Re: Future improvment in Mozilla 1.9

[email protected] wrote:
> Tony Mechelynck wrote:
>> Someone replied to me privately but not on the newsgroup, with a 
>> from-address of [email protected] . That address is invalid. Please 
>> post on the newsgroup instead.
>>
>>
>>
>> Best regards,
>> Tony.
> 
> That would have been me. I am [email protected].

I got your private "test" email, and just like this one, it still had the 
wrong from-line. If I were you, I'd check my account settings for all "mail" 
and "news" accounts. (In Thunderbird it's on the first page of settings for 
each account.)

> 
> I was seeking information about existing test cases of the i18n 
> functionality. I am being asked to do a sort of general survey. If 
> anyone has questions, ideas, or suggestions, I would be glad to chat.
> 
> I think that some things that are being done now, or that are available 
> now, would make this stuff much more testable and tested.
> 
> I am also on the IRC channels (#qa, #firefox, #bmo, #litmus, &c), 
> usually as "ray".
> 
> thanx - ray

According to your private email to me, you may have come under the impression 
that I had some authority with Mozilla. Let me dissipate this 
misunderstanding: my only relation to Mozilla is that I use Firefox, 
Thunderbird and SeaMonkey, and that I have had some activity on 
bugzilla.mozilla.org (both as reporter and commenter) and also on some of the 
present newsgroups (where my past training as math/physics teacher may have 
shown as a tendency to "pedagogy" in my answers).

My worst gripe with i18n in Mozilla products is exhibited (AFAIK) only in 
Linux but not in Windows (I don't know about Mac). In Linux itself, it's a 
Gecko-only bug in the sense that Konqueror hasn't got it. It is described by 
bug 81367 (reported in May of 2001 and still open), blocking bug 115713 (meta 
bug on bidi layout issues).

The essence of it, is that on Linux (using GTK2+Xft, and probably also when 
using GTK1) Firefox seems not to know that composing characters (such as, 
among others, the short vowels of Hebrew and Arabic, which are not used in 
most everyday texts but are used in sacred writings, in dictionaries, in 
language lessons, and wherever they are necessary to prevent an ambiguity) 
must be drawn "superimposed" to the preceding "spacing" character and not 
"following" it. Bug 81367 has attachments (one in HTML which displays "wrong" 
on Linux and one in JPG which displays "right" everywhere); my welcome page 
http://users.skynet.be/antoine.mechelynck/ has text in several languages; the 
Arabic text left of the photo (below the Russian and above the Chinese) 
displays wrong on Gecko browsers on Linux, but not on Windows and, on Linux, 
not in Konqueror. The rest of the Arabic text (right of the photo and on both 
sides of the counter), which is not vocalised, displays correctly wherever I 
have tested it.

IMHO this bug is making all Gecko browsers useless on Linux for any language 
which uses composing characters: Biblical Hebrew and Koranic Arabic are the 
best-known of these that come to mind. Undoubtedly there are others.

I have tried, without success, do do a screen capture of the "wrong" display 
on this Linux+kde system. AFAICT, the PrintScreen key just does nothing.


Best regards,
Tony.