http://bugs.winehq.org/show_bug.cgi?id=22469
Summary: FileMaker Pro 11: Text is upside down when builtin gdiplus is used Product: Wine Version: 1.1.43 Platform: x86 URL: http://www.filemaker.com/ OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: gdiplus AssignedTo: wine-bugs@winehq.org ReportedBy: wylda@volny.cz
Created an attachment (id=27526) --> (http://bugs.winehq.org/attachment.cgi?id=27526) Filemaker pro 11
Text in boxes is upside down, when the box is selected. This doesn't happen when native gdiplus is used. Attachment shows that better.
To reproduce: 1. Download the trial version 2. Download demo data http://bugs.winehq.org/attachment.cgi?id=26679 3. Run "wine FileMaker\ Pro.exe" and load data from step 2. 4. Click on any textbox, to make it upside down
http://bugs.winehq.org/show_bug.cgi?id=22469
Wylda wylda@volny.cz changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |madewokherd@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #1 from Wylda wylda@volny.cz 2010-04-25 16:08:15 --- Created an attachment (id=27555) --> (http://bugs.winehq.org/attachment.cgi?id=27555) console log from wine-1.1.43-249-g033bb8a +gdiplus
Log taken and during that, i was clicking on different fields to make the inside text upside down.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #2 from Vincent Povirk madewokherd@gmail.com 2010-04-26 13:30:46 --- Can you add +win,+event to the log, and tell me EXACTLY what you did when you generated it, down to the keystroke/mouse button press? I need some way to zero in on the incorrect draws.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #3 from Wylda wylda@volny.cz 2010-04-26 14:49:57 --- Created an attachment (id=27569) --> (http://bugs.winehq.org/attachment.cgi?id=27569) console log trigered by clicking on the box from wine-1.1.43-269-g379cf46 +gdiplus
(In reply to comment #2)
Can you add +win,+event to the log, and tell me EXACTLY what you did when you
Hi Vincent, image app from PNG image attached in comment #0. So when you click in one of the box (for example the one with red arrow on the left) you select it and text which is inside becomes upside down (like the red arrow on the right side).
This does not happen whit native gdiplus.
Also when the box is selected (and the text is upside down), console is permanently flooded by gdiplus stuff until you cancel the selection.
So this second log is taken only with +gdiplus and first line is excatly from the moment when i selected the box (and inside text become upside down).
Is this enough or you would like to see rather +gdiplus,+win,+event ??
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #4 from Vincent Povirk madewokherd@gmail.com 2010-04-27 09:56:02 --- Well, there's nothing obviously wrong in the log, and it's not using gdiplus to draw the text. So I guess it's either using information from gdiplus or relying on some interaction between gdiplus and gdi32.
Can you do the same thing again with +gdiplus,+font? That'll at least tell me when it's drawing the text.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #5 from Wylda wylda@volny.cz 2010-04-27 11:42:50 --- Created an attachment (id=27578) --> (http://bugs.winehq.org/attachment.cgi?id=27578) Console log from wine-1.1.43-321-g94a3c09 +gdiplus,+font
(In reply to comment #4)
Can you do the same thing again with +gdiplus,+font?
Hi Vincent, i tried to do it by same way, so begins when i'm going to click on the box.
BTW: Download is not large and it is easy to reproduce... Just in case it would be more convenient to do it yourself. That does not mean i do not want to do what you ask. I will gladly do it the way as you wish.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #6 from Wylda wylda@volny.cz 2010-04-27 12:06:03 ---
Just a note to a previous comment. The text which gets upside down is:
"Argentinische Milonga mit Tango und anderen Rhythmen"
Just in case it would make it easier to find in log.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #7 from Vincent Povirk madewokherd@gmail.com 2010-05-03 15:41:05 --- Yes, knowing the text helps.
From the log, it appears that gdi32 is drawing the text right-side-up, and I
don't see a reason to suspect gdiplus should be doing something to the DC used to draw the text.
