http://bugs.winehq.org/show_bug.cgi?id=14658
Summary: Font rendering regression Product: Wine Version: 1.1.2 Platform: All OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: gdi32 AssignedTo: wine-bugs@winehq.org ReportedBy: piernik@gmail.com
Created an attachment (id=15088) --> (http://bugs.winehq.org/attachment.cgi?id=15088) Before regression
Font rendering regression occurs at least in Protel 99 SE schematic view since the following commit:
commit df76fe6f0cdd55568ba67489cc725813804061fd Author: Dmitry Timoshkov dmitry@codeweavers.com Date: Tue Jun 24 16:11:40 2008 +0900
gdi32: Perform the world transform separately from font scaling.
Rendering changes through versions but it stays broken. Screenshots will follow.
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #1 from Krzysztof Kotlenga piernik@gmail.com 2008-07-27 05:41:03 --- Created an attachment (id=15089) --> (http://bugs.winehq.org/attachment.cgi?id=15089) Regression
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #2 from Krzysztof Kotlenga piernik@gmail.com 2008-07-27 05:42:03 --- Created an attachment (id=15090) --> (http://bugs.winehq.org/attachment.cgi?id=15090) Still broken in 1.1.2
http://bugs.winehq.org/show_bug.cgi?id=14658
Krzysztof Kotlenga piernik@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |dmitry@codeweavers.com
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #3 from Dmitry Timoshkov dmitry@codeweavers.com 2008-07-28 01:38:19 --- Created an attachment (id=15134) --> (http://bugs.winehq.org/attachment.cgi?id=15134) gdi32: Avoid loading a bitmap if a glyph transformation is being requested
Does the attached patch help?
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #4 from Krzysztof Kotlenga piernik@gmail.com 2008-07-28 03:42:55 --- No, it doesn't make visible difference. (applied against current head = 1.1.2)
http://bugs.winehq.org/show_bug.cgi?id=14658
Krzysztof Kotlenga piernik@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Version|1.1.2 |1.1.3
--- Comment #5 from Krzysztof Kotlenga piernik@gmail.com 2008-08-23 04:00:54 --- Still broken in 1.1.3
http://bugs.winehq.org/show_bug.cgi?id=14658
Dmitry Timoshkov dmitry@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Platform|All |Other Version|1.1.3 |1.1.2
--- Comment #6 from Dmitry Timoshkov dmitry@codeweavers.com 2008-08-23 06:27:50 --- Please don't change the originally reported version.
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #7 from Krzysztof Kotlenga piernik@gmail.com 2008-09-21 14:26:50 --- Created an attachment (id=16203) --> (http://bugs.winehq.org/attachment.cgi?id=16203) gdi32: Partial revert
Trial and error patch bringing back correct font rendering. I have no idea what's going on there but hopefully it will give you some hints for a real fix.
http://bugs.winehq.org/show_bug.cgi?id=14658
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |patch, regression
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #8 from Dmitry Timoshkov dmitry@codeweavers.com 2008-09-23 02:22:22 --- Is there a demo of the program available so that I could play with it?
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #9 from Krzysztof Kotlenga piernik@gmail.com 2008-09-23 06:17:06 --- Yes, trial is available here: http://www.elektroda.net/download/big/p99se_tr.exe Service Pack 6: http://www.protel.com/downloads/files/protel99seservicepack6.exe Installation guide: http://appdb.winehq.org/objectManager.php?sClass=version&iId=4910&iT... Patch (because trial protection doesn't work under wine): http://rapidshare.com/files/147650192/protel99-wine.exe.html
Please open some example project and search for .sch files in the explorer on the left.
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #10 from Dmitry Timoshkov dmitry@codeweavers.com 2008-10-01 17:08:29 --- Does the following patch help? http://www.winehq.org/pipermail/wine-patches/2008-October/062418.html
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #11 from Krzysztof Kotlenga piernik@gmail.com 2008-10-01 18:04:45 --- No, it doesn't (applied against 1.1.5)
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #12 from Dmitry Timoshkov dmitry@codeweavers.com 2008-10-12 03:56:32 --- Please retest with 1.1.6
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #13 from Krzysztof Kotlenga piernik@gmail.com 2008-10-12 12:32:58 --- No change, still buggy in 1.1.6.
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #14 from Krzysztof Kotlenga piernik@gmail.com 2008-10-16 04:14:30 --- Created an attachment (id=16669) --> (http://bugs.winehq.org/attachment.cgi?id=16669) Partial font trace - bad
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #15 from Krzysztof Kotlenga piernik@gmail.com 2008-10-16 04:17:15 --- Created an attachment (id=16670) --> (http://bugs.winehq.org/attachment.cgi?id=16670) Partial font trace - good
1.1.6 trace with partial revert applied. It's line aligned for easy comparison.
