As the MIME Sniffing Standard requires and other browser behave.
<rdar://problem/109211253>
Pull request: https://github.com/WebKit/WebKit/pull/13760
We have these relevant files in LayoutTests: fast/xsl/xslt-enc-cyr.xml fast/xsl/xslt-enc.xml fast/xsl/xslt-enc16.xml fast/xsl/xslt-enc16to16.xml fast/xsl/xslt-entity-enc.xml fast/xsl/xslt-text.html The expectation is that the XSLT is decoded using UTF-8. Per https://github.com/WebKit/WebKit/commit/7c5509f22cb96221a2a556cedb3f8d0ef9abc7cf#diff-64e5dc08b95e4ca6ad90d8b8791caffbcb760805f0c8f4c4bd83cda39537ead9 it appears that's a function of the MIME type in WebKit. Whether that's correct is not entirely clear to me. Might also just be that the XSLT "loader" ought to enforce that. However, I don't plan to investigate this further for now.
Dupe to https://bugs.webkit.org/show_bug.cgi?id=257360 ?
I think there's still a potential issue due to TextResourceDecoder.cpp now including a text/xsl check. At least I'm pretty sure this doesn't match how Gecko handles loading XSLT.
(In reply to Anne van Kesteren from comment #5) > I think there's still a potential issue due to TextResourceDecoder.cpp now > including a text/xsl check. At least I'm pretty sure this doesn't match how > Gecko handles loading XSLT. I see. I had to add this check to keep the tests under fast/xsl passing.