https://bugs.winehq.org/show_bug.cgi?id=47558
Bug ID: 47558 Summary: Ubuntu package for 4.12.1 cannot open project in Adobe Premiere Pro 2.0 Product: Packaging Version: unspecified Hardware: x86 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: wine-packages Assignee: wine-bugs@winehq.org Reporter: julius.schwartzenberg@gmail.com CC: dimesio@earthlink.net, michael@fds-team.de, sebastian@fds-team.de Distribution: ---
Created attachment 64946 --> https://bugs.winehq.org/attachment.cgi?id=64946 output of Wine
When I run Adobe Premiere Pro 2.0 with Wine 4.12.1 in Ubuntu, I get the following dialog when trying to open a project: Sorry, a serious error has occurred that requires Adobe Premiere Pro to shut down. We will attempt to save your current project.
Rebuilding Wine myself from Git, currently the latest commit matches 4.12.1, does not show this problem.
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #1 from Julius Schwartzenberg julius.schwartzenberg@gmail.com --- Note that the winehq-stable package with version 4.0.1 does not have this problem. Only with winehq-devel I see this.
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #2 from Rosanne DiMesio dimesio@earthlink.net --- Which version of Ubuntu?
When you built Wine yourself, was it with MingW support?
https://bugs.winehq.org/show_bug.cgi?id=47558
Julius Schwartzenberg julius.schwartzenberg@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Hardware|x86 |x86-64 Distribution|--- |Ubuntu
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #3 from Julius Schwartzenberg julius.schwartzenberg@gmail.com --- Thanks for the quick response!
I'm using Ubuntu 18.04 Bionic AMD64.
I think the version I built myself used MinGW. I followed the instructions for building on Ubuntu, so I used 'apt build-dep' to get the build dependencies and I think I saw the mingw compilers pass by during the build.
Would it make sense to try building without MinGW to see if I can reproduce the problem then? Or maybe there's another build option that the Ubuntu packages use that deviates from what I did now. Another thing to try to narrow it down could be to try other versions of the winehq-devel packages. What do you think would be the easiest?
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #4 from Rosanne DiMesio dimesio@earthlink.net --- The WineHQ packages are built with MingW support. I thought the problem might be related to bug 40350, but if you built Wine with MingW support, you should also see the crash there, too, if that were the cause.
Please attach the config.log from your build.
Have you tried a clean wineprefix?
Unfortunately, there is no legal, free download of Adobe Premiere 2.0, so I can't test this myself.
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #5 from Julius Schwartzenberg julius.schwartzenberg@gmail.com --- Created attachment 64959 --> https://bugs.winehq.org/attachment.cgi?id=64959 config.log for wine64
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #6 from Julius Schwartzenberg julius.schwartzenberg@gmail.com --- Created attachment 64960 --> https://bugs.winehq.org/attachment.cgi?id=64960 config.log for wine32-tools
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #7 from Julius Schwartzenberg julius.schwartzenberg@gmail.com --- Created attachment 64961 --> https://bugs.winehq.org/attachment.cgi?id=64961 config.log for wine32
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #8 from Julius Schwartzenberg julius.schwartzenberg@gmail.com --- That bug seems to be unrelated. There someone has two variants of the same Windows binary, one compiled with a Microsoft compiler and one with MinGW. They should behave identical, but the MinGW one doesn't work.
On the building of Wine side, there would be a choice between GCC and MinGW. MinGW is a variant of GCC that outputs Windows compatible PE binaries instead of ELF binaries. Wine always used it to build the test suite for running on Windows and I think recently I saw more and more parts of Wine itself are being compiled as PEs as well.
I have attached the three config.log files from the different builds.
I'm using the same wineprefix with both versions and switch back and forth to see it work/break.
It appears that only later versions had a trial that can be used everywhere. They behave very differently from this version.
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #9 from Rosanne DiMesio dimesio@earthlink.net --- What version of mingw32-gcc did you use to build Wine?
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #10 from Julius Schwartzenberg julius.schwartzenberg@gmail.com --- Everyhing I used for the build is standard Ubuntu 18.04. For MinGW it shows this version: 7.3 (20180312).
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #11 from Rosanne DiMesio dimesio@earthlink.net --- Created attachment 64962 --> https://bugs.winehq.org/attachment.cgi?id=64962 OBS bionic i586 4.12.1 build log
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #12 from Rosanne DiMesio dimesio@earthlink.net --- Created attachment 64963 --> https://bugs.winehq.org/attachment.cgi?id=64963 OBS x86_64 bionic 4.12.1 build log
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #13 from Rosanne DiMesio dimesio@earthlink.net --- (In reply to Julius Schwartzenberg from comment #10)
Everyhing I used for the build is standard Ubuntu 18.04. For MinGW it shows this version: 7.3 (20180312).
