() ([]byte, error) {
return []byte(d.String()), nil
}
// UnmarshalText parses b using RFC 3339 to fill d.
func (d *LocalTime) UnmarshalText(b []byte) error {
res, left, err := parseLocalTime(b)
if err == nil && len(left) != 0 {
err = newDecodeError(left, "extra characters")
}
if err != nil {
return LocalTime{}, err
return err
}
return LocalTimeOf(t), nil
*d = res
return nil
}
// String returns the date in the format described in ParseLocalTime. If Nanoseconds
// is zero, no fractional part will be generated. Otherwise, the result will
// end with a fractional part consisting of a decimal point and nine digits.
func (t LocalTime) String() string {
s := fmt.Sprintf("%02d:%02d:%02d", t.Hour, t.Minute, t.Second)
if t.Nanosecond == 0 {
return s
}
return s + fmt.Sprintf(".%09d", t.Nanosecond)
}
// IsValid reports whether the time is valid.
func (t LocalTime) IsValid() bool {
// Construct a non-zero time.
tm := time.Date(2, 2, 2, t.Hour, t.Minute, t.Second, t.Nanosecond, time.UTC)
return LocalTimeOf(tm) == t
}
// MarshalText implements the encoding.TextMarshaler interface.
// The output is the result of t.String().
func (t LocalTime) MarshalText() ([]byte, error) {
return []byte(t.String()), nil
}
// UnmarshalText implements the encoding.TextUnmarshaler interface.
// The time is expected to be a string in a format accepted by ParseLocalTime.
func (t *LocalTime) UnmarshalText(data []byte) error {
var err error
*t, err = ParseLocalTime(string(data))
return err
}
// A LocalDateTime represents a date and time.
//
// This type does not include location information, and therefore does not
// describe a unique moment in time.
// LocalDateTime represents a time of a specific day in no specific timezone.
type LocalDateTime struct {
Date LocalDate
Time LocalTime
LocalDate
LocalTime
}
// Note: We deliberately do not embed LocalDate into LocalDateTime, to avoid promoting AddDays and Sub.
// AsTime converts d into a specific time instance in zone.
func (d LocalDateTime) AsTime(zone *time.Location) time.Time {
return time.Date(d.Year, time.Month(d.Month), d.Day, d.Hour, d.Minute, d.Second, d.Nanosecond, zone)
}
// LocalDateTimeOf returns the LocalDateTime in which a time occurs in that time's location.
func LocalDateTimeOf(t time.Time) LocalDateTime {
return LocalDateTime{
Date: LocalDateOf(t),
Time: LocalTimeOf(t),
// String returns RFC 3339 representation of d.
func (d LocalDateTime) String() string {
return d.LocalDate.String() + "T" + d.LocalTime.String()
}
// MarshalText returns RFC 3339 representation of d.
func (d LocalDateTime) MarshalText() ([]byte, error) {
return []byte(d.String()), nil
}
// UnmarshalText parses b using RFC 3339 to fill d.
func (d *LocalDateTime) UnmarshalText(data []byte) error {
res, left, err := parseLocalDateTime(data)
if err == nil && len(left) != 0 {
err = newDecodeError(left, "extra characters")
}
}
// ParseLocalDateTime parses a string and returns the LocalDateTime it represents.
// ParseLocalDateTime accepts a variant of the RFC3339 date-time format that omits
// the time offset but includes an optional fractional time, as described in
// ParseLocalTime. Informally, the accepted format is
// YYYY-MM-DDTHH:MM:SS[.FFFFFFFFF]
// where the 'T' may be a lower-case 't'.
func ParseLocalDateTime(s string) (LocalDateTime, error) {
t, err := time.Parse("2006-01-02T15:04:05.999999999", s)
if err != nil {
t, err = time.Parse("2006-01-02t15:04:05.999999999", s)
if err != nil {
return LocalDateTime{}, err
}
return err
}
return LocalDateTimeOf(t), nil
}
// String returns the date in the format described in ParseLocalDate.
func (dt LocalDateTime) String() string {
return dt.Date.String() + "T" + dt.Time.String()
}
// IsValid reports whether the datetime is valid.
func (dt LocalDateTime) IsValid() bool {
return dt.Date.IsValid() && dt.Time.IsValid()
}
// In returns the time corresponding to the LocalDateTime in the given location.
//
// If the time is missing or ambigous at the location, In returns the same
// result as time.LocalDate. For example, if loc is America/Indiana/Vincennes, then
// both
// time.LocalDate(1955, time.May, 1, 0, 30, 0, 0, loc)
// and
// civil.LocalDateTime{
// civil.LocalDate{Year: 1955, Month: time.May, Day: 1}},
// civil.LocalTime{Minute: 30}}.In(loc)
// return 23:30:00 on April 30, 1955.
//
// In panics if loc is nil.
func (dt LocalDateTime) In(loc *time.Location) time.Time {
return time.Date(
dt.Date.Year, dt.Date.Month, dt.Date.Day,
dt.Time.Hour, dt.Time.Minute, dt.Time.Second, dt.Time.Nanosecond, loc,
)
}
// Before reports whether dt occurs before future.
func (dt LocalDateTime) Before(future LocalDateTime) bool {
return dt.In(time.UTC).Before(future.In(time.UTC))
}
// After reports whether dt occurs after past.
func (dt LocalDateTime) After(past LocalDateTime) bool {
return past.Before(dt)
}
// MarshalText implements the encoding.TextMarshaler interface.
// The output is the result of dt.String().
func (dt LocalDateTime) MarshalText() ([]byte, error) {
return []byte(dt.String()), nil
}
// UnmarshalText implements the encoding.TextUnmarshaler interface.
// The datetime is expected to be a string in a format accepted by ParseLocalDateTime.
func (dt *LocalDateTime) UnmarshalText(data []byte) error {
var err error
*dt, err = ParseLocalDateTime(string(data))
return err
*d = res
return nil
}
+83 -472
View File
@@ -1,507 +1,118 @@
// Copyright 2016 Google LLC
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package toml
package toml_test
import (
"encoding/json"
"reflect"
"testing"
"time"
"github.com/pelletier/go-toml/v2"
"github.com/stretchr/testify/require"
)
func cmpEqual(x, y interface{}) bool {
return reflect.DeepEqual(x, y)
func TestLocalDate_AsTime(t *testing.T) {
d := toml.LocalDate{2021, 6, 8}
cast := d.AsTime(time.UTC)
require.Equal(t, time.Date(2021, time.June, 8, 0, 0, 0, 0, time.UTC), cast)
}
func TestDates(t *testing.T) {
t.Parallel()
for _, test := range []struct {
date LocalDate
loc *time.Location
wantStr string
wantTime time.Time
}{
{
date: LocalDate{2014, 7, 29},
loc: time.Local,
wantStr: "2014-07-29",
wantTime: time.Date(2014, time.July, 29, 0, 0, 0, 0, time.Local),
},
{
date: LocalDateOf(time.Date(2014, 8, 20, 15, 8, 43, 1, time.Local)),
loc: time.UTC,
wantStr: "2014-08-20",
wantTime: time.Date(2014, 8, 20, 0, 0, 0, 0, time.UTC),
},
{
date: LocalDateOf(time.Date(999, time.January, 26, 0, 0, 0, 0, time.Local)),
loc: time.UTC,
wantStr: "0999-01-26",
wantTime: time.Date(999, 1, 26, 0, 0, 0, 0, time.UTC),
},
} {
if got := test.date.String(); got != test.wantStr {
t.Errorf("%#v.String() = %q, want %q", test.date, got, test.wantStr)
}
if got := test.date.In(test.loc); !got.Equal(test.wantTime) {
t.Errorf("%#v.In(%v) = %v, want %v", test.date, test.loc, got, test.wantTime)
}
}
func TestLocalDate_String(t *testing.T) {
d := toml.LocalDate{2021, 6, 8}
require.Equal(t, "2021-06-08", d.String())
}
func TestDateIsValid(t *testing.T) {
t.Parallel()
for _, test := range []struct {
date LocalDate
want bool
}{
{LocalDate{2014, 7, 29}, true},
{LocalDate{2000, 2, 29}, true},
{LocalDate{10000, 12, 31}, true},
{LocalDate{1, 1, 1}, true},
{LocalDate{0, 1, 1}, true}, // year zero is OK
{LocalDate{-1, 1, 1}, true}, // negative year is OK
{LocalDate{1, 0, 1}, false},
{LocalDate{1, 1, 0}, false},
{LocalDate{2016, 1, 32}, false},
{LocalDate{2016, 13, 1}, false},
{LocalDate{1, -1, 1}, false},
{LocalDate{1, 1, -1}, false},
} {
got := test.date.IsValid()
if got != test.want {
t.Errorf("%#v: got %t, want %t", test.date, got, test.want)
}
}
func TestLocalDate_MarshalText(t *testing.T) {
d := toml.LocalDate{2021, 6, 8}
b, err := d.MarshalText()
require.NoError(t, err)
require.Equal(t, []byte("2021-06-08"), b)
}
func TestParseDate(t *testing.T) {
t.Parallel()
func TestLocalDate_UnmarshalMarshalText(t *testing.T) {
d := toml.LocalDate{}
err := d.UnmarshalText([]byte("2021-06-08"))
require.NoError(t, err)
require.Equal(t, toml.LocalDate{2021, 6, 8}, d)
var emptyDate LocalDate
for _, test := range []struct {
str string
want LocalDate // if empty, expect an error
}{
{"2016-01-02", LocalDate{2016, 1, 2}},
{"2016-12-31", LocalDate{2016, 12, 31}},
{"0003-02-04", LocalDate{3, 2, 4}},
{"999-01-26", emptyDate},
{"", emptyDate},
{"2016-01-02x", emptyDate},
} {
got, err := ParseLocalDate(test.str)
if got != test.want {
t.Errorf("ParseLocalDate(%q) = %+v, want %+v", test.str, got, test.want)
}
if err != nil && test.want != (emptyDate) {
t.Errorf("Unexpected error %v from ParseLocalDate(%q)", err, test.str)
}
}
err = d.UnmarshalText([]byte("what"))
require.Error(t, err)
}
func TestDateArithmetic(t *testing.T) {
t.Parallel()
for _, test := range []struct {
desc string
start LocalDate
end LocalDate
days int
}{
{
desc: "zero days noop",
start: LocalDate{2014, 5, 9},
end: LocalDate{2014, 5, 9},
days: 0,
},
{
desc: "crossing a year boundary",
start: LocalDate{2014, 12, 31},
end: LocalDate{2015, 1, 1},
days: 1,
},
{
desc: "negative number of days",
start: LocalDate{2015, 1, 1},
end: LocalDate{2014, 12, 31},
days: -1,
},
{
desc: "full leap year",
start: LocalDate{2004, 1, 1},
end: LocalDate{2005, 1, 1},
days: 366,
},
{
desc: "full non-leap year",
start: LocalDate{2001, 1, 1},
end: LocalDate{2002, 1, 1},
days: 365,
},
{
desc: "crossing a leap second",
start: LocalDate{1972, 6, 30},
end: LocalDate{1972, 7, 1},
days: 1,
},
{
desc: "dates before the unix epoch",
start: LocalDate{101, 1, 1},
end: LocalDate{102, 1, 1},
days: 365,
},
} {
if got := test.start.AddDays(test.days); got != test.end {
t.Errorf("[%s] %#v.AddDays(%v) = %#v, want %#v", test.desc, test.start, test.days, got, test.end)
}
if got := test.end.DaysSince(test.start); got != test.days {
t.Errorf("[%s] %#v.Sub(%#v) = %v, want %v", test.desc, test.end, test.start, got, test.days)
}
}
func TestLocalTime_String(t *testing.T) {
d := toml.LocalTime{20, 12, 1, 2, 9}
require.Equal(t, "20:12:01.000000002", d.String())
d = toml.LocalTime{20, 12, 1, 0, 0}
require.Equal(t, "20:12:01", d.String())
d = toml.LocalTime{20, 12, 1, 0, 9}
require.Equal(t, "20:12:01.000000000", d.String())
d = toml.LocalTime{20, 12, 1, 100, 0}
require.Equal(t, "20:12:01.0000001", d.String())
}
func TestDateBefore(t *testing.T) {
t.Parallel()
for _, test := range []struct {
d1, d2 LocalDate
want bool
}{
{LocalDate{2016, 12, 31}, LocalDate{2017, 1, 1}, true},
{LocalDate{2016, 1, 1}, LocalDate{2016, 1, 1}, false},
{LocalDate{2016, 12, 30}, LocalDate{2016, 12, 31}, true},
{LocalDate{2016, 1, 30}, LocalDate{2016, 12, 31}, true},
} {
if got := test.d1.Before(test.d2); got != test.want {
t.Errorf("%v.Before(%v): got %t, want %t", test.d1, test.d2, got, test.want)
}
}
func TestLocalTime_MarshalText(t *testing.T) {
d := toml.LocalTime{20, 12, 1, 2, 9}
b, err := d.MarshalText()
require.NoError(t, err)
require.Equal(t, []byte("20:12:01.000000002"), b)
}
func TestDateAfter(t *testing.T) {
t.Parallel()
func TestLocalTime_UnmarshalMarshalText(t *testing.T) {
d := toml.LocalTime{}
err := d.UnmarshalText([]byte("20:12:01.000000002"))
require.NoError(t, err)
require.Equal(t, toml.LocalTime{20, 12, 1, 2, 9}, d)
for _, test := range []struct {
d1, d2 LocalDate
want bool
}{
{LocalDate{2016, 12, 31}, LocalDate{2017, 1, 1}, false},
{LocalDate{2016, 1, 1}, LocalDate{2016, 1, 1}, false},
{LocalDate{2016, 12, 30}, LocalDate{2016, 12, 31}, false},
} {
if got := test.d1.After(test.d2); got != test.want {
t.Errorf("%v.After(%v): got %t, want %t", test.d1, test.d2, got, test.want)
}
}
err = d.UnmarshalText([]byte("what"))
require.Error(t, err)
err = d.UnmarshalText([]byte("20:12:01.000000002 bad"))
require.Error(t, err)
}
func TestTimeToString(t *testing.T) {
t.Parallel()
for _, test := range []struct {
str string
time LocalTime
roundTrip bool // ParseLocalTime(str).String() == str?
