Commit Graph

  • c3e1a2bff4 Increment version number to 0.8.15. Henri Sivonen 2019-01-29 21:32:17 +02:00
  • e371e190c6 Document simd-accel more accurately. Henri Sivonen 2019-01-29 21:31:34 +02:00
  • f7921d677b Fix syntax error in non-SSE2 simd_pack(). Henri Sivonen 2019-01-21 13:03:42 +02:00
  • 21f0d93f2a Increment packed_simd requirement. Henri Sivonen 2019-01-21 11:54:19 +02:00
  • 3acfb1bdea Merge trunk changes into packed_simd Henri Sivonen 2019-01-21 11:45:02 +02:00
  • 77cc8339f2 Increment version number to 0.8.14. Henri Sivonen 2019-01-07 16:52:58 +02:00
  • 1e503be658 Merge trunk changes into packed_simd Henri Sivonen 2018-12-20 14:16:12 +02:00
  • c35e1ad71b Remove clippy:: prefix from lints. Henri Sivonen 2018-12-20 14:12:37 +02:00
  • 58c2c8a517 Update rustfmt. Henri Sivonen 2018-12-20 14:07:00 +02:00
  • 62f00d0668 Make the fuzzer aware of times 3 plus 1 changing to just times 3. Henri Sivonen 2018-12-19 14:33:13 +02:00
  • 698bff0b5a Tweak comment about speculation. Henri Sivonen 2018-12-19 14:19:37 +02:00
  • 22d0dfeeaa Add tests for precise UTF-16 to UTF-8 buffer filling. Henri Sivonen 2018-12-19 13:47:31 +02:00
  • e43d797e46 Change UTF-16 to UTF-8 buffer requirements from times 3 plus 1 to just times 3. Henri Sivonen 2018-12-19 12:03:27 +02:00
  • f886bbd028 Note the buffer filling change in UTF-16 to UTF-8 encode in README. Henri Sivonen 2018-12-19 11:52:32 +02:00
  • 1880eb4d8e Make UTF-16 to UTF-8 encode fill output buffer exactly using inline(never) code for the tail. Henri Sivonen 2018-12-19 11:19:20 +02:00
  • 1773c0f0a6 Make the core of UTF-16 to UTF-8 encoded inline(never). Henri Sivonen 2018-12-19 10:33:49 +02:00
  • 05ae91866e Update rustfmt. Henri Sivonen 2018-12-19 10:31:36 +02:00
  • 743d831aba Link to https://hsivonen.fi/encoding_rs/ from the API docs. Henri Sivonen 2018-12-18 10:17:12 +02:00
  • b18316f1ab Strike over the roadmap item about Bob Steagall's DFA. Henri Sivonen 2018-12-18 10:09:54 +02:00
  • 8446da6b42 Link to relevant things from the README. Henri Sivonen 2018-12-18 10:09:25 +02:00
  • 11b0b46b3c Add code generation for the codepage crate. Henri Sivonen 2018-12-02 17:04:11 +02:00
  • 26061feded Use the standard library for creating &[u16] in fuzz_mem. Henri Sivonen 2018-11-30 11:11:02 +02:00
  • 340cbe1c5c Add missing word to README. Henri Sivonen 2018-11-28 19:33:11 +02:00
  • 473ce0beef Increment the version number to 0.8.13. Henri Sivonen 2018-11-28 16:53:04 +02:00
  • b9ed235460 Compare written with the length of the right slice when convertion UTF-8 to UTF-16. Henri Sivonen 2018-11-28 16:44:29 +02:00
  • 570a23c104 Increment version number to 0.8.12. Henri Sivonen 2018-11-17 20:15:50 +02:00
  • 4388101984 Document the clippy lint name change. Henri Sivonen 2018-11-17 20:14:54 +02:00
  • 84465f3769 remove unstable scoped lint Yusuke Sasaki 2018-11-17 21:47:47 +09:00
  • 87c9397468 Increment version number to 0.8.11. Henri Sivonen 2018-11-16 12:27:11 +02:00
  • 00403b118b Silence warnings and run cargo fmt. Henri Sivonen 2018-11-16 11:36:33 +02:00
  • b482618653 Disable use of SIMD for ASCII validation on aarch64 due to ThunderX performance. Henri Sivonen 2018-11-16 11:28:52 +02:00
  • 874d8bda47 Return correct advance when the second vector in a pair fails. Henri Sivonen 2018-11-13 14:37:42 +02:00
  • 3fca950d8f In unrolled SSE2 ASCII loop, write vectors that are not completely ASCII. Henri Sivonen 2018-11-13 13:48:03 +02:00
  • bbbb222d9e Unroll SSE2 ASCII decode. Henri Sivonen 2018-11-13 13:05:00 +02:00
  • 545feda41c Use SIMD instead of ALU to get into alignment in SSE2 ASCII validation. Henri Sivonen 2018-11-13 11:17:43 +02:00
  • b924bf38ef Unroll SSE2 ASCII validation by two. Henri Sivonen 2018-11-13 10:49:02 +02:00
  • 5ac63e47bc Avoid saying that minimum Rust version change is semver breaking. Henri Sivonen 2018-11-06 10:33:31 +02:00
  • c314d70a24 Merge changes from different computers Henri Sivonen 2018-11-05 12:58:56 +02:00
  • b8e1e5e827 Update fuzzing Cargo.toml. Henri Sivonen 2018-11-05 12:57:24 +02:00
  • dd1f5f88e8 Adjust crates.io categories: remove email and add internationalization. Henri Sivonen 2018-11-05 09:19:53 +02:00
  • d327f5eade Point to the charset crate where relevant. Henri Sivonen 2018-11-05 09:04:12 +02:00
  • 542ad1634b Tests for decode_latin1 and encode_latin1_lossy. Henri Sivonen 2018-11-04 21:46:45 +02:00
  • 47f49ffea7 Document previous changeset in release notes. Henri Sivonen 2018-11-04 21:37:50 +02:00
  • 1c42d5bc38 Add conversions between UTF-8 and Latin1 with Cows for output. Henri Sivonen 2018-11-04 21:36:17 +02:00
  • 7a8b938b8c Tweak release notes. Henri Sivonen 2018-11-02 12:35:17 +02:00
  • bbbaa4ef22 Bob Steagall claims better DFA results than achieved with Björn Höhrmann's DFA, so if investigating DFA, it's better to look at Bob Steagall's DFA than Björn Höhrmann's DFA. Henri Sivonen 2018-11-02 12:30:43 +02:00
  • 22455fdfe7 Remove traces of rayon. Henri Sivonen 2018-11-02 12:13:06 +02:00
  • 3ad7192d65 Revise release notes about UTF-8. Henri Sivonen 2018-11-02 12:12:18 +02:00
  • 0b15425487 Merge UTF-8 data arrays. Henri Sivonen 2018-11-02 12:06:02 +02:00
  • 6206d92ba6 Ensure UTF-8 data is aligned to cache lines. Henri Sivonen 2018-11-02 11:55:02 +02:00
  • 13e6941022 Clean up constants and mem. Henri Sivonen 2018-11-02 11:46:34 +02:00
  • 367b4119ba Avoid table lookup for third and fourth byte in four-byte UTF-8 sequences. Henri Sivonen 2018-11-02 11:14:02 +02:00
  • fe9d54e440 Avoid table lookup for the last trail in three-byte UTF-8 sequences. Henri Sivonen 2018-11-02 10:23:53 +02:00
  • b6c39f3c74 Stop using UTF-8 validation code forked from the standard library. Henri Sivonen 2018-11-01 17:19:28 +02:00
  • d2829e270d Use UTF-8 to UTF-16 code structure for UTF-8 validation. Henri Sivonen 2018-11-01 17:06:32 +02:00
  • 80fc74789d Tweak UTF-8 to UTF-16 branch probabilities. Henri Sivonen 2018-11-01 16:32:45 +02:00
  • fdd4d90830 Have an inner loop only for 3-byte sequences. Henri Sivonen 2018-11-01 16:25:05 +02:00
  • 35fcfa89de Avoid checking ASCII after checking next lead. Henri Sivonen 2018-11-01 16:15:01 +02:00
  • 5b0113baf2 Optimize next UTF-8 sequence length matching previous. Henri Sivonen 2018-11-01 16:07:06 +02:00
  • 85fe2f99d4 Manually inline reading of the next UTF-8 lead byte. Henri Sivonen 2018-11-01 15:57:56 +02:00
  • 43d8c6143d Add UTF-8 to UTF-16 improvement to release notes. Henri Sivonen 2018-11-01 15:45:17 +02:00
  • a48ff1b364 Make the UTF-8 to UTF-16 tail handling not stop too early. Henri Sivonen 2018-11-01 15:36:54 +02:00
  • e56e7025a4 Avoid table for two-byte UTF-8 sequences. Henri Sivonen 2018-11-01 14:53:20 +02:00
  • 601ed56fbe Read UTF-8 second byte range mask from a table. Henri Sivonen 2018-11-01 14:43:23 +02:00
  • d96d43cbbb Use if-else instead of match. Henri Sivonen 2018-11-01 12:25:37 +02:00
  • 48e0a440f6 Remove bound checks from the inner loop of UTF-8 to UTF-16 conversion. Henri Sivonen 2018-11-01 11:44:27 +02:00