So for now, I'm stumped.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #8 from Vincent Povirk madewokherd@gmail.com 2010-05-03 15:43:55 --- I forgot to mention that I suspect gdi32 is drawing the text right-side-up to something that isn't the window, and perhaps there's some top-down versus bottom-up dib confusion going on (or perhaps the program is just transforming the bitmap later for some reason and should be drawing the text inverted). But I can't find any indication of that in the log, and I don't know how else to look for it.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #9 from Wylda wylda@volny.cz 2010-06-01 11:31:14 ---
Still present in wine-1.2-rc2-51-gbbd93e5.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #10 from Emerson Prado emerson.prado.eng@gmail.com 2010-12-07 12:17:38 CST --- Created an attachment (id=32393) --> (http://bugs.winehq.org/attachment.cgi?id=32393) Screenshot showing reversed icons and text
"Me too" with Wine 1.3.8 on Linux Mint Debian Edition (LMDE) kernel 2.6.32-5-686 i686. Not only text from fields being edited is upside down. Images also appear reversed. Pls see attached screenshot. Disregarding the toolbar on top, the 3 icons on top center (+, x and magnifying glass) and the small logo on the right mid-top appear upside down, besides the text inside the field "Nome" ("Spam Corp."). I have no clue on how to check which gdi+ FM11 is using. But, if someone wants me to find this out (or whatever else), pls just provide the directions, and I'll happily follow.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #11 from Emerson Prado emerson.prado.eng@gmail.com 2010-12-08 07:51:46 CST --- Created an attachment (id=32413) --> (http://bugs.winehq.org/attachment.cgi?id=32413) Replaced text: new superimposed over old
BTW, not only the text appears upside-down, but it is also kept in view after deleted, replaced or moved. Characters are just not erased. The attached screenshot is the result of typing "Company one", and then replacing "one" by "two". I preferred mentioning this problem here, rather than opening a new bug, because I guess this may be closely related. I apologize if I guessed wrong.
http://bugs.winehq.org/show_bug.cgi?id=22469
Emerson Prado emerson.prado.eng@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |emerson.prado.eng@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=22469
Joerg Schiermeier newsletter@Schiermeier-Software.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |newsletter@Schiermeier-Soft | |ware.de
--- Comment #12 from Joerg Schiermeier newsletter@Schiermeier-Software.de 2010-12-11 12:03:33 CST --- Still not working (well) in wine-1.3.9 with GDIplus from wine.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #13 from Joerg Schiermeier newsletter@Schiermeier-Software.de 2011-01-20 23:51:35 CST --- According to the description in http://wiki.winehq.org/RegressionTesting I did it and here is my result:
Result of the Regression Testing (2011-01-21): ==============================================
db3d756c1e4b03b0fb33f2863ab3d2275dcf00c2 is the first bad commit commit db3d756c1e4b03b0fb33f2863ab3d2275dcf00c2 Author: Vincent Povirk vincent@codeweavers.com Date: Sun Jan 24 15:18:46 2010 -0600
gdiplus: Prefer the builtin version.
:040000 040000 f697af8a8f1f28485cb41662835069206b23cf58 dc52b6e519488f4fec677577cb113cf08e1da028 M dlls
----------------------------------------------
I used my testing database from http://bugs.winehq.org/attachment.cgi?id=26679. Interesting effect: every run where the text was upside down crashes the application after clicking into a control.
btw: this is the same result than in my other test described here: http://bugs.winehq.org/show_bug.cgi?id=21975#c22
GDI+ has serious a fault. The only workaround is at the moment to use the gdi+ windows dll.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #14 from Emerson Prado emerson.prado.eng@gmail.com 2011-01-21 06:08:05 CST --- (In reply to comment #13)
Interesting effect: every run where the text was upside down crashes the application after clicking into a control.
Would it be what I described in Bug 25461? Maybe it's related.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #15 from Joerg Schiermeier newsletter@Schiermeier-Software.de 2011-01-21 13:32:50 CST --- (In reply to comment #14)
Would it be what I described in Bug 25461? Maybe it's related.
Possible. Both are related to GDI+. Also during the Regression Test there where crashes, when wine-GDI+ was active.
If you what to check it in an own test:
I used as the 'bisect good' the wine version 1.1.37 and as 'bisect bad' 1.1.38. Somewhere in between is the suspect.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #16 from Emerson Prado emerson.prado.eng@gmail.com 2011-01-24 13:53:40 CST --- (In reply to comment #15)
(In reply to comment #14)
Would it be what I described in Bug 25461? Maybe it's related.
Possible.
I tested native gdi+ against crashes (why didn't I think of that?). FileMaker still crashes, so I guess it's not related.
