Commit Graph

  • 60058254eb ASCII-accelerate UTF-8 decode. Henri Sivonen 2016-11-21 10:57:37 +02:00
  • aab17421e2 ASCII-accelerate EUC-JP decode. Henri Sivonen 2016-11-18 11:23:34 +02:00
  • c291593bde Avoid weasel words about Mac and IBM encodings. Henri Sivonen 2016-11-17 12:36:08 +02:00
  • ed468b8de6 Run rustfmt. Henri Sivonen 2016-11-17 12:35:18 +02:00
  • 6624250ab0 Make single-byte encode faster by searching in a smarter order. Henri Sivonen 2016-11-17 12:34:37 +02:00
  • e2fcbc5e05 Rename decommit() to commit() to make the name make sense. Henri Sivonen 2016-11-17 09:17:27 +02:00
  • 2d2b83cf4c ASCII-accelerate gb18030 to UTF-16 decode. Henri Sivonen 2016-11-17 09:15:08 +02:00
  • 3015881f02 Add ideas about speeding up CJK encode without full backwards mapping. Henri Sivonen 2016-11-16 21:56:02 +02:00
  • 208cdcde6f Macroize ASCII-accelerated gb18030 decode. Henri Sivonen 2016-11-16 17:15:04 +02:00
  • 6a4a2f3ea0 Make the usage of > vs. >= consistent between two-byte decoders. Henri Sivonen 2016-11-16 16:29:42 +02:00
  • f1d12a2132 ASCII-accelerate gb18030 to UTF-8 decode. Henri Sivonen 2016-11-16 16:28:55 +02:00
  • 6eb4e8fe92 ASCII-accelerate Shift_JIS decode. Henri Sivonen 2016-11-14 14:31:57 +02:00
  • 2f1969e3a1 Parametrize ASCII punctuation in two-byte decoder macro. Henri Sivonen 2016-11-14 13:13:15 +02:00
  • d34e0e3a70 Macroize Big5 decode. Henri Sivonen 2016-11-14 13:10:17 +02:00
  • ca07bc21a9 Macroize EUC-KR decode. Henri Sivonen 2016-11-14 11:10:11 +02:00
  • ae1da9c5e9 ASCII-accelerate EUC-KR to UTF-8 decode. Henri Sivonen 2016-11-14 12:17:06 +02:00
  • 40b2d6403e ASCII-accelerate UTF-16 to UTF-8 encode. Henri Sivonen 2016-11-11 15:40:06 +02:00
  • 4418a4fd52 Note an opportunity to make gb18030/gbk less branchy. Henri Sivonen 2016-11-11 15:22:49 +02:00
  • 18c2505c64 ASCII-accelerate gb18030 encode. Henri Sivonen 2016-11-11 15:20:44 +02:00
  • 15e89c5699 ASCII-accelerate Big5 encode. Henri Sivonen 2016-11-11 14:25:17 +02:00
  • 3da322c90d ASCII-accelerate Shift_JIS encode. Henri Sivonen 2016-11-11 14:12:14 +02:00
  • 9f9cc865bb Update generate-encoding-data.py to match new type changes. Henri Sivonen 2016-11-11 13:44:55 +02:00
  • c0b2d96581 ASCII-accelerate EUC-JP encode. Henri Sivonen 2016-11-11 13:37:59 +02:00
  • 542b1217c3 Remove unused variable. Henri Sivonen 2016-11-11 13:21:04 +02:00
  • 761b4dba9d Remove useless destination space check. Henri Sivonen 2016-11-11 13:20:28 +02:00
  • c78878c660 ASCII-accelerate UTF-16 to EUC-KR encode. Henri Sivonen 2016-11-11 13:16:28 +02:00
  • d94b881699 Put SIMD behind a feature flag. Henri Sivonen 2016-11-11 10:03:17 +02:00
  • 04fbe8aee8 Introduce EncoderResult::unmappable_from_bmp(). Henri Sivonen 2016-11-10 16:57:17 +02:00
  • 9407b428c3 Macroize UTF-8 to single-byte encode. Henri Sivonen 2016-11-10 16:47:57 +02:00
  • 92be595edf Introduce ascii_compatible_bmp_encoder_function!(). Henri Sivonen 2016-11-10 16:41:14 +02:00
  • 6a1bd0436a Run rustfmt. Henri Sivonen 2016-11-10 16:14:10 +02:00
  • c1e81d6242 Macroize UTF-8 to EUC-KR encode. Henri Sivonen 2016-11-10 16:13:58 +02:00
  • d5f5e579df Silence a warning. Henri Sivonen 2016-11-10 15:43:37 +02:00
  • 9ce8578550 ASCII-accelerated encode from UTF-8 to EUC-KR. Henri Sivonen 2016-11-10 15:43:22 +02:00
