[encoding] Add tests for GBK and gb18030 encoding #26385
Conversation
|
I added another file since inexorabletash's approval. It includes 82 code points that are decoded with GBK differently in Chromium and WebKit before my change. |
|
Thanks for writing these. There was also #20361 but the i18n WG never got around to finishing those and cleaning them up seemed like quite a bit of effort. |
There was a problem hiding this comment.
Sticking with "Approve" but a last minute observation that many variables aren't declared/scoped (with var, let or const) in many of these files/functions. Probably copy/paste from existing tests so not a big deal to fix, but I wouldn't object if they got a sprinkle of keywords.
These tests come from https://bugs.webkit.org/show_bug.cgi?id=218380 All tests pass in Firefox and in WebKit after that change. Chrome has the following deviations from the standard: 1. The URL GBK encoding of the Yen sign (U+00A5) is %A3%A4 instead of %26%23165%3B 2. Decoding 0x80 with gb18030 results in a replacement character instead of the Euro sign (U+20A0) 3. The GBK decoder does not use the gb18030 decoder, which makes it less permissive.
…it/Chromium and spec.
annevk
force-pushed
the
gbk_gb18030_encoding
branch
from
da9f6ab
to
ec62bb8
Compare
|
I have rebased this since a bot seemed to be stuck and I also added a commit to clean up the const/let/semicolon stuff since I think I was responsible for the original bad style here. |
Refs: web-platform-tests/wpt#26385 PR-URL: #36659 Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: Rich Trott <rtrott@gmail.com>
Refs: web-platform-tests/wpt#26385 PR-URL: #36659 Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: Rich Trott <rtrott@gmail.com>
These tests come from https://bugs.webkit.org/show_bug.cgi?id=218380
All tests pass in Firefox and in WebKit after that change.
Chrome has the following deviations from the standard: