ptitSeb
Release 0.3.4+ stalls out with Bookworm + WINE RPi4
Hello, RPi4 64-Bit Bookworm OS with @Itai-Nelken weekly box86 builds stall out when using WINE. Also tried building from source, but when doing the final step of the X86WINE.md directions wine wineboot two box86 lines print and it remains there indefinitely. If I revert to 0.3.3 manually it does not stall out. Bringing to your attention for situational awareness and I will use an older release to solve my issue. Otherwise, thanks for a great product
So this happens also with latest box86 sources? What wine version are you downloading, and it is a 32bits or 64bits version?
Yes, attached is the script I am using to install box86 and WINE on the Raspbian OS. Again, also tried from source and had similar issues. For WINE, I've tried 7.1 devel, 7.0.1 stable, 7.0.2 stable, etc...
sudo wget https://itai-nelken.github.io/weekly-box86-debs/debian/box86.list -O /etc/apt/sources.list.d/box86.list wget -qO- https://itai-nelken.github.io/weekly-box86-debs/debian/KEY.gpg | sudo gpg --dearmor -o /etc/apt/trusted.gpg.d/box86-debs-archive-keyring.gpg sudo apt update && sudo apt install box86:armhf -y
Had the same problem in a bit more complex scenario. The RPi4 on the host is Debian Bookworm, but I was running box86 (and box64) inside of a container running Ubuntu 24.04. The hang was not obvious immediately, because one core gets nailed to 100% even overnight and nothing happens.
I checked out the v0.3.2 tag and built it with the flag from here: https://github.com/ptitSeb/box86/issues/958#issuecomment-2081087153
(I was fine with going with 32 because of hitting problems same as in https://github.com/ptitSeb/box86/issues/600)
So box86 v0.3.2 and box64 v0.3.1 with wine soda 7.0-9 did the job in my case.
The application then started correctly.
I have full stability... no freezing and I'm on pi-apps + rpi os "bookworm"
@gitmanbloggus As I wrote, it's not RPi Bookworm, running inside a container in my case. It should be possible to build from source and get it running without the bug.
I looked through the pi-apps GitHub pipelines to find out what build flags or adaptations they are using. Didn't find the right spot and was short on time to keep looking. There has to be some difference, or maybe in the underlying dependencies during the build that hangs box86.
Well, I'm unsure on what to do with this ticket tbh. I feel like it should be closed now. What do you think @AaronKehl ?
Also, your issue @z0rb seems slightly different, in behaviour and in context, and would prbably be better in a separate ticket.
I thought that it is the same issue just under another context. I get two lines printed and then it hangs indefinitely. The workaround was the same, revert and build the version that @AaronKehl mentioned and exactly that worked. If I check out the tag "v0.3.4" or later and build from source it hangs.
I was hoping newer versions would take care of a memory access exception I get, since I saw that @ptitSeb fixed a "Fixed a nasty bug on memory permission tracking " as the commit message says. :D But can only confirm that or open a new bug once on the most recent version.
I’m fine with the issue being closed out. I don’t plan on trying the pi-apps route and reverting to an older release works for me. I’d much rather time be spent on Pi 5 development than this gremlin :)