}{
{"13:26:33", LocalTime{13, 26, 33, 0}, true},
{"01:02:03.000023456", LocalTime{1, 2, 3, 23456}, true},
{"00:00:00.000000001", LocalTime{0, 0, 0, 1}, true},
{"13:26:03.1", LocalTime{13, 26, 3, 100000000}, false},
{"13:26:33.0000003", LocalTime{13, 26, 33, 300}, false},
} {
gotTime, err := ParseLocalTime(test.str)
if err != nil {
t.Errorf("ParseLocalTime(%q): got error: %v", test.str, err)
continue
}
if gotTime != test.time {
t.Errorf("ParseLocalTime(%q) = %+v, want %+v", test.str, gotTime, test.time)
}
if test.roundTrip {
gotStr := test.time.String()
if gotStr != test.str {
t.Errorf("%#v.String() = %q, want %q", test.time, gotStr, test.str)
}
}
}
func TestLocalTime_RoundTrip(t *testing.T) {
var d struct{ A toml.LocalTime }
err := toml.Unmarshal([]byte("a=20:12:01.500"), &d)
require.NoError(t, err)
require.Equal(t, "20:12:01.500", d.A.String())
}
func TestTimeOf(t *testing.T) {
t.Parallel()
for _, test := range []struct {
time time.Time
want LocalTime
}{
{time.Date(2014, 8, 20, 15, 8, 43, 1, time.Local), LocalTime{15, 8, 43, 1}},
{time.Date(1, 1, 1, 0, 0, 0, 0, time.UTC), LocalTime{0, 0, 0, 0}},
} {
if got := LocalTimeOf(test.time); got != test.want {
t.Errorf("LocalTimeOf(%v) = %+v, want %+v", test.time, got, test.want)
}
func TestLocalDateTime_AsTime(t *testing.T) {
d := toml.LocalDateTime{
toml.LocalDate{2021, 6, 8},
toml.LocalTime{20, 12, 1, 2, 9},
}
cast := d.AsTime(time.UTC)
require.Equal(t, time.Date(2021, time.June, 8, 20, 12, 1, 2, time.UTC), cast)
}
func TestTimeIsValid(t *testing.T) {
t.Parallel()
for _, test := range []struct {
time LocalTime
want bool
}{
{LocalTime{0, 0, 0, 0}, true},
{LocalTime{23, 0, 0, 0}, true},
{LocalTime{23, 59, 59, 999999999}, true},
{LocalTime{24, 59, 59, 999999999}, false},
{LocalTime{23, 60, 59, 999999999}, false},
{LocalTime{23, 59, 60, 999999999}, false},
{LocalTime{23, 59, 59, 1000000000}, false},
{LocalTime{-1, 0, 0, 0}, false},
{LocalTime{0, -1, 0, 0}, false},
{LocalTime{0, 0, -1, 0}, false},
{LocalTime{0, 0, 0, -1}, false},
} {
got := test.time.IsValid()
if got != test.want {
t.Errorf("%#v: got %t, want %t", test.time, got, test.want)
}
func TestLocalDateTime_String(t *testing.T) {
d := toml.LocalDateTime{
toml.LocalDate{2021, 6, 8},
toml.LocalTime{20, 12, 1, 2, 9},
}
require.Equal(t, "2021-06-08T20:12:01.000000002", d.String())
}
func TestDateTimeToString(t *testing.T) {
t.Parallel()
for _, test := range []struct {
str string
dateTime LocalDateTime
roundTrip bool // ParseLocalDateTime(str).String() == str?
}{
{"2016-03-22T13:26:33", LocalDateTime{LocalDate{2016, 3, 22}, LocalTime{13, 26, 33, 0}}, true},
{"2016-03-22T13:26:33.000000600", LocalDateTime{LocalDate{2016, 3, 22}, LocalTime{13, 26, 33, 600}}, true},
{"2016-03-22t13:26:33", LocalDateTime{LocalDate{2016, 3, 22}, LocalTime{13, 26, 33, 0}}, false},
} {
gotDateTime, err := ParseLocalDateTime(test.str)
if err != nil {
t.Errorf("ParseLocalDateTime(%q): got error: %v", test.str, err)
continue
}
if gotDateTime != test.dateTime {
t.Errorf("ParseLocalDateTime(%q) = %+v, want %+v", test.str, gotDateTime, test.dateTime)
}
if test.roundTrip {
gotStr := test.dateTime.String()
if gotStr != test.str {
t.Errorf("%#v.String() = %q, want %q", test.dateTime, gotStr, test.str)
}
}
func TestLocalDateTime_MarshalText(t *testing.T) {
d := toml.LocalDateTime{
toml.LocalDate{2021, 6, 8},
toml.LocalTime{20, 12, 1, 2, 9},
}
b, err := d.MarshalText()
require.NoError(t, err)
require.Equal(t, []byte("2021-06-08T20:12:01.000000002"), b)
}
func TestParseDateTimeErrors(t *testing.T) {
t.Parallel()
func TestLocalDateTime_UnmarshalMarshalText(t *testing.T) {
d := toml.LocalDateTime{}
err := d.UnmarshalText([]byte("2021-06-08 20:12:01.000000002"))
require.NoError(t, err)
require.Equal(t, toml.LocalDateTime{
toml.LocalDate{2021, 6, 8},
toml.LocalTime{20, 12, 1, 2, 9},
}, d)
for _, str := range []string{
"",
"2016-03-22", // just a date
"13:26:33", // just a time
"2016-03-22 13:26:33", // wrong separating character
"2016-03-22T13:26:33x", // extra at end
} {
if _, err := ParseLocalDateTime(str); err == nil {
t.Errorf("ParseLocalDateTime(%q) succeeded, want error", str)
}
}
}
func TestDateTimeOf(t *testing.T) {
t.Parallel()
for _, test := range []struct {
time time.Time
want LocalDateTime
}{
{
time.Date(2014, 8, 20, 15, 8, 43, 1, time.Local),
LocalDateTime{LocalDate{2014, 8, 20}, LocalTime{15, 8, 43, 1}},
},
{
time.Date(1, 1, 1, 0, 0, 0, 0, time.UTC),
LocalDateTime{LocalDate{1, 1, 1}, LocalTime{0, 0, 0, 0}},
},
} {
if got := LocalDateTimeOf(test.time); got != test.want {
t.Errorf("LocalDateTimeOf(%v) = %+v, want %+v", test.time, got, test.want)
}
}
}
func TestDateTimeIsValid(t *testing.T) {
t.Parallel()
// No need to be exhaustive here; it's just LocalDate.IsValid && LocalTime.IsValid.
for _, test := range []struct {
dt LocalDateTime
want bool
}{
{LocalDateTime{LocalDate{2016, 3, 20}, LocalTime{0, 0, 0, 0}}, true},
{LocalDateTime{LocalDate{2016, -3, 20}, LocalTime{0, 0, 0, 0}}, false},
{LocalDateTime{LocalDate{2016, 3, 20}, LocalTime{24, 0, 0, 0}}, false},
} {
got := test.dt.IsValid()
if got != test.want {
t.Errorf("%#v: got %t, want %t", test.dt, got, test.want)
}
}
}
func TestDateTimeIn(t *testing.T) {
t.Parallel()
dt := LocalDateTime{LocalDate{2016, 1, 2}, LocalTime{3, 4, 5, 6}}
want := time.Date(2016, 1, 2, 3, 4, 5, 6, time.UTC)
if got := dt.In(time.UTC); !got.Equal(want) {
t.Errorf("got %v, want %v", got, want)
}
}
func TestDateTimeBefore(t *testing.T) {
t.Parallel()
d1 := LocalDate{2016, 12, 31}
d2 := LocalDate{2017, 1, 1}
t1 := LocalTime{5, 6, 7, 8}
t2 := LocalTime{5, 6, 7, 9}
for _, test := range []struct {
dt1, dt2 LocalDateTime
want bool
}{
{LocalDateTime{d1, t1}, LocalDateTime{d2, t1}, true},
{LocalDateTime{d1, t1}, LocalDateTime{d1, t2}, true},
{LocalDateTime{d2, t1}, LocalDateTime{d1, t1}, false},
{LocalDateTime{d2, t1}, LocalDateTime{d2, t1}, false},
} {
if got := test.dt1.Before(test.dt2); got != test.want {
t.Errorf("%v.Before(%v): got %t, want %t", test.dt1, test.dt2, got, test.want)
}
}
}
func TestDateTimeAfter(t *testing.T) {
t.Parallel()
d1 := LocalDate{2016, 12, 31}
d2 := LocalDate{2017, 1, 1}
t1 := LocalTime{5, 6, 7, 8}
t2 := LocalTime{5, 6, 7, 9}
for _, test := range []struct {
dt1, dt2 LocalDateTime
want bool
}{
{LocalDateTime{d1, t1}, LocalDateTime{d2, t1}, false},
{LocalDateTime{d1, t1}, LocalDateTime{d1, t2}, false},
{LocalDateTime{d2, t1}, LocalDateTime{d1, t1}, true},
{LocalDateTime{d2, t1}, LocalDateTime{d2, t1}, false},
} {
if got := test.dt1.After(test.dt2); got != test.want {
t.Errorf("%v.After(%v): got %t, want %t", test.dt1, test.dt2, got, test.want)
}
}
}
func TestMarshalJSON(t *testing.T) {
t.Parallel()
for _, test := range []struct {
value interface{}
want string
}{
{LocalDate{1987, 4, 15}, `"1987-04-15"`},
{LocalTime{18, 54, 2, 0}, `"18:54:02"`},
{LocalDateTime{LocalDate{1987, 4, 15}, LocalTime{18, 54, 2, 0}}, `"1987-04-15T18:54:02"`},
} {
bgot, err := json.Marshal(test.value)
if err != nil {
t.Fatal(err)
}
if got := string(bgot); got != test.want {
t.Errorf("%#v: got %s, want %s", test.value, got, test.want)
}
}
}
func TestUnmarshalJSON(t *testing.T) {
t.Parallel()
var (
d LocalDate
tm LocalTime
dt LocalDateTime
)
for _, test := range []struct {
data string
ptr interface{}
want interface{}
}{
{`"1987-04-15"`, &d, &LocalDate{1987, 4, 15}},
{`"1987-04-\u0031\u0035"`, &d, &LocalDate{1987, 4, 15}},
{`"18:54:02"`, &tm, &LocalTime{18, 54, 2, 0}},
{`"1987-04-15T18:54:02"`, &dt, &LocalDateTime{LocalDate{1987, 4, 15}, LocalTime{18, 54, 2, 0}}},
} {
if err := json.Unmarshal([]byte(test.data), test.ptr); err != nil {
t.Fatalf("%s: %v", test.data, err)
}
if !cmpEqual(test.ptr, test.want) {
t.Errorf("%s: got %#v, want %#v", test.data, test.ptr, test.want)
}
}
for _, bad := range []string{
"", `""`, `"bad"`, `"1987-04-15x"`,
`19870415`, // a JSON number
`11987-04-15x`, // not a JSON string
} {
if json.Unmarshal([]byte(bad), &d) == nil {
t.Errorf("%q, LocalDate: got nil, want error", bad)
}
if json.Unmarshal([]byte(bad), &tm) == nil {
t.Errorf("%q, LocalTime: got nil, want error", bad)
}
if json.Unmarshal([]byte(bad), &dt) == nil {
t.Errorf("%q, LocalDateTime: got nil, want error", bad)
}
}
err = d.UnmarshalText([]byte("what"))
require.Error(t, err)
err = d.UnmarshalText([]byte("2021-06-08 20:12:01.000000002 bad"))
require.Error(t, err)
}
+22 -8
View File
@@ -5,6 +5,7 @@ import (
"encoding"
"fmt"
"io"
"math"
"reflect"
"sort"
"strconv"
@@ -53,8 +54,9 @@ func NewEncoder(w io.Writer) *Encoder {
// inline tag:
//
// MyField `inline:"true"`
func (enc *Encoder) SetTablesInline(inline bool) {
func (enc *Encoder) SetTablesInline(inline bool) *Encoder {
enc.tablesInline = inline
return enc
}
// SetArraysMultiline forces the encoder to emit all arrays with one element per
@@ -63,20 +65,23 @@ func (enc *Encoder) SetTablesInline(inline bool) {
// This behavior can be controlled on an individual struct field basis with the multiline tag:
//
// MyField `multiline:"true"`
func (enc *Encoder) SetArraysMultiline(multiline bool) {
func (enc *Encoder) SetArraysMultiline(multiline bool) *Encoder {
enc.arraysMultiline = multiline
return enc
}
// SetIndentSymbol defines the string that should be used for indentation. The
// provided string is repeated for each indentation level. Defaults to two
// spaces.
func (enc *Encoder) SetIndentSymbol(s string) {
func (enc *Encoder) SetIndentSymbol(s string) *Encoder {
enc.indentSymbol = s
return enc
}
// SetIndentTables forces the encoder to intent tables and array tables.
func (enc *Encoder) SetIndentTables(indent bool) {
func (enc *Encoder) SetIndentTables(indent bool) *Encoder {
enc.indentTables = indent
return enc
}
// Encode writes a TOML representation of v to the stream.
@@ -244,9 +249,17 @@ func (enc *Encoder) encode(b []byte, ctx encoderCtx, v reflect.Value) ([]byte, e
case reflect.String:
b = enc.encodeString(b, v.String(), ctx.options)
case reflect.Float32:
b = strconv.AppendFloat(b, v.Float(), 'f', -1, 32)
if math.Trunc(v.Float()) == v.Float() {
b = strconv.AppendFloat(b, v.Float(), 'f', 1, 32)
} else {
b = strconv.AppendFloat(b, v.Float(), 'f', -1, 32)
}
case reflect.Float64:
b = strconv.AppendFloat(b, v.Float(), 'f', -1, 64)
if math.Trunc(v.Float()) == v.Float() {
b = strconv.AppendFloat(b, v.Float(), 'f', 1, 64)
} else {
b = strconv.AppendFloat(b, v.Float(), 'f', -1, 64)
}
case reflect.Bool:
if v.Bool() {
b = append(b, "true"...)
