Replies: 5 comments 17 replies
-
|
scroll should never crash unless there is a graphics driver or other low level bug, which doesn't seem to be the case. I don't like to have crashing bugs, so I appreciate you raising this issue, and any help you can provide debugging it. With your description, it will be a bit hard for me to replicate the problem. I would also need to have a look at the script you are using to launch Playwright and the Chrome worker, etc. So, if you can make it crash reliably, it is much better you run a debug version of scroll and post a stack trace when it crashes, so we can start from there. git clone https://github.com/dawsers/scroll.git
cd scroll
meson setup --buildtype debug build-debug
ninja -C build-debug/
# Launch scroll and make it crash
./build-debug/sway/scroll --debugEdit: And then, of course, post the stack trace coredumpctl dump scrollor if you have |
Beta Was this translation helpful? Give feedback.
-
|
I'll build and run the debug version tomorrow. Thanks!
Apr 26, 2026 4:55:51 PM dawsers ***@***.***>:
…
/scroll/ should never crash unless there is a graphics driver or other low level bug, which doesn't seem to be the case. I don't like to have crashing bugs, so I appreciate you raising this issue, and any help you can provide debugging it.
With your description, it will be a bit hard for me to replicate the problem. I would also need to have a look at the script you are using to launch Playwright and the Chrome worker, etc. So, if you can make it crash reliably, it is much better you run a debug version of scroll and post a stack trace when it crashes, so we can start from there.
git clone https://github.com/dawsers/scroll.git
cd scroll
meson setup --buildtype debug build-debug
ninja -C build-debug/
# Launch scroll and make it crash
./build-debug/sway/scroll --debug
—
Reply to this email directly, view it on GitHub[#261 (comment)], or unsubscribe[https://github.com/notifications/unsubscribe-auth/BGF44FMVZLAGFCNOPPKKCND4XZZUVAVCNFSM6AAAAACYHCSTL6VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTMNZSGE3DAMQ].
Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS[https://github.com/notifications/mobile/ios/BGF44FJVE2DNG3JRPKA3OH34XZZUVA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCNRXGIYTMMBSUZZGKYLTN5XKMYLVORUG64VFMV3GK3TUVJTG633UMVZF62LPOM] and Android[https://github.com/notifications/mobile/android/BGF44FMB4BW3N7G2KEONA5T4XZZUVA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCNRXGIYTMMBSUZZGKYLTN5XKMYLVORUG64VFMV3GK3TUVZTG633UMVZF6YLOMRZG62LE]. Download it today!
You are receiving this because you authored the thread.
[Tracking image][https://github.com/notifications/beacon/BGF44FLW3IZBDB3EHBGYORL4XZZUVBFCNFSM6AAAAACYHCSTL6WGG33NNVSW45C7OR4XAZNRIRUXGY3VONZWS33OINXW23LFNZ2KUY3PNVWWK3TUL5UWJTQA74TMFJTSMVQXG33OUZQXK5DIN5ZA.gif]
|
Beta Was this translation helpful? Give feedback.
-
|
Ok, I'll give that a try, thanks!
Apr 29, 2026 4:21:17 AM dawsers ***@***.***>:
…
If you aren't getting any crashes, it could also be your build was referring to some library that was updated (you mentioned wayland 1.25 too), and having built the debug version, you fixed that by linking against the new version.
If you use the AUR, it doesn't automatically flag packages out of date when one of their dependencies has been updated, so there are cases where the package may be in an unstable state and cause crashes. You can always rebuild/reinstall the package if you suspect that may be the case:
paru -S sway-scroll
# or, if you use the git version
paru -S sway-scroll-git
—
Reply to this email directly, view it on GitHub[#261 (reply in thread)], or unsubscribe[https://github.com/notifications/unsubscribe-auth/BGF44FPQAJT2IGEJFFQMCIL4YG3PFAVCNFSM6AAAAACYHCSTL6VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTMNZVGMYDQNA].
Triage notifications on the go with GitHub Mobile for iOS[https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675] or Android[https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub].
You are receiving this because you authored the thread.
[Tracking image][https://github.com/notifications/beacon/BGF44FL3Y4OCQESYND2NRT34YG3PFBFCNFSM6AAAAACYHCSTL6WGG33NNVSW45C7OR4XAZNRIRUXGY3VONZWS33OINXW23LFNZ2KUY3PNVWWK3TUL5UWJTQA76Q3ZJTSMVQXG33OUZQXK5DIN5ZA.gif]
|
Beta Was this translation helpful? Give feedback.
-
|
Ok, seems worth a shot. Thanks.
