You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Mounts the AppImage using ./Ghostty-1.1.0-x86_64.AppImage --appimage-mount
Goes to sleep
Wakes from sleep
Tries to access the files at the mountpoint - are the files still mounted? Are there errors when opening/reading them? Or is there only an error when trying to run AppRun?
On which operating systems and sleep modes can this be reproduced?
About the sleep mode, I am not sure. I can tell that if I click on Sleep in KDE and resume a few seconds later, there is no issue.
But I have a crash every morning with the computer sleeping since the previous end of day.
See the suspend and resume systemd logs in case you can find anything useful about the sleep mode.
No @probonopd the issue is happening with the type2-runtime and we fixed it by switching to the uruntime, the appimage with the uruntime no longer crashes on sleep.
WIth that said I don't know why @romainreignier tested here with version 1.1.2 which there are 3 separate releases and only the oldest one of that one is using the type2-runtime.
Here we were told that there was no crash anymore on sleep, that artifact is using the uruntime.
I later asked @romainreignierto test the AppImage that crashes on sleep, which is v1.1.2+1and uses the type2-runtimeand there was no crash anymore. when extracted but it crashes when run normally.
TLDR:
v1.1.2+1 and older crash on sleep unless extracted, these use type2-runtime.
v1.1.2+2 and newer do not crash on sleep under any condition, these use uruntime.
On top of that I had already seen this issue at linuxdeploy which happened after it was switched to use the static type2-runtime, and in that case the dev says that the crash does not happen when extracted as well.
Me? I can't but that's likely because I have one of those chinese X99 motherboards that don't have the ultra slow power states enabled, when I send the PC to sleep it doesn't really go to sleep lol.
I tried in any case, here is what I was going to test:
You can find out if an appimage is using the uruntime by grepping the URUNTIME_MOUNT string which is an internal env variable of it which is not present in the type2-runtime.
Also notice that the mount point names are different, while the type2-runtime it is .mount_ghosttHlAdkm in the uruntime it is a bunch of numbers .mount_5354263854163925111.
So @romainreignier was indeed testing the right appimage here because --appimage-mount had the .mount_Ghostt*** output.
Hi @probonopd indeed, I was testing with the version v1.1.2+1 using type2-untime.
If you need it, how should I use appimagetool-x86_64.AppImage to keep it running while the PC goes to sleep?
Try to access the files at the mountpoint - are the files still mounted? Are there errors when opening/reading them? Or is there only an error when trying to run AppRun?
Come on, it's not that I've deleted it. Just hidden as off-topic.
If there i a bug in type2-runtime, then that bug needs to be fixed - pointing to inofficial third-party runtimes doesn't really solve it.
Or, said differently: The "solution" is not to use an inofficial third-party runtime, the solution would be to sit down, reproduce the issue, and submit a PR that resolves this bug in type2 runtime. (If I was still using Linux, I'd probably have solved it by now - but I am not using Linux anymore.)
pointing to inofficial third-party runtimes doesn't really solve it.
It literally solves it... Just not for this project's codebase, specifically (not yet). Workarounds are part of every bug tracker, and knowing the history of why this workaround isn't upstreamed is valuable information. This feels petty and given how much research and explanation went into @Samueru-sama's messages, gives me a very uneasy feeling.
1. set system to sleep after 5 minutes
2. open the AppImage and leave it opened
3. do not turn off the PC for any reason
4. go in the kitchen and cook something, or go out for a walk
5. go back to the PC and, if it sleeps, awake it
If the AppImage crashes, this means that the issue is still there. If not, repeat the steps above 2 or 3 times.
@ivan-hc yes I had this issue repeatedly with the latest version of the runtime. With the reverted runtime, with a version of Librewolf that I built locally, I no longer have this issue.
Activity
probonopd commentedon Mar 3, 2025
After looking at pkgforge-dev/ghostty-appimage#27 (comment), it would be interesting to see what happens if the user
./Ghostty-1.1.0-x86_64.AppImage --appimage-mountAppRun?On which operating systems and sleep modes can this be reproduced?
Samueru-sama commentedon Mar 3, 2025
@romainreignier Can you try the above with the AppImage that crashes on sleep?
romainreignier commentedon Mar 5, 2025
Hi @probonopd
I have tried your procedure and once resumed, the temp directory is not there anymore.
This is on Kubuntu 22.04.5:
About the sleep mode, I am not sure. I can tell that if I click on Sleep in KDE and resume a few seconds later, there is no issue.
But I have a crash every morning with the computer sleeping since the previous end of day.
See the suspend and resume systemd logs in case you can find anything useful about the sleep mode.
resume_logs.txt
probonopd commentedon Mar 5, 2025
Thanks @romainreignier. It seems like your AppImage is not using https://github.com/AppImage/type2-runtime but https://github.com/VHSgunzo/uruntime. So closing here, please report at https://github.com/VHSgunzo/uruntime/issues.
Thanks!
Reference:
probonopd commentedon Mar 5, 2025
If you can reproduce the issue using the AppImages from https://github.com/AppImage/appimagetool/releases/tag/continuous, then type2-runtime is also affected and we can reopen this.