  • fc7ce9af78 Make some warnings go away. Henri Sivonen 2016-11-10 13:49:17 +02:00
  • 886a16d0d1 Implement accelerated UTF-8 to single-byte encode. Henri Sivonen 2016-11-10 11:09:22 +02:00
  • c918057b62 SIMD-accelerate UTF-16 to single-byte encode. Henri Sivonen 2016-11-09 17:09:57 +02:00
  • 2c0f81dff5 Avoid double bound checks in handles.rs. Henri Sivonen 2016-11-09 15:09:38 +02:00
  • 37d69256fe Make single-byte to UTF-8 work with SIMD and punctuation optimization. Henri Sivonen 2016-11-09 14:50:41 +02:00
  • 43e3075bad Make the single-byte to UTF-16 punctuation optimization actually work. Henri Sivonen 2016-11-09 14:17:12 +02:00
  • 29b00bd3ca SIMD-accelerate ASCII in single-byte to UTF-8 decoding. Henri Sivonen 2016-11-08 13:52:47 +02:00
  • 6cb05a9387 Write down a note to try reading SIMD lanes on ASCII failure. Henri Sivonen 2016-11-08 12:32:43 +02:00
  • d2e113e6a1 Handle ASCII is single-byte to UTF-16 decode using SSE2. Henri Sivonen 2016-11-08 12:27:52 +02:00
  • f1fae7d482 SSE2-accelerated ASCII. Henri Sivonen 2016-11-04 11:58:59 +02:00
  • 8ddfcbaf96 Run rustfmt Henri Sivonen 2016-11-02 13:10:07 +02:00
  • e4a97082d4 Avoid going back to ASCII acceration just for space/punctuation. Henri Sivonen 2016-11-02 15:04:55 +02:00
  • e75fc9f9f4 Don't differentiate with big and little-endian ascii_to_ascii stride. Henri Sivonen 2016-11-01 14:49:28 +02:00
  • d0f2ec68c0 Use ALU word acceleration for ascii_to_ascii on 64-bit LE. Henri Sivonen 2016-11-01 14:47:23 +02:00
  • af779fd98a Use ALU word acceleration for ascii_to_basic_latin on 64-bit LE. Henri Sivonen 2016-11-01 14:43:57 +02:00
  • f88bd1eac4 Simplify ascii_naive. Henri Sivonen 2016-11-01 13:10:27 +02:00
  • 50f44c4972 Test the correctness of ASCII acceleration. Henri Sivonen 2016-11-01 13:06:59 +02:00
  • 84f0ae0add Test encode and decode with 32 different ASCII paddings. Henri Sivonen 2016-11-01 12:36:15 +02:00
  • d2123d660a Mention encoding_bench in README. Henri Sivonen 2016-10-31 17:27:39 +02:00
  • 8efd0baa21 Run unit test on convenience encode/decode, too. Henri Sivonen 2016-10-31 14:53:30 +02:00
  • 114cdf590b Simplify ASCII functions. Henri Sivonen 2016-10-31 14:30:09 +02:00
  • 4729f02b2c When using convenience encode() to UTF-8, return borrow of input. Henri Sivonen 2016-10-31 14:26:37 +02:00
  • df2dfa2827 Merge changes to the benchmark baseline. Henri Sivonen 2016-10-31 13:16:45 +02:00
  • b9c57c2787 Make overlong bogus labels not panic. Henri Sivonen 2016-10-31 11:34:01 +02:00
  • a8e2a7add9 Map 0xCA to U+05BA in windows-1255 per whatwg/encoding#73. Henri Sivonen 2016-10-31 09:35:13 +02:00
  • 33beb57ba9 Make the Rust-only convenience methods returns Cows. Henri Sivonen 2016-10-27 12:45:50 +03:00
  • 6037132a96 Make the code compile again after accidentally landed broken state. Henri Sivonen 2016-10-27 09:28:52 +03:00
  • 3e6dd013b2 Merge changes to Ideas.md. Henri Sivonen 2016-10-27 09:18:06 +03:00
  • f01fd30235 Start ASCII ALU optimizations. Henri Sivonen 2016-10-27 09:17:23 +03:00
  • a37003b32a Add an idea about avoiding a premature return to the SIMD ASCII track. Henri Sivonen 2016-10-21 18:12:13 +03:00
  • 25d9695cfa Move single-byte to UTF-16 to ASCII acceleration. Henri Sivonen 2016-10-21 14:33:50 +03:00
