479548 - Need "Agfa AC3 Font Compression and Decompression" support

Status: RESOLVED WONTFIX

(Firefox for Android Graveyard :: General, defect)

Description

[Hiroyuki Ikezoe (:hiro)]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; ja; rv:1.8.0.4) Gecko/20060508 Firefox/3.0
Build Identifier: 

Windows Mobile for East Asia uses AC3 Font <http://msdn.microsoft.com/en-us/library/ms925032.aspx>.

We need it to display our own characters.


Reproducible: Always

Comment 1

[Doug Turner (:dougt)]

vlad: is this something we need?

Comment 2

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

We might, it turns out, though it's not going to be critical initially.  But the device CJK fonts are probably going to be AC3 compressed, so we'll either have to figure out how to get it into freetype, or, if available, hook into the custom font renderer stuff on CE6 (and presumably in the winmo based on CE6).

Comment 3

[Brad Lassey [:blassey] (use needinfo?)]

for what its worth, there was discussion on the ft-devel list about this and its apparently a no go for them because of patent issues.  I haven't looked at it, but I suspect we'd have the same issues.

http://www.mail-archive.com/[e-mail]/msg02440.html

Comment 4

[Hiroyuki Ikezoe (:hiro)]

As far as I can investigate, the specification of AC3 font (i.e. MicroType® Express (MTX) Font Format) was submitted to W3C <http://www.w3.org/Submission/MTX/>. But unfortunately Monotype Imaging has the patent.

So we can handle AC3 font only through Windows APIs. I guess the APIs must be Uniscribe on CE6 as Vladimir said.

Comment 5

[John Daggett (:jtd)]

The W3C submission only covers part of this, the compression in that case covers the entire glyph table.  The docs describe AC3 as allowing per-glyph compression/decompression.  My guess is we need more info from Microsoft and/or Monotype for this, or we need to reverse engineer it.  MicroType in general is patent-encumbered, so any implementation will probably have legal issues.  The format has been submitted to W3C but until it is included in a recommendation we would probably need a license from Monotype to implement.

A hacky way around this might be to use GetFontData to pull out an uncompressed version of the glyp table and reconstruct and normal uncompressed TrueType font.

Comment 6

[Hiroyuki Ikezoe (:hiro)]

Wow! If GetFontData can pull out uncompressed data, we can presumably use FT_Stream.

Comment 7

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

I can't imagine that it can -- it just bangs whatever the table is in the file back at you, no?  http://msdn.microsoft.com/en-us/library/aa911428.aspx doesn't mention it being smart in any way.

Comment 8

[Takuro Ashie [:ashie]]

Attachment: A test code to load fonts through GetFontData.

Although I confirmed that an uncompressed image for an AC3 font can be
obtained using GetFontData (you may confirm it with attached dirty patch),
unfortunately this way seems far from usable.
Because calling GetFontData() for AC3 fonts seems very unstable.
Sometimes succeeds, but sometimes fails. 
And once a function call is failed, it never returns AC3 font data anymore
until the user reset the device.
Loading some bigger tables in a AC3 font can easily reproduce this issue.
(You'll fail it when you load fonts in msgothic.ac3 file twice.)

TTF fonts seems fine with this code although it's meaningless...

I'm suspecting the AC3 font driver...

Comment 9

[Takuro Ashie [:ashie]]

Attachment: Test patch V2

Added some missing TTF tables to fix broken layout with MS Gothic.

BTW, to test thie patch, you have to do soft-reset after the first lauch of the fennec due to the issue I mentioned before.

Comment 10

[Brad Lassey [:blassey] (use needinfo?)]

We're going to use the application fonts implemented in bug 491305 to work around this problem for fennec 1.0, so it is no longer blocking.

Comment 11

[Brad Lassey [:blassey] (use needinfo?)]

As noted above, we have a work around for this so there is no reason to block for it.

Comment 12

[Doug Turner (:dougt)]

This year we mothballed windows mobile development.  See:

  http://blog.pavlov.net/2010/03/22/stopping-development-for-windows-mobile/

Marking bugs in the windows mobile / windows ce bucket as WONTFIX.