Samueru-sama commentedon Mar 5, 2025
No @probonopd the issue is happening with the type2-runtime and we fixed it by switching to the uruntime, the appimage with the uruntime no longer crashes on sleep.
WIth that said I don't know why @romainreignier tested here with version
1.1.2which there are 3 separate releases and only the oldest one of that one is using the type2-runtime.Here we were told that there was no crash anymore on sleep, that artifact is using the uruntime.
I later asked @romainreignier to test the AppImage that crashes on sleep, which is
v1.1.2+1and uses the type2-runtime and there was no crash anymore. when extracted but it crashes when run normally.TLDR:
v1.1.2+1and older crash on sleep unless extracted, these use type2-runtime.v1.1.2+2and newer do not crash on sleep under any condition, these use uruntime.Samueru-sama commentedon Mar 5, 2025
On top of that I had already seen this issue at linuxdeploy which happened after it was switched to use the static type2-runtime, and in that case the dev says that the crash does not happen when extracted as well.
probonopd commentedon Mar 5, 2025
Sorry, my bad. Was tired from dayjob work ;)
probonopd commentedon Mar 5, 2025
Can you reproduce the issue using the AppImages from https://github.com/AppImage/appimagetool/releases/tag/continuous though?
Samueru-sama commentedon Mar 5, 2025
Me? I can't but that's likely because I have one of those chinese X99 motherboards that don't have the ultra slow power states enabled, when I send the PC to sleep it doesn't really go to sleep lol.
I tried in any case, here is what I was going to test:
You can find out if an appimage is using the uruntime by grepping the
URUNTIME_MOUNTstring which is an internal env variable of it which is not present in the type2-runtime.Also notice that the mount point names are different, while the type2-runtime it is
.mount_ghosttHlAdkmin the uruntime it is a bunch of numbers.mount_5354263854163925111.So @romainreignier was indeed testing the right appimage here because
--appimage-mounthad the.mount_Ghostt***output.romainreignier commentedon Mar 6, 2025
Hi @probonopd indeed, I was testing with the version
v1.1.2+1using type2-untime.If you need it, how should I use
appimagetool-x86_64.AppImageto keep it running while the PC goes to sleep?probonopd commentedon Mar 6, 2025
As above,
appimagetool-x86_64.AppImage --appimage-mountromainreignier commentedon Mar 7, 2025
I can confirm the same behavior.
After the sleep:
16 remaining items
probonopd commentedon Oct 21, 2025
Come on, it's not that I've deleted it. Just hidden as off-topic.
If there i a bug in type2-runtime, then that bug needs to be fixed - pointing to inofficial third-party runtimes doesn't really solve it.
Or, said differently: The "solution" is not to use an inofficial third-party runtime, the solution would be to sit down, reproduce the issue, and submit a PR that resolves this bug in type2 runtime. (If I was still using Linux, I'd probably have solved it by now - but I am not using Linux anymore.)
tresf commentedon Oct 21, 2025
It literally solves it... Just not for this project's codebase, specifically (not yet). Workarounds are part of every bug tracker, and knowing the history of why this workaround isn't upstreamed is valuable information. This feels petty and given how much research and explanation went into @Samueru-sama's messages, gives me a very uneasy feeling.
probonopd commentedon Oct 21, 2025
Is there a fix for type2-runtime that can be upstreamed? If so, I have missed it.
Samueru-sama commentedon Oct 21, 2025
Revert this 😭
probonopd commentedon Oct 21, 2025
Thanks @Samueru-sama. 💯 I had missed that indeed.
Could someone who can reproduce the issue please test with the runtime here:
If there issue doesn't persist there, then we need to revert #93 and address #92 (which was the reason to introduce it) in some other way.
davidvedvick commentedon Oct 22, 2025
I am not having the issue yet with the reverted runtime and a custom appimage build of LibreWolf, but it's only been a few sleep/resume cycles so far.
davidvedvick commentedon Oct 30, 2025
Still no issues after resume here :)
ivan-hc commentedon Oct 30, 2025
@davidvedvick to reproduce the issue, you should:
If the AppImage crashes, this means that the issue is still there. If not, repeat the steps above 2 or 3 times.
probonopd commentedon Nov 1, 2025
@Samueru-sama would you be able to confirm that #134 resolves this issue? Thank you very much.
davidvedvick commentedon Nov 1, 2025
@ivan-hc yes I had this issue repeatedly with the latest version of the runtime. With the reverted runtime, with a version of Librewolf that I built locally, I no longer have this issue.
Samueru-sama commentedon Nov 1, 2025
I was never able to replicate the issue as mentioned here
TheAssassin commentedon Nov 7, 2025
Let's merge this then. I don't think a few stale mountpoints or processes are a huge deal outside some embedded world with limited RAM.
probonopd commentedon Nov 8, 2025
Thanks everyone who contributed to the solution, especially @Samueru-sama 👍
probonopd commentedon Nov 8, 2025
Please see/test the alternative solution to the original problem, #92: