https://bugs.winehq.org/show_bug.cgi?id=56044
Bug ID: 56044 Summary: Wrong RTF text saved by application Product: Wine Version: 9.0-rc1 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs@winehq.org Reporter: leyyyyy@gmail.com Distribution: ---
Created attachment 75699 --> https://bugs.winehq.org/attachment.cgi?id=75699 Comparison in Meld
In windows application will update only the following lines in the file:
<a:ModificationDate>1632770938</a:ModificationDate> A=char(1)
But Wine adds its own data which breaks original application.
You can check screenshot for details.
Error happens at least in versions 8/9 (staging is used).
https://bugs.winehq.org/show_bug.cgi?id=56044
Eric Pouech eric.pouech@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |eric.pouech@gmail.com
--- Comment #1 from Eric Pouech eric.pouech@gmail.com --- can you attach the result of running your example with WINEDEBUG=+richedit ./wine meld.exe [[your args]]
(there's a flag in richedit control to tell whether to output in RTF format or in pure text... your image shows that it outputs in RTF format, while you'd expect pure text). still need to figure out why.
TIA
https://bugs.winehq.org/show_bug.cgi?id=56044
Bernhard Übelacker bernhardu@mailbox.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |bernhardu@mailbox.org
--- Comment #2 from Bernhard Übelacker bernhardu@mailbox.org --- Hello Maxim, I tried to reproduce this issue, but plain text files with the content visible in your screenshot, with the "<a:Comment>" section equal, but I did just receive the first and the third difference.
If publication of those files is no problem, you could attach the two files which produce the wrong difference in Meld.
My test was with this version: https://download.gnome.org/binaries/win32/meld/3.22/Meld-3.22.0-mingw.msi
https://bugs.winehq.org/show_bug.cgi?id=56044
--- Comment #3 from Maxim leyyyyy@gmail.com --- I do not know why my post not saved.
I wrote that program which saves file is Power Designer.
It happens when you edit RDBMS resources.
Meld is used only to show which wrong lines are added to resource file. So it is just FYI.
Original issue is related to PD not Meld.
https://bugs.winehq.org/show_bug.cgi?id=56044
--- Comment #4 from Bernhard Übelacker bernhardu@mailbox.org --- Thanks for this clarification.
So attaching terminal output here with the WINEDEBUG environment set might still be informative (https://wiki.winehq.org/FAQ#General):
WINEDEBUG=+richedit wine "Power Designer.exe" >> /tmp/output.txt 2>&1
Or a free and legal download for the application, and some instructions how to generate such a rtf text file from it, could be helpful.
https://bugs.winehq.org/show_bug.cgi?id=56044
--- Comment #5 from Maxim leyyyyy@gmail.com --- Created attachment 75843 --> https://bugs.winehq.org/attachment.cgi?id=75843 Output
https://bugs.winehq.org/show_bug.cgi?id=56044
--- Comment #6 from Maxim leyyyyy@gmail.com --- Created attachment 75844 --> https://bugs.winehq.org/attachment.cgi?id=75844 Related text changes (it makes file broken)
https://bugs.winehq.org/show_bug.cgi?id=56044
--- Comment #7 from Eric Pouech eric.pouech@gmail.com --- Created attachment 75900 --> https://bugs.winehq.org/attachment.cgi?id=75900 exploration patch
Thanks for the log.
Unfortunately, from retrying with the messages logged in #2, I'm always getting RTF content from EM_STREAMOUT.
I can see two possible cause of the bug: - I couldn't repro some of the messages because from passed structures were not dumped. Esp. in CHARFORMAT2A where there are bits regarding revision marks and modification, which the bug is about. Could be correlated, could be not. - the caller of EM_STREAMOUT reparses the RTF output. There are a couple of differences in generated RTF between Wine's, but the parser shouldn't be too tied to some constructs. I spotted one diff that could cause some issue (but that's a very long long shot). (and maybe it's the combination of the two items here, where the caller grabs the RTF for revised code!)
So, the attached patch should help dig into this two paths. Could you please regenerate the logs as done in #5 with Wine recompiled with this patch.
TIA
https://bugs.winehq.org/show_bug.cgi?id=56044
--- Comment #8 from Maxim leyyyyy@gmail.com --- I never compiled Wine before. I just used it via pre-built staging package (OpenSUSE Thumbleweed).
Staging version is required, it has some patches which allows to run this app.
I will try...
-
WineHQ Bugzilla