  • b2c0941ce6 Add newline to end of file. Henri Sivonen 2018-10-23 20:30:29 +03:00
  • e2da3a8923 Write down a design for table-driven UTF-8 validation and decoding. Henri Sivonen 2018-10-23 20:29:50 +03:00
  • 542b914e3a Remove unsafe from copy_utf8_up_to_invalid_from. Henri Sivonen 2018-10-19 15:05:20 +03:00
  • 0f99225469 Remove unsafe from UTF-8 to UTF-8 encode. Henri Sivonen 2018-10-19 15:00:23 +03:00
  • d8ee055e8f Remove the remaining transmutes. Henri Sivonen 2018-10-19 10:11:47 +03:00
  • 8c4b3d2ee3 Merge branch 'noinlinemem' into packed_simd Henri Sivonen 2018-10-18 12:26:29 +03:00
  • 3edd3bcd52 Avoid offering functions that get SIMD code inlined into them as candidates to be inlined into other crates. Henri Sivonen 2018-10-18 12:21:35 +03:00
  • 7de454ea16 Restore compatibility with Rust 1.29.x. Henri Sivonen 2018-10-17 14:47:16 +03:00
  • 16902ccf09 Run rustfmt. Henri Sivonen 2018-10-17 14:29:20 +03:00
  • b87f0e49b7 Address clippy lints. Henri Sivonen 2018-10-17 14:28:16 +03:00
  • f1d04bde7a Add method is_single_byte() on Encoding. Henri Sivonen 2018-10-16 15:57:14 +03:00
  • d2f2002516 Merge trunk into packed_simd. Henri Sivonen 2018-10-16 13:23:28 +03:00
  • 36f6444642 Update rustfmt. Henri Sivonen 2018-10-16 13:16:15 +03:00
  • 3410c189b8 Use full github url for safe_encoding_rs_mem Mike Hommey 2018-10-04 09:18:51 +09:00
  • 29d1cae317 Document the prerequisites for running the code generation script. Henri Sivonen 2018-10-16 13:05:13 +03:00
  • caeda336ec Add a word to README. Henri Sivonen 2018-10-16 12:34:07 +03:00
  • ca6724644a Document the new CJK legacy encode speed options. Henri Sivonen 2018-10-16 12:27:03 +03:00
  • a3303be49a Merge the fastcjkencode branch. Henri Sivonen 2018-10-16 10:28:49 +03:00
  • 577380c0fa Add a fast-legacy-encode meta feature. Henri Sivonen 2018-10-08 18:52:57 +03:00
  • 77dab08378 Add option for fast-hanja-encode. Henri Sivonen 2018-10-08 18:17:56 +03:00
  • 110267beaa Use pair of u8 instead of pair of usize in Hangul encode. Henri Sivonen 2018-10-08 17:58:32 +03:00
  • 79532d0f3c Avoid storing non-generated parts of data.rs inside the Python script. Henri Sivonen 2018-10-08 17:47:01 +03:00
  • bc369a0b00 Add fast-big5-hanzi-encode compile-time option. Henri Sivonen 2018-10-08 17:40:42 +03:00
  • 759e8aa812 Increment version number to 0.8.10. Henri Sivonen 2018-10-07 12:30:59 +03:00
  • 955443cf70 Make a test that should assert condititonal on the enablement condition for the assertion. Henri Sivonen 2018-10-07 12:27:21 +03:00
  • 0d7b9d18ca Update rustfmt. Henri Sivonen 2018-10-01 13:57:58 +03:00
  • ae7235fa1a Add data table for fast Big5 Unified Ideograph encode. Henri Sivonen 2018-10-01 13:11:11 +03:00
  • eb3ec27f06 Implement fast-gb-hanzi-encode. Henri Sivonen 2018-09-25 17:01:36 +03:00
  • 0e851eb628 Make fast-kanji-encode additive over less-slow-kanji-encode without warnings. Henri Sivonen 2018-09-25 16:21:29 +03:00
  • e9683096e2 Make fast-kanji-encode faster for EUC-JP. Henri Sivonen 2018-09-25 15:32:02 +03:00
  • bc4cc15253 Non-optimal EUC-JP support for fast-kanji-encode. Henri Sivonen 2018-09-25 15:06:54 +03:00
  • 8b84c6f1df Add fast-kanji-encode for Shift_JIS. Henri Sivonen 2018-09-25 13:58:21 +03:00
  • beaff16757 Add cargo feature fast-hangul-encode. Henri Sivonen 2018-09-25 13:22:54 +03:00
  • 858789ea9a Change minimum Rust version to 1.29.0. Henri Sivonen 2018-10-04 14:09:42 +03:00