Allow direct creation of replacement codec
Created attachment 407732 [details] Patch
Some tests need rebasing, some tests need investigation (fast/encoding/charset-replacement.html).
Created attachment 407765 [details] Patch
Comment on attachment 407765 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=407765&action=review > LayoutTests/imported/w3c/web-platform-tests/encoding/replacement-encodings.any-expected.txt:12 > +PASS replacement - non-empty input decodes to one replacement character. imported/w3c/web-platform-tests/encoding/replacement-encodings.any.worker-expected.txt might need rebasing as well.
Created attachment 407770 [details] Patch
http://trac.webkit.org/r266466
<rdar://problem/68212529>
This test is broken on Windows now: fast/encoding/charset-replacement.html I think it needs to be rebased for Windows.
https://trac.webkit.org/changeset/266504/webkit
And iOS https://trac.webkit.org/changeset/266522/webkit
Comment on attachment 407770 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=407770&action=review > LayoutTests/ChangeLog:9 > + * fast/encoding/charset-replacement-expected.txt: > + * fast/encoding/charset-replacement.html: Why was it necessary to make this test a render tree dump?
Comment on attachment 407770 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=407770&action=review >> LayoutTests/ChangeLog:9 >> + * fast/encoding/charset-replacement.html: > > Why was it necessary to make this test a render tree dump? That does seem like a step in the wrong direction.