If you what to check it in an own test:
I used as the 'bisect good' the wine version 1.1.37 and as 'bisect bad' 1.1.38. Somewhere in between is the suspect.
I did that, but my test failed. All versions tested were bad (text upside down and not being erased from screen during edition). Plus, it started crashing easier - clicking anything or just opening a remote database was enough - after 1 or 2 bisects - related to a "File not found" error while installing FileMaker. At least I could see for myself FileMaker working with native gdi+. No upside down text, correct erasing from screen, and even a speed gain.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #17 from Emerson Prado emerson.prado.eng@gmail.com 2011-01-31 10:12:34 CST --- Tried to add author of suspected patch for his awareness, but Bugzilla couldn't find name or address mentioned. If someone knows a path, pls help.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #18 from Emerson Prado emerson.prado.eng@gmail.com 2011-01-31 10:14:02 CST --- (In reply to comment #17)
Tried to add author of suspected patch for his awareness, but Bugzilla couldn't find name or address mentioned. If someone knows a path, pls help.
Oops, forget it. I just found out "madewokherd" is him...
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #19 from Joerg Schiermeier newsletter@Schiermeier-Software.de 2011-01-31 12:05:39 CST --- (In reply to comment #18)
Tried to add author of suspected patch for his awareness, but Bugzilla find name or address mentioned. If someone knows a path, pls help.
Oops, forget it. I just found out "madewokherd" is him...
ACK.
http://bugs.winehq.org/show_bug.cgi?id=22469
Joerg Schiermeier newsletter@Schiermeier-Software.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |NEW Ever Confirmed|0 |1
--- Comment #20 from Joerg Schiermeier newsletter@Schiermeier-Software.de 2011-01-31 12:41:03 CST --- *** This bug has been confirmed by popular vote. ***
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #21 from Joerg Schiermeier newsletter@Schiermeier-Software.de 2011-01-31 13:14:02 CST --- (In reply to comment #19)
ACK.
Hmm, this small patch doesn't really lead us into the real problem:
---cite--- switch(reason) { - case DLL_WINE_PREATTACH: - return FALSE; /* prefer native version */ - case DLL_PROCESS_ATTACH: DisableThreadLibraryCalls( hinst ); break; ---/cite---
Maybe we will have to send some logs. This patch also change something here: bug 21975, comment 23 and made the application (Quark) unuseable. There I send a log where I traced GDI+ ('WINEDEBUG=fixme-all,trace+gdiplus'). Maybe we will have to do it here similarly.
But until now I didn't dive so deep to the code of gdiplus. It's speaking about unknown things.
!Hasta luego!
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #22 from Joerg Schiermeier newsletter@Schiermeier-Software.de 2011-01-31 20:15:50 CST --- Created an attachment (id=33081) --> (http://bugs.winehq.org/attachment.cgi?id=33081) Test database in FileMaker at 200%
I compaered the two views of both settings: one time with wines GDI+ library and one time with a library from WindowsXP SP3 with a magnification of 200%, done by FileMaker itself.
This shows some difference: -the letterings are placed wrong -the buttons are not three-dimensional
To avoid hunting for ghosts I also created a new database, but there was also the text turning upside down when I clicked into a datafield.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #23 from Emerson Prado emerson.prado.eng@gmail.com 2011-02-07 05:50:30 CST --- This bug seems to depend on, or be related to, Bug 16138 and Bug 24653. These seem to look at the problem closer than this bug and all remaining ones complaining about gdi+. Would it be a good idea to dup this and others to those central bugs? Or, at least, include this in their dependency tree? Maybe if we all focus in the core of this issue, we'll be happier than isolatedly trying to find solutions for specific apps.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #24 from Emerson Prado emerson.prado.eng@gmail.com 2011-02-07 07:33:16 CST --- BTW, I did some search in gdiplus, trying to get some light on the issue, though I'm an absolute beginner when it comes to programming. I'll list below my findings, just to bring them into discussion and not to lose them (I know this must be too basic for real developers, just trying to level out...).