@@ -640,9 +653,10 @@ func (enc *Encoder) encodeTableInline(b []byte, ctx encoderCtx, t table) ([]byte
return b, nil
}
var textMarshalerType = reflect.TypeOf(new(encoding.TextMarshaler)).Elem()
func willConvertToTable(ctx encoderCtx, v reflect.Value) bool {
if !v.IsValid() {
return false
}
if v.Type() == timeType || v.Type().Implements(textMarshalerType) {
return false
}
+40 -23
View File
@@ -14,8 +14,6 @@ import (
//nolint:funlen
func TestMarshal(t *testing.T) {
t.Parallel()
someInt := 42
type structInline struct {
@@ -516,8 +514,6 @@ K = 42`,
for _, e := range examples {
e := e
t.Run(e.desc, func(t *testing.T) {
t.Parallel()
b, err := toml.Marshal(e.v)
if e.err {
require.Error(t, err)
@@ -555,7 +551,7 @@ K = 42`,
type flagsSetters []struct {
name string
f func(enc *toml.Encoder, flag bool)
f func(enc *toml.Encoder, flag bool) *toml.Encoder
}
var allFlags = flagsSetters{
@@ -609,8 +605,6 @@ func equalStringsIgnoreNewlines(t *testing.T, expected string, actual string) {
//nolint:funlen
func TestMarshalIndentTables(t *testing.T) {
t.Parallel()
examples := []struct {
desc string
v interface{}
@@ -661,8 +655,6 @@ root = 'value0'
for _, e := range examples {
e := e
t.Run(e.desc, func(t *testing.T) {
t.Parallel()
var buf strings.Builder
enc := toml.NewEncoder(&buf)
enc.SetIndentTables(true)
@@ -685,24 +677,18 @@ func (c *customTextMarshaler) MarshalText() ([]byte, error) {
}
func TestMarshalTextMarshaler_NoRoot(t *testing.T) {
t.Parallel()
c := customTextMarshaler{}
_, err := toml.Marshal(&c)
require.Error(t, err)
}
func TestMarshalTextMarshaler_Error(t *testing.T) {
t.Parallel()
m := map[string]interface{}{"a": &customTextMarshaler{value: 1}}
_, err := toml.Marshal(m)
require.Error(t, err)
}
func TestMarshalTextMarshaler_ErrorInline(t *testing.T) {
t.Parallel()
type s struct {
A map[string]interface{} `inline:"true"`
}
@@ -716,8 +702,6 @@ func TestMarshalTextMarshaler_ErrorInline(t *testing.T) {
}
func TestMarshalTextMarshaler(t *testing.T) {
t.Parallel()
m := map[string]interface{}{"a": &customTextMarshaler{value: 2}}
r, err := toml.Marshal(m)
require.NoError(t, err)
@@ -731,7 +715,6 @@ func (b *brokenWriter) Write([]byte) (int, error) {
}
func TestEncodeToBrokenWriter(t *testing.T) {
t.Parallel()
w := brokenWriter{}
enc := toml.NewEncoder(&w)
err := enc.Encode(map[string]string{"hello": "world"})
@@ -739,7 +722,6 @@ func TestEncodeToBrokenWriter(t *testing.T) {
}
func TestEncoderSetIndentSymbol(t *testing.T) {
t.Parallel()
var w strings.Builder
enc := toml.NewEncoder(&w)
enc.SetIndentTables(true)
@@ -753,8 +735,6 @@ func TestEncoderSetIndentSymbol(t *testing.T) {
}
func TestIssue436(t *testing.T) {
t.Parallel()
data := []byte(`{"a": [ { "b": { "c": "d" } } ]}`)
var v interface{}
@@ -774,8 +754,6 @@ c = 'd'
}
func TestIssue424(t *testing.T) {
t.Parallel()
type Message1 struct {
Text string
}
@@ -804,6 +782,22 @@ func TestIssue424(t *testing.T) {
require.Equal(t, msg2, msg2parsed)
}
func TestIssue567(t *testing.T) {
var m map[string]interface{}
err := toml.Unmarshal([]byte("A = 12:08:05"), &m)
require.NoError(t, err)
require.IsType(t, m["A"], toml.LocalTime{})
}
func TestIssue590(t *testing.T) {
type CustomType int
var cfg struct {
Option CustomType `toml:"option"`
}
err := toml.Unmarshal([]byte("option = 42"), &cfg)
require.NoError(t, err)
}
func ExampleMarshal() {
type MyConfig struct {
Version int
@@ -828,3 +822,26 @@ func ExampleMarshal() {
// Name = 'go-toml'
// Tags = ['go', 'toml']
}
func TestIssue571(t *testing.T) {
type Foo struct {
Float32 float32
Float64 float64
}
const closeEnough = 1e-9
foo := Foo{
Float32: 42,
Float64: 43,
}
b, err := toml.Marshal(foo)
require.NoError(t, err)
var foo2 Foo
err = toml.Unmarshal(b, &foo2)
require.NoError(t, err)
assert.InDelta(t, 42, foo2.Float32, closeEnough)
assert.InDelta(t, 43, foo2.Float64, closeEnough)
}
+199 -93
View File
@@ -2,9 +2,10 @@ package toml
import (
"bytes"
"strconv"
"unicode"
"github.com/pelletier/go-toml/v2/internal/ast"
"github.com/pelletier/go-toml/v2/internal/danger"
)
type parser struct {
@@ -16,9 +17,20 @@ type parser struct {
first bool
}
func (p *parser) Range(b []byte) ast.Range {
return ast.Range{
Offset: uint32(danger.SubsliceOffset(p.data, b)),
Length: uint32(len(b)),
}
}
func (p *parser) Raw(raw ast.Range) []byte {
return p.data[raw.Offset : raw.Offset+raw.Length]
}
func (p *parser) Reset(b []byte) {
p.builder.Reset()
p.ref = ast.Reference{}
p.ref = ast.InvalidReference
p.data = b
p.left = b
p.err = nil
@@ -32,7 +44,7 @@ func (p *parser) NextExpression() bool {
}
p.builder.Reset()
p.ref = ast.Reference{}
p.ref = ast.InvalidReference
for {
if len(p.left) == 0 || p.err != nil {
@@ -61,7 +73,7 @@ func (p *parser) NextExpression() bool {
}
}
func (p *parser) Expression() ast.Node {
func (p *parser) Expression() *ast.Node {
return p.builder.NodeAt(p.ref)
}
@@ -86,7 +98,7 @@ func (p *parser) parseExpression(b []byte) (ast.Reference, []byte, error) {
// expression = ws [ comment ]
// expression =/ ws keyval ws [ comment ]
// expression =/ ws table ws [ comment ]
var ref ast.Reference
ref := ast.InvalidReference
b = p.parseWhitespace(b)
@@ -95,9 +107,8 @@ func (p *parser) parseExpression(b []byte) (ast.Reference, []byte, error) {
}
if b[0] == '#' {
_, rest := scanComment(b)
return ref, rest, nil
_, rest, err := scanComment(b)
return ref, rest, err
}
if b[0] == '\n' || b[0] == '\r' {
@@ -118,9 +129,8 @@ func (p *parser) parseExpression(b []byte) (ast.Reference, []byte, error) {
b = p.parseWhitespace(b)
if len(b) > 0 && b[0] == '#' {
_, rest := scanComment(b)
return ref, rest, nil
_, rest, err := scanComment(b)
return ref, rest, err
}
return ref, b, nil
@@ -197,7 +207,7 @@ func (p *parser) parseKeyval(b []byte) (ast.Reference, []byte, error) {
key, b, err := p.parseKey(b)
if err != nil {
return ast.Reference{}, nil, err
return ast.InvalidReference, nil, err
}
// keyval-sep = ws %x3D ws ; =
@@ -205,12 +215,12 @@ func (p *parser) parseKeyval(b []byte) (ast.Reference, []byte, error) {
b = p.parseWhitespace(b)
if len(b) == 0 {
return ast.Reference{}, nil, newDecodeError(b, "expected = after a key, but the document ends there")
return ast.InvalidReference, nil, newDecodeError(b, "expected = after a key, but the document ends there")
}
b, err = expect('=', b)
if err != nil {
return ast.Reference{}, nil, err
return ast.InvalidReference, nil, err
}
b = p.parseWhitespace(b)
@@ -229,7 +239,7 @@ func (p *parser) parseKeyval(b []byte) (ast.Reference, []byte, error) {
//nolint:cyclop,funlen
func (p *parser) parseVal(b []byte) (ast.Reference, []byte, error) {
// val = string / boolean / array / inline-table / date-time / float / integer
var ref ast.Reference
ref := ast.InvalidReference
if len(b) == 0 {
return ref, nil, newDecodeError(b, "expected value, not eof")
@@ -240,32 +250,36 @@ func (p *parser) parseVal(b []byte) (ast.Reference, []byte, error) {
switch c {
case '"':
var raw []byte
var v []byte
if scanFollowsMultilineBasicStringDelimiter(b) {
v, b, err = p.parseMultilineBasicString(b)
raw, v, b, err = p.parseMultilineBasicString(b)
} else {
v, b, err = p.parseBasicString(b)
raw, v, b, err = p.parseBasicString(b)
}
if err == nil {
ref = p.builder.Push(ast.Node{
Kind: ast.String,
Raw: p.Range(raw),
Data: v,
})
}
return ref, b, err
case '\'':
var raw []byte
var v []byte
if scanFollowsMultilineLiteralStringDelimiter(b) {
v, b, err = p.parseMultilineLiteralString(b)
raw, v, b, err = p.parseMultilineLiteralString(b)
} else {
v, b, err = p.parseLiteralString(b)
raw, v, b, err = p.parseLiteralString(b)
}
if err == nil {
ref = p.builder.Push(ast.Node{
Kind: ast.String,
Raw: p.Range(raw),
Data: v,
})
}
@@ -310,13 +324,13 @@ func atmost(b []byte, n int) []byte {
return b[:n]
}
func (p *parser) parseLiteralString(b []byte) ([]byte, []byte, error) {
func (p *parser) parseLiteralString(b []byte) ([]byte, []byte, []byte, error) {
v, rest, err := scanLiteralString(b)
if err != nil {
return nil, nil, err
return nil, nil, nil, err
}
return v[1 : len(v)-1], rest, nil
return v, v[1 : len(v)-1], rest, nil
}
func (p *parser) parseInlineTable(b []byte) (ast.Reference, []byte, error) {
@@ -338,7 +352,13 @@ func (p *parser) parseInlineTable(b []byte) (ast.Reference, []byte, error) {
var err error
for len(b) > 0 {
previousB := b
b = p.parseWhitespace(b)
if len(b) == 0 {
return parent, nil, newDecodeError(previousB[:1], "inline table is incomplete")
}
if b[0] == '}' {
break
}
@@ -382,6 +402,7 @@ func (p *parser) parseValArray(b []byte) (ast.Reference, []byte, error) {
// array-values =/ ws-comment-newline val ws-comment-newline [ array-sep ]
// array-sep = %x2C ; , Comma
// ws-comment-newline = *( wschar / [ comment ] newline )
arrayStart := b
b = b[1:]
parent := p.builder.Push(ast.Node{
@@ -400,7 +421,7 @@ func (p *parser) parseValArray(b []byte) (ast.Reference, []byte, error) {
}
if len(b) == 0 {
return parent, nil, newDecodeError(b, "array is incomplete")
return parent, nil, newDecodeError(arrayStart[:1], "array is incomplete")
}
if b[0] == ']' {
@@ -417,6 +438,8 @@ func (p *parser) parseValArray(b []byte) (ast.Reference, []byte, error) {
if err != nil {
return parent, nil, err
}
} else if !first {
return parent, nil, newDecodeError(b[0:1], "array elements must be separated by commas")
}
// TOML allows trailing commas in arrays.
@@ -425,7 +448,6 @@ func (p *parser) parseValArray(b []byte) (ast.Reference, []byte, error) {
}
var valueRef ast.Reference
valueRef, b, err = p.parseVal(b)
if err != nil {
return parent, nil, err
@@ -456,7 +478,10 @@ func (p *parser) parseOptionalWhitespaceCommentNewline(b []byte) ([]byte, error)
b = p.parseWhitespace(b)
if len(b) > 0 && b[0] == '#' {
_, b = scanComment(b)
_, b, err = scanComment(b)
if err != nil {
return nil, err
}
}
if len(b) == 0 {
@@ -476,10 +501,10 @@ func (p *parser) parseOptionalWhitespaceCommentNewline(b []byte) ([]byte, error)
return b, nil
}
func (p *parser) parseMultilineLiteralString(b []byte) ([]byte, []byte, error) {
func (p *parser) parseMultilineLiteralString(b []byte) ([]byte, []byte, []byte, error) {
token, rest, err := scanMultilineLiteralString(b)
if err != nil {
return nil, nil, err
return nil, nil, nil, err
}
i := 3
@@ -491,11 +516,11 @@ func (p *parser) parseMultilineLiteralString(b []byte) ([]byte, []byte, error) {
i += 2
}
return token[i : len(token)-3], rest, err
return token, token[i : len(token)-3], rest, err
}
//nolint:funlen,gocognit,cyclop
func (p *parser) parseMultilineBasicString(b []byte) ([]byte, []byte, error) {
func (p *parser) parseMultilineBasicString(b []byte) ([]byte, []byte, []byte, error) {
// ml-basic-string = ml-basic-string-delim [ newline ] ml-basic-body
// ml-basic-string-delim
// ml-basic-string-delim = 3quotation-mark
@@ -506,13 +531,11 @@ func (p *parser) parseMultilineBasicString(b []byte) ([]byte, []byte, error) {
// mlb-quotes = 1*2quotation-mark
// mlb-unescaped = wschar / %x21 / %x23-5B / %x5D-7E / non-ascii
// mlb-escaped-nl = escape ws newline *( wschar / newline )
token, rest, err := scanMultilineBasicString(b)
token, escaped, rest, err := scanMultilineBasicString(b)
if err != nil {
return nil, nil, err
return nil, nil, nil, err
}
var builder bytes.Buffer
i := 3
// skip the immediate new line
@@ -522,9 +545,24 @@ func (p *parser) parseMultilineBasicString(b []byte) ([]byte, []byte, error) {
i += 2
}
// fast path
startIdx := i
endIdx := len(token) - len(`"""`)
if !escaped {
str := token[startIdx:endIdx]
verr := utf8TomlValidAlreadyEscaped(str)
if verr.Zero() {
return token, str, rest, nil
}
return nil, nil, nil, newDecodeError(str[verr.Index:verr.Index+verr.Size], "invalid UTF-8")
}
var builder bytes.Buffer
// The scanner ensures that the token starts and ends with quotes and that
// escapes are balanced.
for ; i < len(token)-3; i++ {
for i < len(token)-3 {
c := token[i]
//nolint:nestif
@@ -532,17 +570,29 @@ func (p *parser) parseMultilineBasicString(b []byte) ([]byte, []byte, error) {
// When the last non-whitespace character on a line is an unescaped \,
// it will be trimmed along with all whitespace (including newlines) up
// to the next non-whitespace character or closing delimiter.
if token[i+1] == '\n' || (token[i+1] == '\r' && token[i+2] == '\n') {
i++ // skip the \
isLastNonWhitespaceOnLine := false
j := 1
findEOLLoop:
for ; j < len(token)-3-i; j++ {
switch token[i+j] {
case ' ', '\t':
continue
case '\n':
isLastNonWhitespaceOnLine = true
}
break findEOLLoop
}
if isLastNonWhitespaceOnLine {
i += j
for ; i < len(token)-3; i++ {
c := token[i]
if !(c == '\n' || c == '\r' || c == ' ' || c == '\t') {
i--
break
}
}
i++
continue
}
@@ -564,30 +614,35 @@ func (p *parser) parseMultilineBasicString(b []byte) ([]byte, []byte, error) {
case 't':
builder.WriteByte('\t')
case 'u':
x, err := hexToString(atmost(token[i+1:], 4), 4)
x, err := hexToRune(atmost(token[i+1:], 4), 4)
if err != nil {
return nil, nil, err
return nil, nil, nil, err
}
builder.WriteString(x)
builder.WriteRune(x)
i += 4
case 'U':
x, err := hexToString(atmost(token[i+1:], 8), 8)
x, err := hexToRune(atmost(token[i+1:], 8), 8)
if err != nil {
return nil, nil, err
return nil, nil, nil, err
}
builder.WriteString(x)
builder.WriteRune(x)
i += 8
default:
return nil, nil, newDecodeError(token[i:i+1], "invalid escaped character %#U", c)
return nil, nil, nil, newDecodeError(token[i:i+1], "invalid escaped character %#U", c)
}
i++
} else {
builder.WriteByte(c)
size := utf8ValidNext(token[i:])
if size == 0 {
return nil, nil, nil, newDecodeError(token[i:i+1], "invalid character %#U", c)
}
builder.Write(token[i : i+size])
i += size
}
}
return builder.Bytes(), rest, nil
return token, builder.Bytes(), rest, nil
}
func (p *parser) parseKey(b []byte) (ast.Reference, []byte, error) {
@@ -599,13 +654,14 @@ func (p *parser) parseKey(b []byte) (ast.Reference, []byte, error) {
// dotted-key = simple-key 1*( dot-sep simple-key )
//
// dot-sep = ws %x2E ws ; . Period
key, b, err := p.parseSimpleKey(b)
raw, key, b, err := p.parseSimpleKey(b)
if err != nil {
return ast.Reference{}, nil, err
return ast.InvalidReference, nil, err
}
ref := p.builder.Push(ast.Node{
Kind: ast.Key,
Raw: p.Range(raw),
Data: key,
})
@@ -614,13 +670,14 @@ func (p *parser) parseKey(b []byte) (ast.Reference, []byte, error) {
if len(b) > 0 && b[0] == '.' {
b = p.parseWhitespace(b[1:])
key, b, err = p.parseSimpleKey(b)
raw, key, b, err = p.parseSimpleKey(b)
if err != nil {
return ref, nil, err
}
p.builder.PushAndChain(ast.Node{
Kind: ast.Key,
Raw: p.Range(raw),
Data: key,
})
} else {
@@ -631,14 +688,10 @@ func (p *parser) parseKey(b []byte) (ast.Reference, []byte, error) {
return ref, b, nil
}
func (p *parser) parseSimpleKey(b []byte) (key, rest []byte, err error) {
func (p *parser) parseSimpleKey(b []byte) (raw, key, rest []byte, err error) {
// simple-key = quoted-key / unquoted-key
// unquoted-key = 1*( ALPHA / DIGIT / %x2D / %x5F ) ; A-Z / a-z / 0-9 / - / _
// quoted-key = basic-string / literal-string
if len(b) == 0 {
return nil, nil, newDecodeError(b, "key is incomplete")
}
switch {
case b[0] == '\'':
return p.parseLiteralString(b)
@@ -646,14 +699,14 @@ func (p *parser) parseSimpleKey(b []byte) (key, rest []byte, err error) {
return p.parseBasicString(b)
case isUnquotedKeyChar(b[0]):
key, rest = scanUnquotedKey(b)
return key, rest, nil
return key, key, rest, nil
default:
return nil, nil, newDecodeError(b[0:1], "invalid character at start of key: %c", b[0])
return nil, nil, nil, newDecodeError(b[0:1], "invalid character at start of key: %c", b[0])
}
}
//nolint:funlen,cyclop
func (p *parser) parseBasicString(b []byte) ([]byte, []byte, error) {
func (p *parser) parseBasicString(b []byte) ([]byte, []byte, []byte, error) {
// basic-string = quotation-mark *basic-char quotation-mark
// quotation-mark = %x22 ; "
// basic-char = basic-unescaped / escaped
@@ -668,16 +721,33 @@ func (p *parser) parseBasicString(b []byte) ([]byte, []byte, error) {
// escape-seq-char =/ %x74 ; t tab U+0009
// escape-seq-char =/ %x75 4HEXDIG ; uXXXX U+XXXX
// escape-seq-char =/ %x55 8HEXDIG ; UXXXXXXXX U+XXXXXXXX
token, rest, err := scanBasicString(b)
token, escaped, rest, err := scanBasicString(b)
if err != nil {
return nil, nil, err
return nil, nil, nil, err
}
startIdx := len(`"`)
endIdx := len(token) - len(`"`)
// Fast path. If there is no escape sequence, the string should just be
// an UTF-8 encoded string, which is the same as Go. In that case,
// validate the string and return a direct reference to the buffer.
if !escaped {
str := token[startIdx:endIdx]
verr := utf8TomlValidAlreadyEscaped(str)
if verr.Zero() {
return token, str, rest, nil
}
return nil, nil, nil, newDecodeError(str[verr.Index:verr.Index+verr.Size], "invalid UTF-8")
}
i := startIdx
var builder bytes.Buffer
// The scanner ensures that the token starts and ends with quotes and that
// escapes are balanced.
for i := 1; i < len(token)-1; i++ {
for i < len(token)-1 {
c := token[i]
if c == '\\' {
i++
@@ -697,46 +767,65 @@ func (p *parser) parseBasicString(b []byte) ([]byte, []byte, error) {
case 't':
builder.WriteByte('\t')
case 'u':
x, err := hexToString(token[i+1:len(token)-1], 4)
x, err := hexToRune(token[i+1:len(token)-1], 4)
if err != nil {
return nil, nil, err
return nil, nil, nil, err
}
builder.WriteString(x)
builder.WriteRune(x)
i += 4
case 'U':
x, err := hexToString(token[i+1:len(token)-1], 8)
x, err := hexToRune(token[i+1:len(token)-1], 8)
if err != nil {
return nil, nil, err
return nil, nil, nil, err
}
builder.WriteString(x)
builder.WriteRune(x)
i += 8
default:
return nil, nil, newDecodeError(token[i:i+1], "invalid escaped character %#U", c)
return nil, nil, nil, newDecodeError(token[i:i+1], "invalid escaped character %#U", c)
}
i++
} else {
builder.WriteByte(c)
size := utf8ValidNext(token[i:])
if size == 0 {
return nil, nil, nil, newDecodeError(token[i:i+1], "invalid character %#U", c)
}
builder.Write(token[i : i+size])
i += size
}
}
return builder.Bytes(), rest, nil
return token, builder.Bytes(), rest, nil
}
func hexToString(b []byte, length int) (string, error) {
func hexToRune(b []byte, length int) (rune, error) {
if len(b) < length {
return "", newDecodeError(b, "unicode point needs %d character, not %d", length, len(b))
return -1, newDecodeError(b, "unicode point needs %d character, not %d", length, len(b))
}
b = b[:length]
//nolint:godox
// TODO: slow
intcode, err := strconv.ParseInt(string(b), 16, 32)
if err != nil {
return "", newDecodeError(b, "couldn't parse hexadecimal number: %w", err)
var r uint32
for i, c := range b {
d := uint32(0)
switch {
case '0' <= c && c <= '9':
d = uint32(c - '0')
case 'a' <= c && c <= 'f':
d = uint32(c - 'a' + 10)
case 'A' <= c && c <= 'F':
d = uint32(c - 'A' + 10)
default:
return -1, newDecodeError(b[i:i+1], "non-hex character")
}
r = r*16 + d
}
return string(rune(intcode)), nil
if r > unicode.MaxRune || 0xD800 <= r && r < 0xE000 {
return -1, newDecodeError(b, "escape sequence is invalid Unicode code point")
}
return rune(r), nil
}
func (p *parser) parseWhitespace(b []byte) []byte {
@@ -753,7 +842,7 @@ func (p *parser) parseIntOrFloatOrDateTime(b []byte) (ast.Reference, []byte, err
switch b[0] {
case 'i':
if !scanFollowsInf(b) {
return ast.Reference{}, nil, newDecodeError(atmost(b, 3), "expected 'inf'")
return ast.InvalidReference, nil, newDecodeError(atmost(b, 3), "expected 'inf'")
}
return p.builder.Push(ast.Node{
@@ -762,7 +851,7 @@ func (p *parser) parseIntOrFloatOrDateTime(b []byte) (ast.Reference, []byte, err
}), b[3:], nil
case 'n':
if !scanFollowsNan(b) {
return ast.Reference{}, nil, newDecodeError(atmost(b, 3), "expected 'nan'")
return ast.InvalidReference, nil, newDecodeError(atmost(b, 3), "expected 'nan'")
}
return p.builder.Push(ast.Node{
@@ -791,6 +880,8 @@ func (p *parser) parseIntOrFloatOrDateTime(b []byte) (ast.Reference, []byte, err
if idx == 2 && c == ':' || (idx == 4 && c == '-') {
return p.scanDateTime(b)
}
break
}
return p.scanIntOrFloat(b)
@@ -811,6 +902,7 @@ func digitsToInt(b []byte) int {
func (p *parser) scanDateTime(b []byte) (ast.Reference, []byte, error) {
// scans for contiguous characters in [0-9T:Z.+-], and up to one space if
// followed by a digit.
hasDate := false
hasTime := false
hasTz := false
seenSpace := false
@@ -823,17 +915,23 @@ byteLoop:
switch {
case isDigit(c):
case c == '-':
const offsetOfTz = 19
if i == offsetOfTz {
hasDate = true
const minOffsetOfTz = 8
if i >= minOffsetOfTz {
hasTz = true
}
case c == 'T' || c == ':' || c == '.':
case c == 'T' || c == 't' || c == ':' || c == '.':
hasTime = true
case c == '+' || c == '-' || c == 'Z':
case c == '+' || c == '-' || c == 'Z' || c == 'z':
hasTz = true
case c == ' ':
if !seenSpace && i+1 < len(b) && isDigit(b[i+1]) {
i += 2
// Avoid reaching past the end of the document in case the time
// is malformed. See TestIssue585.