  • 3bb083d037 Mention uconv and ICU looking up even ASCII from a table. Henri Sivonen 2016-10-17 16:58:03 +03:00
  • 0f9a0cb0ee Mention high-bit check in a 16-character stride. Henri Sivonen 2016-10-12 11:06:53 +03:00
  • b790688723 Add a note of what makes ICU faster with single-byte encodings. Henri Sivonen 2016-10-10 12:42:39 +03:00
  • ac77761250 Document some micro-optimizations that are not being pursued. Henri Sivonen 2016-10-03 15:34:02 +03:00
  • 281ccce844 Increment version number to 0.2.4. Henri Sivonen 2016-09-29 14:09:11 +03:00
  • 57ef54d436 Polish ffi docs. Henri Sivonen 2016-09-29 14:08:17 +03:00
  • c5d1b59d1f Improve documentation. Henri Sivonen 2016-09-13 15:58:59 +03:00
  • 31ce32a55b Improve documentation. Henri Sivonen 2016-09-12 12:28:50 +03:00
  • 2e18edb2f1 Update documentation and homepage URLs to point to docs.rs. Henri Sivonen 2016-09-12 10:59:54 +03:00
  • ea4ffa7956 Increment version number to 0.2.3. Henri Sivonen 2016-09-06 13:51:59 +03:00
  • ac4423f3e6 Point to docs.rs for the API docs. Henri Sivonen 2016-09-06 13:50:13 +03:00
  • 02014fc90a Remove bogus "_without_replacement" from naming. Henri Sivonen 2016-09-06 13:36:39 +03:00
  • ba2347c92a Document unit tests as done. Henri Sivonen 2016-09-06 12:52:18 +03:00
  • 459432f55c Test single-byte encodings. Henri Sivonen 2016-09-06 12:51:31 +03:00
  • ef843bb45b Document the UTF-16 to UTF-8 bug fix in the release notes. Henri Sivonen 2016-09-06 12:07:11 +03:00
  • 370aa9ceec Test UTF-8. Henri Sivonen 2016-09-06 12:05:29 +03:00
  • 3a8f905148 Test empty input. Henri Sivonen 2016-09-06 10:24:06 +03:00
  • 4a8bb720a4 Roll back to rustfmt 0.4.1. Henri Sivonen 2016-09-06 09:39:27 +03:00
  • 2d67eca876 Add usage examples. Henri Sivonen 2016-08-12 11:37:07 +03:00
  • 5b57c4ea63 Add a missing ')'. Henri Sivonen 2016-08-09 14:04:19 +03:00
  • e714a13336 Add more FFI documentation. Henri Sivonen 2016-08-09 13:53:12 +03:00
  • 4bec0323e7 Increment version number to 0.2.2. Henri Sivonen 2016-08-09 11:06:36 +03:00
  • ebe76f16ee More accurate API mapping for raw_feed()/raw_finish(). Henri Sivonen 2016-08-09 11:06:06 +03:00
  • b7ea0ac4c7 Add more docs explaining the design. Henri Sivonen 2016-08-09 10:54:40 +03:00
  • a81f275bb2 Tweak error reporting docs. Henri Sivonen 2016-08-08 16:42:52 +03:00
  • 162759a9b8 Add more documentation for the design. Henri Sivonen 2016-08-08 16:33:54 +03:00
  • 794b0d6d30 Improve rust-encoding-related documentation. Henri Sivonen 2016-08-08 15:09:56 +03:00
  • 33ec35e7de Update README. Henri Sivonen 2016-08-08 14:30:32 +03:00
  • bce68b65a9 Add keywords to Cargo.toml. Henri Sivonen 2016-08-08 14:20:37 +03:00
  • d9c0c828ce Add Encoder::encode_from_utf8_to_vec_without_replacement(). Henri Sivonen 2016-08-08 14:15:37 +03:00
  • c16a1b01b2 Start documenting the API mapping from rust-encoding. Henri Sivonen 2016-08-04 14:49:14 +03:00
  • 1a0ea139c6 Document memory and buffering behavior. Henri Sivonen 2016-08-04 14:01:09 +03:00
  • 65a709d371 Clarify performance non-goal for UTF-16. Henri Sivonen 2016-08-04 13:18:51 +03:00
  • bb42978d3f Document Decoder.decode_to_utf[8|16]_without_replacement(). Henri Sivonen 2016-08-04 13:00:28 +03:00
  • 8273f3e2eb Increment version to 0.2.1. Henri Sivonen 2016-07-14 15:07:48 +03:00