Bug 4362 – Add mac-cyrillic to the list of encodings

• Component: DOM
• Status: VERIFIED FIXED

[Alexey Proskuryakov]

This site (one of the largest banks in Russia) sends charset=mac-cyrillic
instead of x-mac-cyrillic. Safari doesn't support such an alias.

I have tried adding an alias to WebCore/kwq/mac-encodings.txt, but for whatever
reason, this makes the build fail:

MacCyrillic: x-mac-cyrillic, mac-cyrillic, xmacukrainian

    /bin/sh -c
/Users/ap/WebKit/WebKitBuild/WebCore.build/Deployment/WebCore.build/Script-932FC0D20824A2C4005B3C75.sh
mac-cyrillic, in mac-encodings.txt, has illegal characters in it (must be all
lowercase alphanumeric)

Also, xmacukrainian (and x-mac-ukrainian) look like aliases to MacUkrainian, not
MacCyrillic.

[Alexey Proskuryakov]

The make-charset-table issue may have been obsoleted by switching to ICU (I'm not really sure, though).

I have filed a request to add mac-cyrillic as an ICU Jitterbug 4859 (<http://dev.icu-project.org/cgi-bin/
icu-bugs>).

[Darin Adler]

All you have to do is add "maccyrillic" without the hyphen. Non-alphanumeric characters are ignored when 
doing the table match and only the first item in each table is allowed to contain them (because it's used in 
reverse-mapping an encoding to its name).

So fixing this bug is as simple as adding that element to the table.

[Alexey Proskuryakov]

Created attachment 5072 [details]
proposed patch

Thank you, I didn't grasp the difference between the first and subsequent
elements :)

I'm left wondering about rdar://3327157 (x-mac-ukrainian should be mapped to
MacCyrillic) - could it be a workaround for some problem that no longer exists?

[Alexey Proskuryakov]

A layout test is probably not necessary here (there are no tests for the presence of charset aliases) - 
correct me if I'm wrong.

[Darin Adler]

Comment on attachment 5072 [details]
proposed patch

r=me

[Adele Peterson]

I've committed this.