if i >= len(b) {
i--
}
seenSpace = true
hasTime = true
} else {
@@ -847,10 +945,14 @@ byteLoop:
var kind ast.Kind
if hasTime {
if hasTz {
kind = ast.DateTime
if hasDate {
if hasTz {
kind = ast.DateTime
} else {
kind = ast.LocalDateTime
}
} else {
kind = ast.LocalDateTime
kind = ast.LocalTime
}
} else {
kind = ast.LocalDate
@@ -866,7 +968,7 @@ byteLoop:
func (p *parser) scanIntOrFloat(b []byte) (ast.Reference, []byte, error) {
i := 0
if len(b) > 2 && b[0] == '0' && b[1] != '.' {
if len(b) > 2 && b[0] == '0' && b[1] != '.' && b[1] != 'e' && b[1] != 'E' {
var isValidRune validRuneFn
switch b[1] {
@@ -918,7 +1020,7 @@ func (p *parser) scanIntOrFloat(b []byte) (ast.Reference, []byte, error) {
}), b[i+3:], nil
}
return ast.Reference{}, nil, newDecodeError(b[i:i+1], "unexpected character 'i' while scanning for a number")
return ast.InvalidReference, nil, newDecodeError(b[i:i+1], "unexpected character 'i' while scanning for a number")
}
if c == 'n' {
@@ -929,14 +1031,14 @@ func (p *parser) scanIntOrFloat(b []byte) (ast.Reference, []byte, error) {
}), b[i+3:], nil
}
return ast.Reference{}, nil, newDecodeError(b[i:i+1], "unexpected character 'n' while scanning for a number")
return ast.InvalidReference, nil, newDecodeError(b[i:i+1], "unexpected character 'n' while scanning for a number")
}
break
}
if i == 0 {
return ast.Reference{}, b, newDecodeError(b, "incomplete number")
return ast.InvalidReference, b, newDecodeError(b, "incomplete number")
}
kind := ast.Integer
@@ -973,8 +1075,12 @@ func isValidBinaryRune(r byte) bool {
}
func expect(x byte, b []byte) ([]byte, error) {
if len(b) == 0 {
return nil, newDecodeError(b, "expected character %c but the document ended here", x)
}
if b[0] != x {
return nil, newDecodeError(b[0:1], "expected character %U", x)
return nil, newDecodeError(b[0:1], "expected character %c", x)
}
return b[1:], nil
+101 -7
View File
@@ -1,6 +1,8 @@
package toml
import (
"strconv"
"strings"
"testing"
"github.com/pelletier/go-toml/v2/internal/ast"
@@ -9,8 +11,6 @@ import (
//nolint:funlen
func TestParser_AST_Numbers(t *testing.T) {
t.Parallel()
examples := []struct {
desc string
input string
@@ -137,7 +137,6 @@ func TestParser_AST_Numbers(t *testing.T) {
for _, e := range examples {
e := e
t.Run(e.desc, func(t *testing.T) {
t.Parallel()
p := parser{}
p.Reset([]byte(`A = ` + e.input))
p.NextExpression()
@@ -168,7 +167,7 @@ type (
}
)
func compareNode(t *testing.T, e astNode, n ast.Node) {
func compareNode(t *testing.T, e astNode, n *ast.Node) {
t.Helper()
require.Equal(t, e.Kind, n.Kind)
require.Equal(t, e.Data, n.Data)
@@ -200,8 +199,6 @@ func compareIterator(t *testing.T, expected []astNode, actual ast.Iterator) {
//nolint:funlen
func TestParser_AST(t *testing.T) {
t.Parallel()
examples := []struct {
desc string
input string
@@ -340,7 +337,6 @@ func TestParser_AST(t *testing.T) {
for _, e := range examples {
e := e
t.Run(e.desc, func(t *testing.T) {
t.Parallel()
p := parser{}
p.Reset([]byte(e.input))
p.NextExpression()
@@ -354,3 +350,101 @@ func TestParser_AST(t *testing.T) {
})
}
}
func BenchmarkParseBasicStringWithUnicode(b *testing.B) {
p := &parser{}
b.Run("4", func(b *testing.B) {
input := []byte(`"\u1234\u5678\u9ABC\u1234\u5678\u9ABC"`)
b.ReportAllocs()
b.SetBytes(int64(len(input)))
for i := 0; i < b.N; i++ {
p.parseBasicString(input)
}
})
b.Run("8", func(b *testing.B) {
input := []byte(`"\u12345678\u9ABCDEF0\u12345678\u9ABCDEF0"`)
b.ReportAllocs()
b.SetBytes(int64(len(input)))
for i := 0; i < b.N; i++ {
p.parseBasicString(input)
}
})
}
func BenchmarkParseBasicStringsEasy(b *testing.B) {
p := &parser{}
for _, size := range []int{1, 4, 8, 16, 21} {
b.Run(strconv.Itoa(size), func(b *testing.B) {
input := []byte(`"` + strings.Repeat("A", size) + `"`)
b.ReportAllocs()
b.SetBytes(int64(len(input)))
for i := 0; i < b.N; i++ {
p.parseBasicString(input)
}
})
}
}
func TestParser_AST_DateTimes(t *testing.T) {
examples := []struct {
desc string
input string
kind ast.Kind
err bool
}{
{
desc: "offset-date-time with delim 'T' and UTC offset",
input: `2021-07-21T12:08:05Z`,
kind: ast.DateTime,
},
{
desc: "offset-date-time with space delim and +8hours offset",
input: `2021-07-21 12:08:05+08:00`,
kind: ast.DateTime,
},
{
desc: "local-date-time with nano second",
input: `2021-07-21T12:08:05.666666666`,
kind: ast.LocalDateTime,
},
{
desc: "local-date-time",
input: `2021-07-21T12:08:05`,
kind: ast.LocalDateTime,
},
{
desc: "local-date",
input: `2021-07-21`,
kind: ast.LocalDate,
},
}
for _, e := range examples {
e := e
t.Run(e.desc, func(t *testing.T) {
p := parser{}
p.Reset([]byte(`A = ` + e.input))
p.NextExpression()
err := p.Error()
if e.err {
require.Error(t, err)
} else {
require.NoError(t, err)
expected := astNode{
Kind: ast.KeyValue,
Children: []astNode{
{Kind: e.kind, Data: []byte(e.input)},
{Kind: ast.Key, Data: []byte(`A`)},
},
}
compareNode(t, expected, p.Expression())
}
})
}
}
+96 -20
View File
@@ -49,13 +49,18 @@ func scanLiteralString(b []byte) ([]byte, []byte, error) {
// literal-string = apostrophe *literal-char apostrophe
// apostrophe = %x27 ; ' apostrophe
// literal-char = %x09 / %x20-26 / %x28-7E / non-ascii
for i := 1; i < len(b); i++ {
for i := 1; i < len(b); {
switch b[i] {
case '\'':
return b[:i+1], b[i+1:], nil
case '\n':
return nil, nil, newDecodeError(b[i:i+1], "literal strings cannot have new lines")
}
size := utf8ValidNext(b[i:])
if size == 0 {
return nil, nil, newDecodeError(b[i:i+1], "invalid character")
}
i += size
}
return nil, nil, newDecodeError(b[len(b):], "unterminated literal string")
@@ -70,10 +75,37 @@ func scanMultilineLiteralString(b []byte) ([]byte, []byte, error) {
// mll-content = mll-char / newline
// mll-char = %x09 / %x20-26 / %x28-7E / non-ascii
// mll-quotes = 1*2apostrophe
for i := 3; i < len(b); i++ {
if b[i] == '\'' && scanFollowsMultilineLiteralStringDelimiter(b[i:]) {
return b[:i+3], b[i+3:], nil
for i := 3; i < len(b); {
if scanFollowsMultilineLiteralStringDelimiter(b[i:]) {
i += 3
// At that point we found 3 apostrophe, and i is the
// index of the byte after the third one. The scanner
// needs to be eager, because there can be an extra 2
// apostrophe that can be accepted at the end of the
// string.
if i >= len(b) || b[i] != '\'' {
return b[:i], b[i:], nil
}
i++
if i >= len(b) || b[i] != '\'' {
return b[:i], b[i:], nil
}
i++
if i < len(b) && b[i] == '\'' {
return nil, nil, newDecodeError(b[i-3:i+1], "''' not allowed in multiline literal string")
}
return b[:i], b[i:], nil
}
size := utf8ValidNext(b[i:])
if size == 0 {
return nil, nil, newDecodeError(b[i:i+1], "invalid character")
}
i += size
}
return nil, nil, newDecodeError(b[len(b):], `multiline literal string not terminated by '''`)
@@ -106,45 +138,62 @@ func scanWhitespace(b []byte) ([]byte, []byte) {
}
//nolint:unparam
func scanComment(b []byte) ([]byte, []byte) {
func scanComment(b []byte) ([]byte, []byte, error) {
// comment-start-symbol = %x23 ; #
// non-ascii = %x80-D7FF / %xE000-10FFFF
// non-eol = %x09 / %x20-7F / non-ascii
//
// comment = comment-start-symbol *non-eol
for i := 1; i < len(b); i++ {
for i := 1; i < len(b); {
if b[i] == '\n' {
return b[:i], b[i:]
return b[:i], b[i:], nil
}
if b[i] == '\r' {
if i+1 < len(b) && b[i+1] == '\n' {
return b[:i+1], b[i+1:], nil
}
return nil, nil, newDecodeError(b[i:i+1], "invalid character in comment")
}
size := utf8ValidNext(b[i:])
if size == 0 {
return nil, nil, newDecodeError(b[i:i+1], "invalid character in comment")
}
i += size
}
return b, nil
return b, b[len(b):], nil
}
func scanBasicString(b []byte) ([]byte, []byte, error) {
func scanBasicString(b []byte) ([]byte, bool, []byte, error) {
// basic-string = quotation-mark *basic-char quotation-mark
// quotation-mark = %x22 ; "
// basic-char = basic-unescaped / escaped
// basic-unescaped = wschar / %x21 / %x23-5B / %x5D-7E / non-ascii
// escaped = escape escape-seq-char
for i := 1; i < len(b); i++ {
escaped := false
i := 1
for ; i < len(b); i++ {
switch b[i] {
case '"':
return b[:i+1], b[i+1:], nil
case '\n':
return nil, nil, newDecodeError(b[i:i+1], "basic strings cannot have new lines")
return b[:i+1], escaped, b[i+1:], nil
case '\n', '\r':
return nil, escaped, nil, newDecodeError(b[i:i+1], "basic strings cannot have new lines")
case '\\':
if len(b) < i+2 {
return nil, nil, newDecodeError(b[i:i+1], "need a character after \\")
return nil, escaped, nil, newDecodeError(b[i:i+1], "need a character after \\")
}
escaped = true
i++ // skip the next character
}
}
return nil, nil, newDecodeError(b[len(b):], `basic string not terminated by "`)
return nil, escaped, nil, newDecodeError(b[len(b):], `basic string not terminated by "`)
}
func scanMultilineBasicString(b []byte) ([]byte, []byte, error) {
func scanMultilineBasicString(b []byte) ([]byte, bool, []byte, error) {
// ml-basic-string = ml-basic-string-delim [ newline ] ml-basic-body
// ml-basic-string-delim
// ml-basic-string-delim = 3quotation-mark
@@ -155,19 +204,46 @@ func scanMultilineBasicString(b []byte) ([]byte, []byte, error) {
// mlb-quotes = 1*2quotation-mark
// mlb-unescaped = wschar / %x21 / %x23-5B / %x5D-7E / non-ascii
// mlb-escaped-nl = escape ws newline *( wschar / newline )
for i := 3; i < len(b); i++ {
escaped := false
i := 3
for ; i < len(b); i++ {
switch b[i] {
case '"':
if scanFollowsMultilineBasicStringDelimiter(b[i:]) {
return b[:i+3], b[i+3:], nil
i += 3
// At that point we found 3 apostrophe, and i is the
// index of the byte after the third one. The scanner
// needs to be eager, because there can be an extra 2
// apostrophe that can be accepted at the end of the
// string.
if i >= len(b) || b[i] != '"' {
return b[:i], escaped, b[i:], nil
}
i++
if i >= len(b) || b[i] != '"' {
return b[:i], escaped, b[i:], nil
}
i++
if i < len(b) && b[i] == '"' {
return nil, escaped, nil, newDecodeError(b[i-3:i+1], `""" not allowed in multiline basic string`)
}
return b[:i], escaped, b[i:], nil
}
case '\\':
if len(b) < i+2 {
return nil, nil, newDecodeError(b[len(b):], "need a character after \\")
return nil, escaped, nil, newDecodeError(b[len(b):], "need a character after \\")
}
escaped = true
i++ // skip the next character
}
}
return nil, nil, newDecodeError(b[len(b):], `multiline basic string not terminated by """`)
return nil, escaped, nil, newDecodeError(b[len(b):], `multiline basic string not terminated by """`)
}
+25 -6
View File
@@ -2,6 +2,7 @@ package toml
import (
"github.com/pelletier/go-toml/v2/internal/ast"
"github.com/pelletier/go-toml/v2/internal/danger"
"github.com/pelletier/go-toml/v2/internal/tracker"
)
@@ -14,7 +15,7 @@ type strict struct {
missing []decodeError
}
func (s *strict) EnterTable(node ast.Node) {
func (s *strict) EnterTable(node *ast.Node) {
if !s.Enabled {
return
}
@@ -22,7 +23,7 @@ func (s *strict) EnterTable(node ast.Node) {
s.key.UpdateTable(node)
}
func (s *strict) EnterArrayTable(node ast.Node) {
func (s *strict) EnterArrayTable(node *ast.Node) {
if !s.Enabled {
return
}
@@ -30,7 +31,7 @@ func (s *strict) EnterArrayTable(node ast.Node) {
s.key.UpdateArrayTable(node)
}
func (s *strict) EnterKeyValue(node ast.Node) {
func (s *strict) EnterKeyValue(node *ast.Node) {
if !s.Enabled {
return
}
@@ -38,7 +39,7 @@ func (s *strict) EnterKeyValue(node ast.Node) {
s.key.Push(node)
}
func (s *strict) ExitKeyValue(node ast.Node) {
func (s *strict) ExitKeyValue(node *ast.Node) {
if !s.Enabled {
return
}
@@ -46,7 +47,7 @@ func (s *strict) ExitKeyValue(node ast.Node) {
s.key.Pop(node)
}
func (s *strict) MissingTable(node ast.Node) {
func (s *strict) MissingTable(node *ast.Node) {
if !s.Enabled {
return
}
@@ -58,7 +59,7 @@ func (s *strict) MissingTable(node ast.Node) {
})
}
func (s *strict) MissingField(node ast.Node) {
func (s *strict) MissingField(node *ast.Node) {
if !s.Enabled {
return
}
@@ -86,3 +87,21 @@ func (s *strict) Error(doc []byte) error {
return err
}
func keyLocation(node *ast.Node) []byte {
k := node.Key()
hasOne := k.Next()
if !hasOne {
panic("should not be called with empty key")
}
start := k.Node().Data
end := k.Node().Data
for k.Next() {
end = k.Node().Data
}
return danger.BytesRange(start, end)
}
-536
View File
@@ -1,536 +0,0 @@
package toml
import (
"fmt"
"math"
"reflect"
"strings"
"sync"
)
type target interface {
// Dereferences the target.
get() reflect.Value
// Store a string at the target.
setString(v string)
// Store a boolean at the target
setBool(v bool)
// Store an int64 at the target
setInt64(v int64)
// Store a float64 at the target
setFloat64(v float64)
// Stores any value at the target
set(v reflect.Value)
}
// valueTarget just contains a reflect.Value that can be set.
// It is used for struct fields.
type valueTarget reflect.Value
func (t valueTarget) get() reflect.Value {
return reflect.Value(t)
}
func (t valueTarget) set(v reflect.Value) {
reflect.Value(t).Set(v)
}
func (t valueTarget) setString(v string) {
t.get().SetString(v)
}
func (t valueTarget) setBool(v bool) {
t.get().SetBool(v)
}
func (t valueTarget) setInt64(v int64) {
t.get().SetInt(v)
}
func (t valueTarget) setFloat64(v float64) {
t.get().SetFloat(v)
}
// interfaceTarget wraps an other target to dereference on get.
type interfaceTarget struct {
x target
}
func (t interfaceTarget) get() reflect.Value {
return t.x.get().Elem()
}
func (t interfaceTarget) set(v reflect.Value) {
t.x.set(v)
}
func (t interfaceTarget) setString(v string) {
panic("interface targets should always go through set")
}
func (t interfaceTarget) setBool(v bool) {
panic("interface targets should always go through set")
}
func (t interfaceTarget) setInt64(v int64) {
panic("interface targets should always go through set")
}
func (t interfaceTarget) setFloat64(v float64) {
panic("interface targets should always go through set")
}
// mapTarget targets a specific key of a map.
type mapTarget struct {
v reflect.Value
k reflect.Value
}
func (t mapTarget) get() reflect.Value {
return t.v.MapIndex(t.k)
}
func (t mapTarget) set(v reflect.Value) {
t.v.SetMapIndex(t.k, v)
}
func (t mapTarget) setString(v string) {
t.set(reflect.ValueOf(v))
}
func (t mapTarget) setBool(v bool) {
t.set(reflect.ValueOf(v))
}
func (t mapTarget) setInt64(v int64) {
t.set(reflect.ValueOf(v))
}
func (t mapTarget) setFloat64(v float64) {
t.set(reflect.ValueOf(v))
}
// makes sure that the value pointed at by t is indexable (Slice, Array), or
// dereferences to an indexable (Ptr, Interface).
func ensureValueIndexable(t target) error {
f := t.get()
switch f.Type().Kind() {
case reflect.Slice:
if f.IsNil() {
t.set(reflect.MakeSlice(f.Type(), 0, 0))
return nil
}
case reflect.Interface:
if f.IsNil() || f.Elem().Type() != sliceInterfaceType {
t.set(reflect.MakeSlice(sliceInterfaceType, 0, 0))
return nil
}
case reflect.Ptr:
panic("pointer should have already been dereferenced")
case reflect.Array:
// arrays are always initialized.
default:
return fmt.Errorf("toml: cannot store array in a %s", f.Kind())
}
return nil
}
var (
sliceInterfaceType = reflect.TypeOf([]interface{}{})
mapStringInterfaceType = reflect.TypeOf(map[string]interface{}{})
)
func ensureMapIfInterface(x target) {
v := x.get()
if v.Kind() == reflect.Interface && v.IsNil() {
newElement := reflect.MakeMap(mapStringInterfaceType)
x.set(newElement)
}
}
func setString(t target, v string) error {
f := t.get()
switch f.Kind() {
case reflect.String:
t.setString(v)
case reflect.Interface:
t.set(reflect.ValueOf(v))
default:
return fmt.Errorf("toml: cannot assign string to a %s", f.Kind())
}
return nil
}
func setBool(t target, v bool) error {
f := t.get()
switch f.Kind() {
case reflect.Bool:
t.setBool(v)
case reflect.Interface:
t.set(reflect.ValueOf(v))
default:
return fmt.Errorf("toml: cannot assign boolean to a %s", f.Kind())
}
return nil
}
const (
maxInt = int64(^uint(0) >> 1)
minInt = -maxInt - 1
)
//nolint:funlen,gocognit,cyclop
func setInt64(t target, v int64) error {
f := t.get()
switch f.Kind() {
case reflect.Int64:
t.setInt64(v)
case reflect.Int32:
if v < math.MinInt32 || v > math.MaxInt32 {
return fmt.Errorf("toml: number %d does not fit in an int32", v)
}
t.set(reflect.ValueOf(int32(v)))
return nil
case reflect.Int16:
if v < math.MinInt16 || v > math.MaxInt16 {
return fmt.Errorf("toml: number %d does not fit in an int16", v)
}
t.set(reflect.ValueOf(int16(v)))
case reflect.Int8:
if v < math.MinInt8 || v > math.MaxInt8 {
return fmt.Errorf("toml: number %d does not fit in an int8", v)
}
t.set(reflect.ValueOf(int8(v)))
case reflect.Int:
if v < minInt || v > maxInt {
return fmt.Errorf("toml: number %d does not fit in an int", v)
}
t.set(reflect.ValueOf(int(v)))
case reflect.Uint64:
if v < 0 {
return fmt.Errorf("toml: negative number %d does not fit in an uint64", v)
}
t.set(reflect.ValueOf(uint64(v)))
case reflect.Uint32:
if v < 0 || v > math.MaxUint32 {
return fmt.Errorf("toml: negative number %d does not fit in an uint32", v)
}
t.set(reflect.ValueOf(uint32(v)))
case reflect.Uint16:
if v < 0 || v > math.MaxUint16 {
return fmt.Errorf("toml: negative number %d does not fit in an uint16", v)
}
t.set(reflect.ValueOf(uint16(v)))
case reflect.Uint8:
if v < 0 || v > math.MaxUint8 {
return fmt.Errorf("toml: negative number %d does not fit in an uint8", v)
}
t.set(reflect.ValueOf(uint8(v)))
case reflect.Uint:
if v < 0 {
return fmt.Errorf("toml: negative number %d does not fit in an uint", v)
}
t.set(reflect.ValueOf(uint(v)))
case reflect.Interface:
t.set(reflect.ValueOf(v))
default:
return fmt.Errorf("toml: integer cannot be assigned to %s", f.Kind())
}
return nil
}
func setFloat64(t target, v float64) error {
f := t.get()
switch f.Kind() {
case reflect.Float64:
t.setFloat64(v)
case reflect.Float32:
if v > math.MaxFloat32 {
return fmt.Errorf("toml: number %f does not fit in a float32", v)
}
t.set(reflect.ValueOf(float32(v)))
case reflect.Interface:
t.set(reflect.ValueOf(v))
default:
return fmt.Errorf("toml: float cannot be assigned to %s", f.Kind())
}
return nil
}
// Returns the element at idx of the value pointed at by target, or an error if
// t does not point to an indexable.
// If the target points to an Array and idx is out of bounds, it returns
// (nil, nil) as this is not a fatal error (the unmarshaler will skip).
func elementAt(t target, idx int) target {
f := t.get()
switch f.Kind() {
case reflect.Slice:
//nolint:godox
// TODO: use the idx function argument and avoid alloc if possible.
idx := f.Len()
t.set(reflect.Append(f, reflect.New(f.Type().Elem()).Elem()))
return valueTarget(t.get().Index(idx))
case reflect.Array:
if idx >= f.Len() {
return nil
}
return valueTarget(f.Index(idx))
case reflect.Interface:
// This function is called after ensureValueIndexable, so it's
// guaranteed that f contains an initialized slice.
ifaceElem := f.Elem()
idx := ifaceElem.Len()
newElem := reflect.New(ifaceElem.Type().Elem()).Elem()
newSlice := reflect.Append(ifaceElem, newElem)
t.set(newSlice)
return valueTarget(t.get().Elem().Index(idx))
default:
// Why ensureValueIndexable let it go through?
panic(fmt.Errorf("elementAt received unhandled value type: %s", f.Kind()))
}
}
func (d *decoder) scopeTableTarget(shouldAppend bool, t target, name string) (target, bool, error) {
x := t.get()
switch x.Kind() {
// Kinds that need to recurse
case reflect.Interface:
t := scopeInterface(shouldAppend, t)
return d.scopeTableTarget(shouldAppend, t, name)
case reflect.Ptr:
t := scopePtr(t)
return d.scopeTableTarget(shouldAppend, t, name)
case reflect.Slice:
t := scopeSlice(shouldAppend, t)
shouldAppend = false
return d.scopeTableTarget(shouldAppend, t, name)
case reflect.Array:
t, err := d.scopeArray(shouldAppend, t)
if err != nil {
return t, false, err
}
shouldAppend = false
return d.scopeTableTarget(shouldAppend, t, name)
// Terminal kinds
case reflect.Struct:
return scopeStruct(x, name)
case reflect.Map:
if x.IsNil() {
t.set(reflect.MakeMap(x.Type()))
x = t.get()
}
return scopeMap(x, name)
default:
panic(fmt.Sprintf("can't scope on a %s", x.Kind()))
}
}
func scopeInterface(shouldAppend bool, t target) target {
initInterface(shouldAppend, t)
return interfaceTarget{t}
}
func scopePtr(t target) target {
initPtr(t)
return valueTarget(t.get().Elem())
}
func initPtr(t target) {
x := t.get()
if !x.IsNil() {
return
}
t.set(reflect.New(x.Type().Elem()))
}
// initInterface makes sure that the interface pointed at by the target is not
// nil.
// Returns the target to the initialized value of the target.
func initInterface(shouldAppend bool, t target) {
x := t.get()
if x.Kind() != reflect.Interface {
panic("this should only be called on interfaces")
}
if !x.IsNil() && (x.Elem().Type() == sliceInterfaceType || x.Elem().Type() == mapStringInterfaceType) {
return
}
var newElement reflect.Value
if shouldAppend {
newElement = reflect.MakeSlice(sliceInterfaceType, 0, 0)
} else {
newElement = reflect.MakeMap(mapStringInterfaceType)
}
t.set(newElement)
}
func scopeSlice(shouldAppend bool, t target) target {
v := t.get()
if shouldAppend {
newElem := reflect.New(v.Type().Elem())
newSlice := reflect.Append(v, newElem.Elem())
t.set(newSlice)
v = t.get()
}
return valueTarget(v.Index(v.Len() - 1))
}
func (d *decoder) scopeArray(shouldAppend bool, t target) (target, error) {
v := t.get()
idx := d.arrayIndex(shouldAppend, v)
if idx >= v.Len() {
return nil, fmt.Errorf("toml: impossible to insert element beyond array's size: %d", v.Len())
}
return valueTarget(v.Index(idx)), nil
}
func scopeMap(v reflect.Value, name string) (target, bool, error) {
k := reflect.ValueOf(name)
keyType := v.Type().Key()
if !k.Type().AssignableTo(keyType) {
if !k.Type().ConvertibleTo(keyType) {
return nil, false, fmt.Errorf("toml: cannot convert map key of type %s to expected type %s", k.Type(), keyType)
}
k = k.Convert(keyType)
}
if !v.MapIndex(k).IsValid() {
newElem := reflect.New(v.Type().Elem())
v.SetMapIndex(k, newElem.Elem())
}
return mapTarget{
v: v,
k: k,
}, true, nil
}
type fieldPathsMap = map[string][]int
type fieldPathsCache struct {
m map[reflect.Type]fieldPathsMap
l sync.RWMutex
}
func (c *fieldPathsCache) get(t reflect.Type) (fieldPathsMap, bool) {
c.l.RLock()
paths, ok := c.m[t]
c.l.RUnlock()
return paths, ok
}
func (c *fieldPathsCache) set(t reflect.Type, m fieldPathsMap) {
c.l.Lock()
c.m[t] = m
c.l.Unlock()
}
var globalFieldPathsCache = fieldPathsCache{
m: map[reflect.Type]fieldPathsMap{},
l: sync.RWMutex{},
}
func scopeStruct(v reflect.Value, name string) (target, bool, error) {
//nolint:godox
// TODO: cache this, and reduce allocations
fieldPaths, ok := globalFieldPathsCache.get(v.Type())
if !ok {
fieldPaths = map[string][]int{}
path := make([]int, 0, 16)
var walk func(reflect.Value)
walk = func(v reflect.Value) {
t := v.Type()
for i := 0; i < t.NumField(); i++ {
l := len(path)
path = append(path, i)
f := t.Field(i)
if f.Anonymous {
walk(v.Field(i))
} else if f.PkgPath == "" {
// only consider exported fields
fieldName, ok := f.Tag.Lookup("toml")
if !ok {
fieldName = f.Name
}
pathCopy := make([]int, len(path))
copy(pathCopy, path)
fieldPaths[fieldName] = pathCopy
// extra copy for the case-insensitive match
fieldPaths[strings.ToLower(fieldName)] = pathCopy
}
path = path[:l]
}
}
walk(v)
globalFieldPathsCache.set(v.Type(), fieldPaths)
}
path, ok := fieldPaths[name]
if !ok {
path, ok = fieldPaths[strings.ToLower(name)]
}
if !ok {
return nil, false, nil
}
return valueTarget(v.FieldByIndex(path)), true, nil
}
-207
View File
@@ -1,207 +0,0 @@
package toml
import (
"reflect"
"testing"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
func TestStructTarget_Ensure(t *testing.T) {
t.Parallel()
examples := []struct {
desc string
input reflect.Value
name string
test func(v reflect.Value, err error)
}{
{
desc: "handle a nil slice of string",
input: reflect.ValueOf(&struct{ A []string }{}).Elem(),
name: "A",
test: func(v reflect.Value, err error) {
assert.NoError(t, err)
assert.False(t, v.IsNil())
},
},
{
desc: "handle an existing slice of string",
input: reflect.ValueOf(&struct{ A []string }{A: []string{"foo"}}).Elem(),
name: "A",
test: func(v reflect.Value, err error) {
assert.NoError(t, err)
require.False(t, v.IsNil())
s, ok := v.Interface().([]string)
if !ok {
t.Errorf("interface %v should be castable into []string", s)
return
}
assert.Equal(t, []string{"foo"}, s)
},
},
}
for _, e := range examples {
e := e
t.Run(e.desc, func(t *testing.T) {
t.Parallel()
d := decoder{}
target, _, err := d.scopeTableTarget(false, valueTarget(e.input), e.name)
require.NoError(t, err)
err = ensureValueIndexable(target)
v := target.get()
e.test(v, err)
})
}
}
func TestStructTarget_SetString(t *testing.T) {
t.Parallel()
str := "value"
examples := []struct {
desc string
input reflect.Value
name string
test func(v reflect.Value, err error)
}{
{
desc: "sets a string",
input: reflect.ValueOf(&struct{ A string }{}).Elem(),
name: "A",
test: func(v reflect.Value, err error) {
assert.NoError(t, err)
assert.Equal(t, str, v.String())
},
},
{
desc: "fails on a float",
input: reflect.ValueOf(&struct{ A float64 }{}).Elem(),
name: "A",
test: func(v reflect.Value, err error) {
assert.Error(t, err)
},
},
{
desc: "fails on a slice",
input: reflect.ValueOf(&struct{ A []string }{}).Elem(),
name: "A",
test: func(v reflect.Value, err error) {
assert.Error(t, err)
},
},
}
for _, e := range examples {
e := e
t.Run(e.desc, func(t *testing.T) {
t.Parallel()
d := decoder{}
target, _, err := d.scopeTableTarget(false, valueTarget(e.input), e.name)
require.NoError(t, err)
err = setString(target, str)
v := target.get()
e.test(v, err)
})
}
}
func TestPushNew(t *testing.T) {
t.Parallel()
t.Run("slice of strings", func(t *testing.T) {
t.Parallel()
type Doc struct {
A []string
}
d := Doc{}
dec := decoder{}
x, _, err := dec.scopeTableTarget(false, valueTarget(reflect.ValueOf(&d).Elem()), "A")
require.NoError(t, err)
n := elementAt(x, 0)
n.setString("hello")
require.Equal(t, []string{"hello"}, d.A)
n = elementAt(x, 1)
n.setString("world")
require.Equal(t, []string{"hello", "world"}, d.A)
})
t.Run("slice of interfaces", func(t *testing.T) {
t.Parallel()
type Doc struct {
A []interface{}
}
d := Doc{}
dec := decoder{}
x, _, err := dec.scopeTableTarget(false, valueTarget(reflect.ValueOf(&d).Elem()), "A")
require.NoError(t, err)
n := elementAt(x, 0)
require.NoError(t, setString(n, "hello"))
require.Equal(t, []interface{}{"hello"}, d.A)
n = elementAt(x, 1)
require.NoError(t, setString(n, "world"))
require.Equal(t, []interface{}{"hello", "world"}, d.A)
})
}
func TestScope_Struct(t *testing.T) {
t.Parallel()
examples := []struct {
desc string
input reflect.Value
name string
err bool
found bool
idx []int
}{
{
desc: "simple field",
input: reflect.ValueOf(&struct{ A string }{}).Elem(),
name: "A",
idx: []int{0},
found: true,
},
{
desc: "fails not-exported field",
input: reflect.ValueOf(&struct{ a string }{}).Elem(),
name: "a",
err: false,
found: false,
},
}
for _, e := range examples {
e := e
t.Run(e.desc, func(t *testing.T) {
t.Parallel()
dec := decoder{}
x, found, err := dec.scopeTableTarget(false, valueTarget(e.input), e.name)
assert.Equal(t, e.found, found)
if e.err {
assert.Error(t, err)
}
if found {
x2, ok := x.(valueTarget)
require.True(t, ok)
x2.get()
}
})
}
}
+74
View File
@@ -0,0 +1,74 @@
package testsuite
import (
"fmt"
"math"
"time"
"github.com/pelletier/go-toml/v2"
)
// addTag adds JSON tags to a data structure as expected by toml-test.