That's what's used in building the packages. I've attached the OBS build logs.
One thing I forgot to ask: you reported this for 4.12.1; does that mean the problem did not occur in previous winehq-devel packages? If it did, can you narrow down which package the problem first appeared in?
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #14 from Julius Schwartzenberg julius.schwartzenberg@gmail.com --- I can only say now that with 4.0.1, so winehq-stable, this does not occur. Do you happen to know if there's an easy trick to install older versions of winehq-devel?
I'll try those to see how they behave.
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #15 from Rosanne DiMesio dimesio@earthlink.net --- You can force a downgrade by specifying the version in the command line, e.g.,
sudo apt install winehq-devel=4.11~bionic wine-devel=4.11~bionic wine-devel-amd64=4.11~bionic wine-devel-i386=4.11~bionic
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #16 from Julius Schwartzenberg julius.schwartzenberg@gmail.com --- Fantastic, that works great! Much quicker than 'git bisect' to initially narrow something down too!!
I found that 4.9 still works, but with 4.10 it broke.
At first sight, I do not see any notable differences between the OBS build and the one I did.
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #17 from Rosanne DiMesio dimesio@earthlink.net --- (In reply to Julius Schwartzenberg from comment #16)
I found that 4.9 still works, but with 4.10 it broke.
At first sight, I do not see any notable differences between the OBS build and the one I did.
There were no changes to the packaging scripts between 4.9 and 4.10 other than the Wine version. Possibly something changed in Wine that doesn't play well with the OBS build environment, but I have no idea what it might be.
https://bugs.winehq.org/show_bug.cgi?id=47558
Zebediah Figura z.figura12@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |z.figura12@gmail.com
--- Comment #18 from Zebediah Figura z.figura12@gmail.com --- (In reply to Rosanne DiMesio from comment #17)
(In reply to Julius Schwartzenberg from comment #16)
I found that 4.9 still works, but with 4.10 it broke.
At first sight, I do not see any notable differences between the OBS build and the one I did.
There were no changes to the packaging scripts between 4.9 and 4.10 other than the Wine version. Possibly something changed in Wine that doesn't play well with the OBS build environment, but I have no idea what it might be.
For what it's worth, 4.10 introduced a lot of PE builtin loading code, and initial conversion of builtin DLLs to PE.
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #19 from Julius Schwartzenberg julius.schwartzenberg@gmail.com --- I guess there is no way to revert back to ELF for a specific build? It could suggest that the specific MinGW version has become more important. Does Wine have a way built-in to compare two sets of builds?
Is the Wine test suite being ran against the Ubuntu builds? If not. Maybe running it against both my build and the Ubuntu build could reveal differences.
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #20 from Dmitry Timoshkov dmitry@baikal.ru --- (In reply to Julius Schwartzenberg from comment #19)
I guess there is no way to revert back to ELF for a specific build? It could suggest that the specific MinGW version has become more important. Does Wine have a way built-in to compare two sets of builds?
'configure --without-mingw' makes it possible to build Wine in an old way (without compiling most of dlls as PE binaries).
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #21 from Julius Schwartzenberg julius.schwartzenberg@gmail.com --- I compiled Wine with '--without-mingw', but the resulting build still runs without problems. The latest Ubuntu 4.13 packages still cause a crash.
https://bugs.winehq.org/show_bug.cgi?id=47558
--- Comment #22 from Rosanne DiMesio dimesio@earthlink.net --- Is this still an issue with current current Wine?
https://bugs.winehq.org/show_bug.cgi?id=47558
Rosanne DiMesio dimesio@earthlink.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |NEEDINFO Ever confirmed|0 |1
--- Comment #23 from Rosanne DiMesio dimesio@earthlink.net --- Is this still an issue?
https://bugs.winehq.org/show_bug.cgi?id=47558
Rosanne DiMesio dimesio@earthlink.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Resolution|--- |ABANDONED Status|NEEDINFO |RESOLVED
--- Comment #24 from Rosanne DiMesio dimesio@earthlink.net --- No response in almost two years. Marking abandoned.
https://bugs.winehq.org/show_bug.cgi?id=47558
Rosanne DiMesio dimesio@earthlink.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #25 from Rosanne DiMesio dimesio@earthlink.net --- Closing abandoned.