1.(Obvious) GDI+ consists of three parts: vectors, imaging and typography. From what we saw here, we have a problem with embedded (raster?) images and font drawing, but not with the window drawing itself (probably vectorial). It seems vector drawing is OK, being the other two parts at fault. http://msdn.microsoft.com/en-us/library/ms536384%28v=vs.85%29.aspx 2.There can be a failure while transforming coordinates. Either world to page or page to device. http://msdn.microsoft.com/en-us/library/ms536399%28v=vs.85%29.aspx 3.There could be an error in one or more transformation matrices that do coordinate transformations for imaging and typography (if they exist at all, I really dunno). http://msdn.microsoft.com/en-us/library/ms536397%28v=vs.85%29.aspx 4.(More relevant) Class BitmapData has a data member called Stride which affects the horizontal order in which lines are drawn - bottom up or top down. Could this be wrongly interpreted by the builtin GDI+? http://msdn.microsoft.com/en-us/library/ms534421%28v=vs.85%29.aspx 5.Class StringFormat determines how strings are drawn on screen. Similarly to 4, one data member could affect vertical orientation, but I couldn't find which one, if any. http://msdn.microsoft.com/en-us/library/ms534510%28v=vs.85%29.aspx 6.Class ImageCodecInfo stores, um, codec information. Maybe it could be getting something wrong about 4 and 5. http://msdn.microsoft.com/en-us/library/ms534466%28v=vs.85%29.aspx 7.The Wine wiki entry about GDI+ mentions a difference on rendering bitmaps I can't fully understand: "Native gdiplus does all drawing to Bitmap images in software. This means that an application can pass its own buffer, and rendering to the image will go to that buffer. Builtin gdiplus has this capability, but it is not used in most cases because most drawing operations have no software implementation." http://wiki.winehq.org/GdiPlus
Given Bug 16138, everything above must be read as "this can happen in some or all version calls (according to manifest) of GDI+".
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #25 from Joerg Schiermeier newsletter@Schiermeier-Software.de 2011-02-10 21:09:37 CST --- (In reply to comment #23)
This bug seems to depend on, or be related to, Bug 16138 and Bug 24653. These seem to look at the problem closer than this bug and all remaining ones complaining about gdi+. Would it be a good idea to dup this and others to those central bugs? Or, at least, include this in their dependency tree? Maybe if we all focus in the core of this issue, we'll be happier than isolatedly trying to find solutions for specific apps.
This is a good idea. But we both cannot do it: this is either Wyldas job or someone of the BugZilla admins.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #26 from Vincent Povirk madewokherd@gmail.com 2011-02-10 21:34:01 CST --- I don't see any reason why this bug would have anything to do with those other two.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #27 from Emerson Prado emerson.prado.eng@gmail.com 2011-02-11 07:22:41 CST --- (In reply to comment #26)
I don't see any reason why this bug would have anything to do with those other two.
Bug 24653 and this one are about incorrect drawing by gdi+. Bug 16138 is about testing gdi+ correctly. It seems to me the correct testing requested in Bug 16138 would find the incorrect behaviour described in Bug 24653 and this one.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #28 from Joerg Schiermeier newsletter@Schiermeier-Software.de 2011-02-11 12:33:59 CST --- (In reply to comment #26)
I don't see any reason why this bug would have anything to do with those other two.
Maybe - or even not. We don't know it until we test it. I think, Emerson and I are ready to do this testings - but we need infos about what we should search for.
Than we will see more clearly.
Thanks! (This is also for GDI+ dependent bug 21975.)
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #29 from Vincent Povirk madewokherd@gmail.com 2011-02-11 12:34:24 CST --- (In reply to comment #27)
Bug 24653 and this one are about incorrect drawing by gdi+. Bug 16138 is about testing gdi+ correctly. It seems to me the correct testing requested in Bug 16138 would find the incorrect behaviour described in Bug 24653 and this one.
Bug 24653 is only about pixel-accurate positioning of things that already draw. Bug 16138 is about testing all versions of a dll when they are available, but as it happens we already get proper coverage of gdiplus versions across windows versions (and most programs use/expect version 1.0, the most tested version).
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #30 from Joerg Schiermeier newsletter@Schiermeier-Software.de 2011-02-11 12:39:08 CST --- (In reply to comment #28)
(This is also for GDI+ dependent bug 21975.)
btw: FileMaker and Quark were originally developt for MacOS before MacOS X. I don't want to place wrong traces but maybe this is a hint.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #31 from Emerson Prado emerson.prado.eng@gmail.com 2011-02-11 13:10:30 CST --- (In reply to comment #29)
Bug 24653 is only about pixel-accurate positioning of things that already draw. Bug 16138 is about testing all versions of a dll when they are available, but as it happens we already get proper coverage of gdiplus versions across windows versions (and most programs use/expect version 1.0, the most tested version).