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #16 from Krzysztof Kotlenga piernik@gmail.com 2008-10-30 14:06:14 --- Dmitry, did you see the good/bad trace diff? It's the same string ("hello im a bug") rendered twice. Differences in computations are clearly visible.
Output of freetype.c:1194 (text metrics) - good: InternalLeading = 78 Ascent = 467 Descent = 117 Height = 583
vs all zeros in "bad".
Also output of freetype.c:4533: TRACE("%u,%u,%s,%d,%d\n", lpgm->gmBlackBoxX, lpgm->gmBlackBoxY, wine_dbgstr_point(&lpgm->gmptGlyphOrigin), lpgm->gmCellIncX, lpgm->gmCellIncY);
differs significantly.
If this isn't enough, how else can I help?
http://bugs.winehq.org/show_bug.cgi?id=14658
Dmitry Timoshkov dmitry@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |NEW Ever Confirmed|0 |1
--- Comment #17 from Dmitry Timoshkov dmitry@codeweavers.com 2008-10-30 21:51:33 --- It would be best if I could reproduce the problem. I did several attempts but unfortunately the app doesn't start for me, I'm using the appdb instructions + exe patch. Does it start for you with 1.1.7 or cr current Wine git?
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #18 from Krzysztof Kotlenga piernik@gmail.com 2008-10-31 00:57:55 --- It starts with 1.1.7 (just tested entire installation). Did you patch .exe and few .dlls in System/ too? Also be sure to uncheck "NT Drivers" during installation. If it isn't it, what error do you get?
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #19 from Dmitry Timoshkov dmitry@codeweavers.com 2008-10-31 04:24:08 --- Looks like the patch hasn't been applied to .dll files correctly, now the app runs here. But I don't see any problem, everything renders just fine for me with today's git. Is there any specific .sch file should I look at?
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #20 from Krzysztof Kotlenga piernik@gmail.com 2008-10-31 12:59:32 --- Created an attachment (id=17009) --> (http://bugs.winehq.org/attachment.cgi?id=17009) It's not supposed to look like that
Strange, for me it still looks bad with today's git (see above). Any .sch file will do.
What system are you running? Perhaps I could install it on a virtual machine and see how it behaves there. Maybe it's caused by triggering some freetype bug...
Thanks for your attention on this.
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #21 from Dmitry Timoshkov dmitry@codeweavers.com 2008-11-01 00:42:12 --- Created an attachment (id=17020) --> (http://bugs.winehq.org/attachment.cgi?id=17020) How it looks for me with today's wine git under Ubuntu 7.04 64-bit
Perhaps you miss some font the app is using and it gets substituted by an ugly one. Do you have MS core truetype fonts installed?
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #22 from Dmitry Timoshkov dmitry@codeweavers.com 2008-11-01 04:56:43 --- The following patch should fix the painting slowdown regression in this app: http://www.winehq.org/pipermail/wine-patches/2008-November/064056.html
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #23 from Krzysztof Kotlenga piernik@gmail.com 2008-11-01 06:38:25 --- Of course I have needed fonts and my environment didn't change between checks. Something else must make a difference too. I will investigate further later today on Ubuntu.
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #24 from Krzysztof Kotlenga piernik@gmail.com 2008-11-02 17:19:09 --- I had some troubles* with Ubuntu 7.04 and tried 8.10 today - it renders *wrong*. I'll be back to 7.04 testing tomorrow.
* I'd have saved much time if I knew it just crashes without necessary fonts.
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #25 from Krzysztof Kotlenga piernik@gmail.com 2008-11-03 08:07:12 --- Created an attachment (id=17065) --> (http://bugs.winehq.org/attachment.cgi?id=17065) Ubuntu 7.04 32-bit
Surprise :) Maybe it's 32-bit architecture dependent?
http://bugs.winehq.org/show_bug.cgi?id=14658
--- Comment #26 from Dmitry Timoshkov dmitry@codeweavers.com 2008-11-03 09:17:23 --- Please generate a +dc,+font log with today's git, 'bzip2 -9' it, and attach to the bug.
http://bugs.winehq.org/show_bug.cgi?id=14658
Krzysztof Kotlenga piernik@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED
--- Comment #27 from Krzysztof Kotlenga piernik@gmail.com 2008-11-03 12:56:36 --- It got accidentally fixed together with bug #14983 in commit 376adf9438907f4e6147d8b2c33ea941aca6835d. Bizarre?
http://bugs.winehq.org/show_bug.cgi?id=14658
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #28 from Alexandre Julliard julliard@winehq.org 2008-11-07 10:46:58 --- Closing bugs fixed in 1.1.8.