208213 - GTK:FT2 should use fontEncoding.properties file instead of hardcoding them

Status: RESOLVED WONTFIX

(Core Graveyard :: GFX: Gtk, defect)

Description

[Jungshik Shin]

While investigating the possibility of supporting wide custom font encodings in
GFX:FT2 (a la bug 176290, bug 177877) I realized that nsFreeType.cpp has the
hard-coded list of fonts and their font-encodings. This should be factored out
into fontEncoding.properties file as is done for Win,Mac and Xft. 

Applications of this include bug 204039, bug 203052, bug 204286, bug 176315.

Comment 1

[Jungshik Shin]

Attachment: work in progress

It more or less works, but needs some factoring (and a couple of problems to
fix). I just want to put an interim version somewhere safer than my machine.


In this patch, I went beyond making FT2 code refer to fontEncoding.properties
file (instead of hard-coded arrays). I enabled FT2 code to deal with 'wide
custom' fonts as well as 'narrow' custom fonts. The latter is supported by the
existing code and is for MathML while the former is for CTL support.

Comment 2

[Jungshik Shin]

One line patch for bug 212400 fixed a seemingly hard-to-understand problem 
(with 'the' Korean custom font working while other custom fonts not working).  
 

Comment 3

[Jungshik Shin]

Attachment: v2 patch (still needs some more refining)

Comment 4

[Jungshik Shin]

Attachment: patch v3

I shouldn't have aimed too high :-).  It took a lot more tweaking than I
expected.
With this patch, I fixed bug 179946 and bug 195038 (see also bug 94319) as well
as this bug. I kinda littered files with debug statements and I'll clean them
up later.
In the meantime, I'm uploading this for the peace of mind :-)

Comment 5

[Jungshik Shin]

Attachment: a new patch

basically the same as attachment 129347 [details] [diff] [review] with a bit more extensive comments

Comment 6

[Jungshik Shin]

Brian and Masaki,
Can either of your review the latest patch and move this forward? If both of you
can't at the moment, who would be qualified to review it(someone at Sun)? Thanks.

Comment 7

[Jungshik Shin]

Pete, can you review the latest patch? 

Comment 8

[rbs]

*** Bug 226121 has been marked as a duplicate of this bug. ***

Comment 9

[Jungshik Shin]

bryner, would you care to take a look? It's long ! I have to update the patch
against the trunk (not much change, though) and perhaps have to comment more. 

Comment 10

[rbs]

Comment on attachment 131648 [details] [diff] [review]
a new patch

Removing the sr request to simplify my queue. Care to re-ask once you get the
r.

Comment 11

[Jungshik Shin]

Comment on attachment 131648 [details] [diff] [review]
a new patch

I'm making a new patch against the trunk.

Comment 12

[Jungshik Shin]

Attachment: a new patch updated to the trunk

I updated attachment 131648 [details] [diff] [review] to the trunk (+ my patch for nsAutoBuffer that will
go in as soon as the HEAD is opened up for 1.7beta development)

I haven't yet managed to test it because a new version of freetype library
doesn't have FT_Image_Cache* APIs (bug 234035).

Comment 13

[Reed Loden [:reed]]

jshin, is still needed after 4 years?

Comment 14

[Zack Weinberg (:zwol)]

I am inclined to think this has been rendered moot by subsequent changes, in that I cannot find the offending hardwired table in the code nor can I find any fontEncoding.properties files.