func addTag(key string, tomlData interface{}) interface{} {
// Switch on the data type.
switch orig := tomlData.(type) {
default:
//return map[string]interface{}{}
panic(fmt.Sprintf("Unknown type: %T", tomlData))
// A table: we don't need to add any tags, just recurse for every table
// entry.
case map[string]interface{}:
typed := make(map[string]interface{}, len(orig))
for k, v := range orig {
typed[k] = addTag(k, v)
}
return typed
// An array: we don't need to add any tags, just recurse for every table
// entry.
case []map[string]interface{}:
typed := make([]map[string]interface{}, len(orig))
for i, v := range orig {
typed[i] = addTag("", v).(map[string]interface{})
}
return typed
case []interface{}:
typed := make([]interface{}, len(orig))
for i, v := range orig {
typed[i] = addTag("", v)
}
return typed
// Datetime: tag as datetime.
case toml.LocalTime:
return tag("time-local", orig.String())
case toml.LocalDate:
return tag("date-local", orig.String())
case toml.LocalDateTime:
return tag("datetime-local", orig.String())
case time.Time:
return tag("datetime", orig.Format("2006-01-02T15:04:05.999999999Z07:00"))
// Tag primitive values: bool, string, int, and float64.
case bool:
return tag("bool", fmt.Sprintf("%v", orig))
case string:
return tag("string", orig)
case int64:
return tag("integer", fmt.Sprintf("%d", orig))
case float64:
// Special case for nan since NaN == NaN is false.
if math.IsNaN(orig) {
return tag("float", "nan")
}
return tag("float", fmt.Sprintf("%v", orig))
}
}
func tag(typeName string, data interface{}) map[string]interface{} {
return map[string]interface{}{
"type": typeName,
"value": data,
}
}
+244
View File
@@ -0,0 +1,244 @@
package testsuite
import (
"fmt"
"strconv"
"strings"
"testing"
"time"
)
func CmpJSON(t *testing.T, key string, want, have interface{}) {
switch w := want.(type) {
case map[string]interface{}:
cmpJSONMaps(t, key, w, have)
case []interface{}:
cmpJSONArrays(t, key, w, have)
default:
t.Errorf(
"Key '%s' in expected output should be a map or a list of maps, but it's a %T",
key, want)
}
}
func cmpJSONMaps(t *testing.T, key string, want map[string]interface{}, have interface{}) {
haveMap, ok := have.(map[string]interface{})
if !ok {
mismatch(t, key, "table", want, haveMap)
return
}
// Check to make sure both or neither are values.
if isValue(want) && !isValue(haveMap) {
t.Fatalf("Key '%s' is supposed to be a value, but the parser reports it as a table", key)
}
if !isValue(want) && isValue(haveMap) {
t.Fatalf("Key '%s' is supposed to be a table, but the parser reports it as a value", key)
}
if isValue(want) && isValue(haveMap) {
cmpJSONValues(t, key, want, haveMap)
return
}
// Check that the keys of each map are equivalent.
for k := range want {
if _, ok := haveMap[k]; !ok {
bunk := kjoin(key, k)
t.Fatalf("Could not find key '%s' in parser output.", bunk)
}
}
for k := range haveMap {
if _, ok := want[k]; !ok {
bunk := kjoin(key, k)
t.Fatalf("Could not find key '%s' in expected output.", bunk)
}
}
// Okay, now make sure that each value is equivalent.
for k := range want {
CmpJSON(t, kjoin(key, k), want[k], haveMap[k])
}
}
func cmpJSONArrays(t *testing.T, key string, want, have interface{}) {
wantSlice, ok := want.([]interface{})
if !ok {
panic(fmt.Sprintf("'value' should be a JSON array when 'type=array', but it is a %T", want))
}
haveSlice, ok := have.([]interface{})
if !ok {
t.Fatalf("Malformed output from your encoder: 'value' is not a JSON array: %T", have)
}
if len(wantSlice) != len(haveSlice) {
t.Fatalf("Array lengths differ for key '%s':\n"+
" Expected: %d\n"+
" Your encoder: %d",
key, len(wantSlice), len(haveSlice))
}
for i := 0; i < len(wantSlice); i++ {
CmpJSON(t, key, wantSlice[i], haveSlice[i])
}
}
func cmpJSONValues(t *testing.T, key string, want, have map[string]interface{}) {
wantType, ok := want["type"].(string)
if !ok {
panic(fmt.Sprintf("'type' should be a string, but it is a %T", want["type"]))
}
haveType, ok := have["type"].(string)
if !ok {
t.Fatalf("Malformed output from your encoder: 'type' is not a string: %T", have["type"])
}
if wantType != haveType {
valMismatch(t, key, wantType, haveType, want, have)
}
// If this is an array, then we've got to do some work to check equality.
if wantType == "array" {
cmpJSONArrays(t, key, want, have)
return
}
// Atomic values are always strings
wantVal, ok := want["value"].(string)
if !ok {
panic(fmt.Sprintf("'value' %v should be a string, but it is a %[1]T", want["value"]))
}
haveVal, ok := have["value"].(string)
if !ok {
panic(fmt.Sprintf("Malformed output from your encoder: %T is not a string", have["value"]))
}
// Excepting floats and datetimes, other values can be compared as strings.
switch wantType {
case "float":
cmpFloats(t, key, wantVal, haveVal)
case "datetime", "datetime-local", "date-local", "time-local":
cmpAsDatetimes(t, key, wantType, wantVal, haveVal)
default:
cmpAsStrings(t, key, wantVal, haveVal)
}
}
func cmpAsStrings(t *testing.T, key string, want, have string) {
if want != have {
t.Fatalf("Values for key '%s' don't match:\n"+
" Expected: %s\n"+
" Your encoder: %s",
key, want, have)
}
}
func cmpFloats(t *testing.T, key string, want, have string) {
// Special case for NaN, since NaN != NaN.
if strings.HasSuffix(want, "nan") || strings.HasSuffix(have, "nan") {
if want != have {
t.Fatalf("Values for key '%s' don't match:\n"+
" Expected: %v\n"+
" Your encoder: %v",
key, want, have)
}
return
}
wantF, err := strconv.ParseFloat(want, 64)
if err != nil {
panic(fmt.Sprintf("Could not read '%s' as a float value for key '%s'", want, key))
}
haveF, err := strconv.ParseFloat(have, 64)
if err != nil {
panic(fmt.Sprintf("Malformed output from your encoder: key '%s' is not a float: '%s'", key, have))
}
if wantF != haveF {
t.Fatalf("Values for key '%s' don't match:\n"+
" Expected: %v\n"+
" Your encoder: %v",
key, wantF, haveF)
}
}
var datetimeRepl = strings.NewReplacer(
" ", "T",
"t", "T",
"z", "Z")
var layouts = map[string]string{
"datetime": time.RFC3339Nano,
"datetime-local": "2006-01-02T15:04:05.999999999",
"date-local": "2006-01-02",
"time-local": "15:04:05",
}
func cmpAsDatetimes(t *testing.T, key string, kind, want, have string) {
layout, ok := layouts[kind]
if !ok {
panic("should never happen")
}
wantT, err := time.Parse(layout, datetimeRepl.Replace(want))
if err != nil {
panic(fmt.Sprintf("Could not read '%s' as a datetime value for key '%s'", want, key))
}
haveT, err := time.Parse(layout, datetimeRepl.Replace(want))
if err != nil {
t.Fatalf("Malformed output from your encoder: key '%s' is not a datetime: '%s'", key, have)
return
}
if !wantT.Equal(haveT) {
t.Fatalf("Values for key '%s' don't match:\n"+
" Expected: %v\n"+
" Your encoder: %v",
key, wantT, haveT)
}
}
func cmpAsDatetimesLocal(t *testing.T, key string, want, have string) {
if datetimeRepl.Replace(want) != datetimeRepl.Replace(have) {
t.Fatalf("Values for key '%s' don't match:\n"+
" Expected: %v\n"+
" Your encoder: %v",
key, want, have)
}
}
func kjoin(old, key string) string {
if len(old) == 0 {
return key
}
return old + "." + key
}
func isValue(m map[string]interface{}) bool {
if len(m) != 2 {
return false
}
if _, ok := m["type"]; !ok {
return false
}
if _, ok := m["value"]; !ok {
return false
}
return true
}
func mismatch(t *testing.T, key string, wantType string, want, have interface{}) {
t.Fatalf("Key '%s' is not an %s but %[4]T:\n"+
" Expected: %#[3]v\n"+
" Your encoder: %#[4]v",
key, wantType, want, have)
}
func valMismatch(t *testing.T, key string, wantType, haveType string, want, have interface{}) {
t.Fatalf("Key '%s' is not an %s but %s:\n"+
" Expected: %#[3]v\n"+
" Your encoder: %#[4]v",
key, wantType, want, have)
}
+69
View File
@@ -0,0 +1,69 @@
package testsuite
import (
"bytes"
"encoding/json"
"fmt"
"github.com/pelletier/go-toml/v2"
)
type parser struct{}
func (p parser) Decode(input string) (output string, outputIsError bool, retErr error) {
defer func() {
if r := recover(); r != nil {
switch rr := r.(type) {
case error:
retErr = rr
default:
retErr = fmt.Errorf("%s", rr)
}
}
}()
var v interface{}
if err := toml.Unmarshal([]byte(input), &v); err != nil {
return err.Error(), true, nil
}
j, err := json.MarshalIndent(addTag("", v), "", " ")
if err != nil {
return "", false, retErr
}
return string(j), false, retErr
}
func (p parser) Encode(input string) (output string, outputIsError bool, retErr error) {
defer func() {
if r := recover(); r != nil {
switch rr := r.(type) {
case error:
retErr = rr
default:
retErr = fmt.Errorf("%s", rr)
}
}
}()
var tmp interface{}
err := json.Unmarshal([]byte(input), &tmp)
if err != nil {
return "", false, err
}
rm, err := rmTag(tmp)
if err != nil {
return err.Error(), true, retErr
}
buf := new(bytes.Buffer)
err = toml.NewEncoder(buf).Encode(rm)
if err != nil {
return err.Error(), true, retErr
}
return buf.String(), false, retErr
}
+110
View File
@@ -0,0 +1,110 @@
package testsuite
import (
"fmt"
"strconv"
"time"
)
// Remove JSON tags to a data structure as returned by toml-test.
func rmTag(typedJson interface{}) (interface{}, error) {
// Check if key is in the table m.
in := func(key string, m map[string]interface{}) bool {
_, ok := m[key]
return ok
}
// Switch on the data type.
switch v := typedJson.(type) {
// Object: this can either be a TOML table or a primitive with tags.
case map[string]interface{}:
// This value represents a primitive: remove the tags and return just
// the primitive value.
if len(v) == 2 && in("type", v) && in("value", v) {
ut, err := untag(v)
if err != nil {
return ut, fmt.Errorf("tag.Remove: %w", err)
}
return ut, nil
}
// Table: remove tags on all children.
m := make(map[string]interface{}, len(v))
for k, v2 := range v {
var err error
m[k], err = rmTag(v2)
if err != nil {
return nil, err
}
}
return m, nil
// Array: remove tags from all itenm.
case []interface{}:
a := make([]interface{}, len(v))
for i := range v {
var err error
a[i], err = rmTag(v[i])
if err != nil {
return nil, err
}
}
return a, nil
}
// The top level must be an object or array.
return nil, fmt.Errorf("unrecognized JSON format '%T'", typedJson)
}
// Return a primitive: read the "type" and convert the "value" to that.
func untag(typed map[string]interface{}) (interface{}, error) {
t := typed["type"].(string)
v := typed["value"].(string)
switch t {
case "string":
return v, nil
case "integer":
n, err := strconv.ParseInt(v, 10, 64)
if err != nil {
return nil, fmt.Errorf("untag: %w", err)
}
return n, nil
case "float":
f, err := strconv.ParseFloat(v, 64)
if err != nil {
return nil, fmt.Errorf("untag: %w", err)
}
return f, nil
case "datetime":
return parseTime(v, "2006-01-02T15:04:05.999999999Z07:00", false)
case "datetime-local":
return parseTime(v, "2006-01-02T15:04:05.999999999", true)
case "date-local":
return parseTime(v, "2006-01-02", true)
case "time-local":
return parseTime(v, "15:04:05.999999999", true)
case "bool":
switch v {
case "true":
return true, nil
case "false":
return false, nil
}
return nil, fmt.Errorf("untag: could not parse %q as a boolean", v)
}
return nil, fmt.Errorf("untag: unrecognized tag type %q", t)
}
func parseTime(v, format string, local bool) (t time.Time, err error) {
if local {
t, err = time.ParseInLocation(format, v, time.Local)
} else {
t, err = time.Parse(format, v)
}
if err != nil {
return time.Time{}, fmt.Errorf("Could not parse %q as a datetime: %w", v, err)
}
return t, nil
}
+50
View File
@@ -0,0 +1,50 @@
// Package testsuite provides helper functions for interoperating with the
// language-agnostic TOML test suite at github.com/BurntSushi/toml-test.
package testsuite
import (
"encoding/json"
"fmt"
"os"
"github.com/pelletier/go-toml/v2"
)
// Marshal is a helpfer function for calling toml.Marshal
//
// Only needed to avoid package import loops.
func Marshal(v interface{}) ([]byte, error) {
return toml.Marshal(v)
}
// Unmarshal is a helper function for calling toml.Unmarshal.