May 9, 2026 4:58:34 PM dawsers ***@***.***>:
…
OK. If you want, you can build a more "sensitive" debug version that should crash easily if there is any memory problem. I don't recommend running it in production, because it makes the system a bit slower. But if you want to test it to see if you finally get the crash, just replace the meson step above for the following one:
meson setup --buildtype debug build-debug -Db_sanitize=address
Like I said, if it doesn't produce the crash the first day you run it, you shouldn't run it again. It has quite a bit of overhead.
—
Reply to this email directly, view it on GitHub[#261 (reply in thread)], or unsubscribe[https://github.com/notifications/unsubscribe-auth/BGF44FLRU3PQVU3JYXL2VXT4Z6LWZAVCNFSM6AAAAACYHCSTL6VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTMOBWGUYDENI].
Triage notifications on the go with GitHub Mobile for iOS[https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675] or Android[https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub].
You are receiving this because you authored the thread.
[Tracking image][https://github.com/notifications/beacon/BGF44FNQPMEVYAHC4VI5DDT4Z6LWZBFCNFSM6AAAAACYHCSTL6WGG33NNVSW45C7OR4XAZNRIRUXGY3VONZWS33OINXW23LFNZ2KUY3PNVWWK3TUL5UWJTQBAFLQDJTSMVQXG33OUZQXK5DIN5ZA.gif]
|
Beta Was this translation helpful? Give feedback.
-
|
No problem whatsoever. I'll run the git version tomorrow.
May 12, 2026 7:12:25 PM dawsers ***@***.***>:
…
Thank you. That is exactly what I needed. The problem now is we don't know what version of scroll the stack trace corresponds to. I tried compiling several versions from around May 1st, which is the date of the crash, but I haven't been able to find anything useful (the addresses don't correspond to those builds).
So I am going to need to ask you to install the latest *git* version, write down which version hash you are running, make it crash, and produce the same information you wrote above (stack trace from coredumpctl). Maybe that will get us closer to the problem.
I am sorry this is becoming a lot of work for you. Thank you.
—
Reply to this email directly, view it on GitHub[#261 (reply in thread)], or unsubscribe[https://github.com/notifications/unsubscribe-auth/BGF44FKRCD3G47ACMIB3DOD42OVU3AVCNFSM6AAAAACYHCSTL6VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTMOBZG44TEMQ].
Triage notifications on the go with GitHub Mobile for iOS[https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675] or Android[https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub].
You are receiving this because you authored the thread.
[Tracking image][https://github.com/notifications/beacon/BGF44FNPV73A6V7TKEIMGLL42OVU3BFCNFSM6AAAAACYHCSTL6WGG33NNVSW45C7OR4XAZNRIRUXGY3VONZWS33OINXW23LFNZ2KUY3PNVWWK3TUL5UWJTQBAHLYFJTSMVQXG33OUZQXK5DIN5ZA.gif]
|
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
I am crashing regularly (1 to 2 times a day) for the past week+. It might not be a scroll bug, but it's certainly causing scroll to crash, reliably. Anecdotally, it seems to have started around the time that Wayland 1.25 was released. I have no idea if that's related. I've also been through at least 2 linux kernel upgrades since it started, and it continues to crash.
The circumstance is a little weird, which is why I'm really hesitant to point the finger at scroll.
I run Chrome Stable and it automatically comes up in WS5.
A few weeks ago I started regularly running a Python script with Playwright. It launches another copy of Chrome (completely separate blank profile in /tmp) and automatically moves it to WS11. This worked for weeks without a single crash. I could work for 10 hours in a day, bouncing around between WS5 and WS11 (and everything else I do during the day) with no issues.
Then, a little over a week ago, with a number of scroll updates as well, at least once a day, occasionally twice, when I hit Mod3+5 to go to the "real" Chrome, scroll crashes. It doesn't matter where I am, meaning I can be in a terminal window in WS1 and hit Mod3+5 and it will crash, it doesn't have to be coming from WS11, though that happens with regularity as well.
I haven't turned on any debugging in scroll yet (he says sheepishly), partially because I can immediately type scroll again, instantiate windows (that all offer "Restore Pages"), etc., so I'm back in business pretty quickly. But, I'm getting tired of it...
Anyway, with no debugging on, and I can't copy/paste the errors on screen (because I'm in a login shell session after the crash), I took an actual photo of one of the crashes, in case anyone has an AHA, that's obvious...

If this looks like something that will only be solved by running in full debug mode, let me know. I can look up my old notes from when I last ran it that way, or you can point me to any current best practices for doing that.
Thanks for taking a look!
P.S. I just looked at the image again and this particular crash came from a Brave instance. Meaning, I pressed Mod3+5 when on WS3 which had Brave opened to Perplexity.ai. I didn't want anyone to think that it was a Chrome to Chrome error, or anything to do with Playwright, etc.
Beta Was this translation helpful? Give feedback.
All reactions