246454 - content-language HTML equivalent doesn't work

Status: VERIFIED FIXED

(SeaMonkey :: General, defect)

Description

[Anne (:annevk)]

It seems that the 'content-language' HTTP header isn't supported as META element:

 <meta http-equiv="content-language" content="x-example">

The code should be changed to make sure we support both.

Comment 1

[Jo Hermans]

Build 2004061108 (Seamonkey)

Works for me if the language was present in the list in
Preferences->Navigator->Languages.

If I add 'x-example' in the list, I see the green line. If I change the language
to 'nl' (which is in my list), I also see green. When I change it to 'de' (not
in my list), I see a red line. When I add 'de' or 'de-at' to the list, I see the
green line again.

Isn't this the expected behaviour ?

Comment 2

[Anne (:annevk)]

Not at all. The following two test cases should output exactly the same with
default settings afaik:

 * <http://annevankesteren.nl/test/css/s/p-c/lang-005.htm.php>
 * <http://annevankesteren.nl/test/css/s/p-c/lang-007.htm>

Comment 3

[Boris Zbarsky [:bzbarsky]]

Anne, you could move content-language header processing out of the document and
into the content sink.  That would fix this.  See
nsContentSink::ProcessHTTPHeaders and nsDocument::RetrieveRelevantHeaders

Comment 4

[Martijn Wargers (dead)]

Attachment: This seems to work for me

Well, this seems to work for me.
Because I absolutely don't know what I'm doing, this is very likely done in the
wrong way.

Comment 5

[Boris Zbarsky [:bzbarsky]]

Martijn, that looks like a pretty good start to me!  I'm not so sure we want to
make that member public -- I'd rather we either added a setter or moved the
actual setting of the string into nsDocument::SetHeaderData.

Comment 6

[Martijn Wargers (dead)]

Attachment: Improved(?) version

Yes, I already thought that was the sucking part, but I didn't know better.
Perhaps this one is better. It seems to be working fine for me.
I did use NS_ConvertUCS2toUTF8(aData). Not really knowing if this is correct.
The other option, adding a setter, I don't think I know how to make that one.

Comment 7

[Christian :Biesinger (don't email me, ping me on IRC)]

a setter would be a function like "void SetContentLanguage(const nsACString& 
aLanguage) { mContentLanguage = aLanguage; }" 
 
(can't comment on the patch you attached as I don't know this code. 
although... shouldn't you still read the default language pref?) 

Comment 8

[Boris Zbarsky [:bzbarsky]]

> I did use NS_ConvertUCS2toUTF8(aData).

Why?  You should have been able to keep using CopyUTF16toUTF8, no?

> shouldn't you still read the default language pref?

Actually, no.  The fact that we read it is a bug (see
http://lists.w3.org/Archives/Public/public-css-testsuite/2004Jun/0002.html).  It
was supposed to have been filed, but looks like that hasn't happened.

Comment 9

[Martijn Wargers (dead)]

Attachment: More improved(?) version

Ehm, yes. I didn't know why I did not use that in the first place. 
This does compile, although I did not test it, because then I have to compile
the whole build again (which I'll do anyway but that takes a while).

I didn't know I would fix another bug with this patch, but now I can see that.
That was completely unintentional (and I don't even know whether I find that
rule logical).

Comment 10

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 152979 [details] [diff] [review]
More improved(?) version

r+sr=bzbarsky.	Do be sure to test this, though, and ping me to check this in
once the tree reopens, please.

Comment 11

[Martijn Wargers (dead)]

Ok, I've tested this now. It seems to work just fine for the testcases. And the
tree is reopened, so...

Comment 12

[Boris Zbarsky [:bzbarsky]]

checked in.

Comment 13

[Anne (:annevk)]

Thanks Martijn! (And Boris for checking it in.)