You surely know better than me, so I take it at face value. I understand what you said this way: all relevant gdi+ versions are already tested, so the issue is not an untested gdi+, but rather an implementation problem.
(In reply to comment #28)
We don't know it until we test it. I think, Emerson and I are ready to do this testings - but we need infos about what we should search for.
Yes, I'm willing to test stuff, just don't know exactly what and how. BTW, I set up a virtual machine (using VirtualBox), installed XP on it (SP2) and currently run FileMaker there. I also kept Wine installation, also with FileMaker (exactly the same version). Since Wine is not my main path to FileMaker, I can mess it at will, including wiping out, reinstalling, regression testing, patching... I guess this is a good field to play in, just needed a coach... ;)
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #32 from Emerson Prado emerson.prado.eng@gmail.com 2011-05-02 13:36:56 CDT --- (In reply to comment #21)
Hmm, this small patch doesn't really lead us into the real problem:
---cite--- switch(reason) {
- case DLL_WINE_PREATTACH:
return FALSE; /* prefer native version */- case DLL_PROCESS_ATTACH: DisableThreadLibraryCalls( hinst ); break;
---/cite---
Could someone tell to which bug was this patch attached? Also, which Wine version was it commited to? I'd like to read the related bug and test the related versions, because this must tell me something. But I just can't translate commit codes...
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #33 from Joerg Schiermeier newsletter@Schiermeier-Software.de 2011-07-22 19:18:15 CDT --- (In reply to comment #32)
(In reply to comment #21)
Hmm, this small patch doesn't really lead us into the real problem:
(..-)
related versions, because this must tell me something. But I just can't translate commit codes...
This was from my regression test about the bug 21975 comment 22: --- Regression since wine-v1.1.38 up to now (v1.1.40) - in wine-v1.1.37 was working. ---
Not this patch is really the problem. We will have to dig deeper, maybe until we leave the basement of Linux into the stony underground of hardware (sic!)...
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #34 from Joerg Schiermeier newsletter@Schiermeier-Software.de 2011-09-09 21:18:17 CDT --- This bug is still in wine v1.3.27.
I use my testing file (test database) and a newly created file. Both are showing the content upside down, if this elements had the focus.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #35 from Joerg Schiermeier mywine@schiermeier-software.de 2011-12-02 18:29:48 CST --- Just tested. Still present in wine v1.3.33.
Any logfiles needed? I can offer time to do this job. Simply tell my which WINEDEBUG variables had to be set.
http://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #36 from Joerg Schiermeier mywine@schiermeier-software.de 2012-01-18 22:25:55 CST --- Just tested with wine v1.3.37. Still present:
this mean: using the (my) test database file (see Wylda's description at the top) I can open and change some entries. But the text inside the active control is still upside down.
And something: this is also the behaviour without any winetricks! So no native gdiplus or so. This bug is now fully related to wine.
https://bugs.winehq.org/show_bug.cgi?id=22469
--- Comment #37 from Austin English austinenglish@gmail.com --- This is your friendly reminder that there has been no bug activity for 2 years. Is this still an issue in current (1.7.16 or newer) wine?
https://bugs.winehq.org/show_bug.cgi?id=22469
Wylda wylda@volny.cz changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution|--- |FIXED
--- Comment #38 from Wylda wylda@volny.cz ---
Original problem is no longer present in wine-1.7.27. Marking fixed.
https://bugs.winehq.org/show_bug.cgi?id=22469
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #39 from Alexandre Julliard julliard@winehq.org --- Closing bugs fixed in 1.7.28.
https://bugs.winehq.org/show_bug.cgi?id=22469
Anastasius Focht focht@gmx.net changed:
What |Removed |Added ---------------------------------------------------------------------------- URL|http://www.filemaker.com/ |https://web.archive.org/web | |/20210318020452/http://fmdl | |.filemaker.com/maint/emea_f | |ba_ftn/fmpa_11.0.3.312.exe CC| |focht@gmx.net Keywords| |download
-
wine-bugs＠winehq.org -
WineHQ Bugzilla