//
// Only needed to avoid package import loops.
func Unmarshal(data []byte, v interface{}) error {
return toml.Unmarshal(data, v)
}
// ValueToTaggedJSON takes a data structure and returns the tagged JSON
// representation.
func ValueToTaggedJSON(doc interface{}) ([]byte, error) {
return json.MarshalIndent(addTag("", doc), "", " ")
}
// DecodeStdin is a helper function for the toml-test binary interface. TOML input
// is read from STDIN and a resulting tagged JSON representation is written to
// STDOUT.
func DecodeStdin() error {
var decoded map[string]interface{}
if err := toml.NewDecoder(os.Stdin).Decode(&decoded); err != nil {
return fmt.Errorf("Error decoding TOML: %s", err)
}
j := json.NewEncoder(os.Stdout)
j.SetIndent("", " ")
if err := j.Encode(addTag("", decoded)); err != nil {
return fmt.Errorf("Error encoding JSON: %s", err)
}
return nil
}
+23 -117
View File
@@ -1,14 +1,14 @@
//go:generate go run ./cmd/tomltestgen/main.go -o toml_testgen_test.go
// This is a support file for toml_testgen_test.go
package toml_test
import (
"encoding/json"
"fmt"
"strconv"
"testing"
"time"
"github.com/pelletier/go-toml/v2"
"github.com/pelletier/go-toml/v2/testsuite"
"github.com/stretchr/testify/require"
)
@@ -17,10 +17,14 @@ func testgenInvalid(t *testing.T, input string) {
t.Logf("Input TOML:\n%s", input)
doc := map[string]interface{}{}
err := toml.Unmarshal([]byte(input), &doc)
err := testsuite.Unmarshal([]byte(input), &doc)
if err == nil {
t.Log(json.Marshal(doc))
out, err := json.Marshal(doc)
if err != nil {
panic("could not marshal map to json")
}
t.Log("JSON output from unmarshal:", string(out))
t.Fatalf("test did not fail")
}
}
@@ -29,124 +33,26 @@ func testgenValid(t *testing.T, input string, jsonRef string) {
t.Helper()
t.Logf("Input TOML:\n%s", input)
doc := map[string]interface{}{}
// TODO: change this to interface{}
var doc map[string]interface{}
err := toml.Unmarshal([]byte(input), &doc)
err := testsuite.Unmarshal([]byte(input), &doc)
if err != nil {
if de, ok := err.(*toml.DecodeError); ok {
t.Logf("%s\n%s", err, de)
}
t.Fatalf("failed parsing toml: %s", err)
}
refDoc := testgenBuildRefDoc(jsonRef)
require.Equal(t, refDoc, doc)
out, err := toml.Marshal(doc)
j, err := testsuite.ValueToTaggedJSON(doc)
require.NoError(t, err)
doc2 := map[string]interface{}{}
err = toml.Unmarshal(out, &doc2)
var ref interface{}
err = json.Unmarshal([]byte(jsonRef), &ref)
require.NoError(t, err)
require.Equal(t, refDoc, doc2)
}
func testgenBuildRefDoc(jsonRef string) map[string]interface{} {
descTree := map[string]interface{}{}
err := json.Unmarshal([]byte(jsonRef), &descTree)
if err != nil {
panic(fmt.Sprintf("reference doc should be valid JSON: %s", err))
}
doc := testGenTranslateDesc(descTree)
if doc == nil {
return map[string]interface{}{}
}
return doc.(map[string]interface{})
}
//nolint:funlen,gocognit,cyclop
func testGenTranslateDesc(input interface{}) interface{} {
a, ok := input.([]interface{})
if ok {
xs := make([]interface{}, len(a))
for i, v := range a {
xs[i] = testGenTranslateDesc(v)
}
return xs
}
d, ok := input.(map[string]interface{})
if !ok {
panic(fmt.Sprintf("input should be valid map[string]: %v", input))
}
var (
dtype string
dvalue interface{}
)
//nolint:nestif
if len(d) == 2 {
dtypeiface, ok := d["type"]
if ok {
dvalue, ok = d["value"]
if ok {
dtype = dtypeiface.(string)
switch dtype {
case "string":
return dvalue.(string)
case "float":
v, err := strconv.ParseFloat(dvalue.(string), 64)
if err != nil {
panic(fmt.Sprintf("invalid float '%s': %s", dvalue, err))
}
return v
case "integer":
v, err := strconv.ParseInt(dvalue.(string), 10, 64)
if err != nil {
panic(fmt.Sprintf("invalid int '%s': %s", dvalue, err))
}
return v
case "bool":
return dvalue.(string) == "true"
case "datetime":
dt, err := time.Parse("2006-01-02T15:04:05Z", dvalue.(string))
if err != nil {
panic(fmt.Sprintf("invalid datetime '%s': %s", dvalue, err))
}
return dt
case "array":
if dvalue == nil {
return nil
}
a := dvalue.([]interface{})
xs := make([]interface{}, len(a))
for i, v := range a {
xs[i] = testGenTranslateDesc(v)
}
return xs
}
panic(fmt.Sprintf("unknown type: %s", dtype))
}
}
}
dest := map[string]interface{}{}
for k, v := range d {
dest[k] = testGenTranslateDesc(v)
}
return dest
var actual interface{}
err = json.Unmarshal([]byte(j), &actual)
require.NoError(t, err)
testsuite.CmpJSON(t, "", ref, actual)
}
+1303 -846
View File
File diff suppressed because it is too large Load Diff
+14
View File
@@ -0,0 +1,14 @@
package toml
import (
"encoding"
"reflect"
"time"
)
var timeType = reflect.TypeOf(time.Time{})
var textMarshalerType = reflect.TypeOf(new(encoding.TextMarshaler)).Elem()
var textUnmarshalerType = reflect.TypeOf(new(encoding.TextUnmarshaler)).Elem()
var mapStringInterfaceType = reflect.TypeOf(map[string]interface{}{})
var sliceInterfaceType = reflect.TypeOf([]interface{}{})
var stringType = reflect.TypeOf("")
+979 -332
View File
File diff suppressed because it is too large Load Diff
+1274 -260
View File
File diff suppressed because it is too large Load Diff
+240
View File
@@ -0,0 +1,240 @@
package toml
import (
"unicode/utf8"
)
type utf8Err struct {
Index int
Size int
}
func (u utf8Err) Zero() bool {
return u.Size == 0
}
// Verified that a given string is only made of valid UTF-8 characters allowed
// by the TOML spec:
//
// Any Unicode character may be used except those that must be escaped:
// quotation mark, backslash, and the control characters other than tab (U+0000
// to U+0008, U+000A to U+001F, U+007F).
//
// It is a copy of the Go 1.17 utf8.Valid implementation, tweaked to exit early
// when a character is not allowed.
//
// The returned utf8Err is Zero() if the string is valid, or contains the byte
// index and size of the invalid character.
//
// quotation mark => already checked
// backslash => already checked
// 0-0x8 => invalid
// 0x9 => tab, ok
// 0xA - 0x1F => invalid
// 0x7F => invalid
func utf8TomlValidAlreadyEscaped(p []byte) (err utf8Err) {
// Fast path. Check for and skip 8 bytes of ASCII characters per iteration.
offset := 0
for len(p) >= 8 {
// Combining two 32 bit loads allows the same code to be used
// for 32 and 64 bit platforms.
// The compiler can generate a 32bit load for first32 and second32
// on many platforms. See test/codegen/memcombine.go.
first32 := uint32(p[0]) | uint32(p[1])<<8 | uint32(p[2])<<16 | uint32(p[3])<<24
second32 := uint32(p[4]) | uint32(p[5])<<8 | uint32(p[6])<<16 | uint32(p[7])<<24
if (first32|second32)&0x80808080 != 0 {
// Found a non ASCII byte (>= RuneSelf).
break
}
for i, b := range p[:8] {
if invalidAscii(b) {
err.Index = offset + i
err.Size = 1
return
}
}
p = p[8:]
offset += 8
}
n := len(p)
for i := 0; i < n; {
pi := p[i]
if pi < utf8.RuneSelf {
if invalidAscii(pi) {
err.Index = offset + i
err.Size = 1
return
}
i++
continue
}
x := first[pi]
if x == xx {
// Illegal starter byte.
err.Index = offset + i
err.Size = 1
return
}
size := int(x & 7)
if i+size > n {
// Short or invalid.
err.Index = offset + i
err.Size = n - i
return
}
accept := acceptRanges[x>>4]
if c := p[i+1]; c < accept.lo || accept.hi < c {
err.Index = offset + i
err.Size = 2
return
} else if size == 2 {
} else if c := p[i+2]; c < locb || hicb < c {
err.Index = offset + i
err.Size = 3
return
} else if size == 3 {
} else if c := p[i+3]; c < locb || hicb < c {
err.Index = offset + i
err.Size = 4
return
}
i += size
}
return
}
// Return the size of the next rune if valid, 0 otherwise.
func utf8ValidNext(p []byte) int {
c := p[0]
if c < utf8.RuneSelf {
if invalidAscii(c) {
return 0
}
return 1
}
x := first[c]
if x == xx {
// Illegal starter byte.
return 0
}
size := int(x & 7)
if size > len(p) {
// Short or invalid.
return 0
}
accept := acceptRanges[x>>4]
if c := p[1]; c < accept.lo || accept.hi < c {
return 0
} else if size == 2 {
} else if c := p[2]; c < locb || hicb < c {
return 0
} else if size == 3 {
} else if c := p[3]; c < locb || hicb < c {
return 0
}
return size
}
var invalidAsciiTable = [256]bool{
0x00: true,
0x01: true,
0x02: true,
0x03: true,
0x04: true,
0x05: true,
0x06: true,
0x07: true,
0x08: true,
// 0x09 TAB
// 0x0A LF
0x0B: true,
0x0C: true,
// 0x0D CR
0x0E: true,
0x0F: true,
0x10: true,
0x11: true,
0x12: true,
0x13: true,
0x14: true,
0x15: true,
0x16: true,
0x17: true,
0x18: true,
0x19: true,
0x1A: true,
0x1B: true,
0x1C: true,
0x1D: true,
0x1E: true,
0x1F: true,
// 0x20 - 0x7E Printable ASCII characters
0x7F: true,
}
func invalidAscii(b byte) bool {
return invalidAsciiTable[b]
}
// acceptRange gives the range of valid values for the second byte in a UTF-8
// sequence.
type acceptRange struct {
lo uint8 // lowest value for second byte.
hi uint8 // highest value for second byte.
}
// acceptRanges has size 16 to avoid bounds checks in the code that uses it.
var acceptRanges = [16]acceptRange{
0: {locb, hicb},
1: {0xA0, hicb},
2: {locb, 0x9F},
3: {0x90, hicb},
4: {locb, 0x8F},
}
// first is information about the first byte in a UTF-8 sequence.
var first = [256]uint8{
// 1 2 3 4 5 6 7 8 9 A B C D E F
as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x00-0x0F
as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x10-0x1F
as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x20-0x2F
as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x30-0x3F
as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x40-0x4F
as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x50-0x5F
as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x60-0x6F
as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x70-0x7F
// 1 2 3 4 5 6 7 8 9 A B C D E F
xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, // 0x80-0x8F
xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, // 0x90-0x9F
xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, // 0xA0-0xAF
xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, // 0xB0-0xBF
xx, xx, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, // 0xC0-0xCF
s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, // 0xD0-0xDF
s2, s3, s3, s3, s3, s3, s3, s3, s3, s3, s3, s3, s3, s4, s3, s3, // 0xE0-0xEF
s5, s6, s6, s6, s7, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, // 0xF0-0xFF
}
const (
// The default lowest and highest continuation byte.
locb = 0b10000000
hicb = 0b10111111
// These names of these constants are chosen to give nice alignment in the
// table below. The first nibble is an index into acceptRanges or F for
// special one-byte cases. The second nibble is the Rune length or the
// Status for the special one-byte case.
xx = 0xF1 // invalid: size 1
as = 0xF0 // ASCII: size 1
s1 = 0x02 // accept 0, size 2
s2 = 0x13 // accept 1, size 3
s3 = 0x03 // accept 0, size 3
s4 = 0x23 // accept 2, size 3
s5 = 0x34 // accept 3, size 4
s6 = 0x04 // accept 0, size 4
s7 = 0x44 // accept 4, size 4
)