msort_with_tmp.part.0 #243

eeems-sentry · 2022-01-13T13:40:05Z

Sentry Issue: OXIDE-H

SIGABRT: Fatal Error: SIGABRT
  File "msort.c", line 99, in msort_with_tmp.part.0
  File "../../include/QtCore/../../../git/src/corelib/io/qdebug.h", line 129, in QDebug::maybeSpace
  File "../../include/QtCore/../../../git/src/corelib/io/qdebug.h", line 166, in QDebug::operator<<
  File "../../include/QtCore/../../../git/src/corelib/io/qdebug.h", line 227, in QtPrivate::printSequentialContainer<T>
  File "../../include/QtCore/../../../git/src/corelib/io/qdebug.h", line 240, in operator<< <T>
...
(1 additional frame(s) were not displayed)

The text was updated successfully, but these errors were encountered:

Eeems · 2022-01-13T13:41:17Z

@LinusCDE could you attach your build of tarnish I can get debug symbols from it?

LinusCDE · 2022-01-13T13:55:59Z

The executed app was a build of inky-2048 v0.2.0. Not exact binary but same source code. The binary and launcher files are attached.

executed_app.tar.gz

Was run on the rM 2. The binary does not have the libremarkable build-in swtfb enabled. So used the usual rm2fb_client wrapper stuff.

Eeems · 2022-01-13T15:57:50Z

Looks like I can reproduce this with the following steps:

Upload inky-2048 and draft file to device
Go to oxide launcher
import apps from draft

eeems-sentry · 2022-01-13T15:59:05Z

Sentry issue: OXIDE-J

Eeems · 2022-01-16T18:26:03Z

@LinusCDE could you test this again with the latest build? I'm no longer getting crashes when importing draft apps.

LinusCDE · 2022-01-17T00:44:01Z

Will update it tomorrow.

LinusCDE · 2022-01-17T13:48:00Z

Updated the build. Tried to confuse it with rapid swiping and also let it wait about 5-10 mins before continuing. Task manager + screenshots worked and i didn't encounter any unexpected crash/restart of the launcher app.

It seems though, that there was a bugreport submitted. I'm not sure what caused it. Seemed to be early on when testing the new build but I'm not sure exactly when.

Tried to get the journallogs for that time, but they usually only go like 10 minutes back. Not sure if this is configured or because the rm2fb warning keeps spamming away older logs. If feasible, it would probably help to just silence this particular warning since this is 99% of the logs on a rM 2:

Jan 17 14:37:05 reMarkable tarnish[8098]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.
Jan 17 14:37:05 reMarkable tarnish[8102]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.
Jan 17 14:37:05 reMarkable tarnish[8106]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.
Jan 17 14:37:08 reMarkable tarnish[8109]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.
Jan 17 14:37:08 reMarkable tarnish[8112]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.
Jan 17 14:37:08 reMarkable tarnish[8113]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.
Jan 17 14:37:08 reMarkable tarnish[8117]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.
Jan 17 14:37:08 reMarkable tarnish[8121]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.
Jan 17 14:37:11 reMarkable tarnish[8124]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.
Jan 17 14:37:11 reMarkable tarnish[8127]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.
Jan 17 14:37:11 reMarkable tarnish[8128]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.
Jan 17 14:37:11 reMarkable tarnish[8132]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.
Jan 17 14:37:11 reMarkable tarnish[8136]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.
Jan 17 14:37:14 reMarkable tarnish[8139]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.
Jan 17 14:37:14 reMarkable tarnish[8142]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.
Jan 17 14:37:14 reMarkable tarnish[8143]: ERROR: ld.so: object '/opt/lib/librm2fb_client.so.1' from LD_PRELOAD cannot be preloaded (internal error): ignored.

Eeems · 2022-01-17T14:29:56Z

Did you test importing draft apps like in my reproduction steps?
As for silencing the preload error, that would require changing how things are called to no longer pass the environment variable to them. Feel free to open a PR that accomplishes that.

LinusCDE · 2022-01-17T14:34:47Z

Did you test importing draft apps like in my reproduction steps?

I might have triggered a "Import Apps", but didn't clean oxides config, so not sure if that would do anything.

As for silencing the preload error, that would require changing how things are called to no longer pass the environment variable to them. Feel free to open a PR that accomplishes that.

Already suspected, that the output is just passed along without any real control over it. Fixing the cause seems also a good idea. Not sure if I'll get to do that though. But should be reasonably doable.

Eeems · 2022-01-17T16:51:29Z

Did you test importing draft apps like in my reproduction steps?

I might have triggered a "Import Apps", but didn't clean oxides config, so not sure if that would do anything.

I was getting it with every import previously. I did try removing one application and triggering it again and was no longer able to replicate.

As for silencing the preload error, that would require changing how things are called to no longer pass the environment variable to them. Feel free to open a PR that accomplishes that.

Already suspected, that the output is just passed along without any real control over it. Fixing the cause seems also a good idea. Not sure if I'll get to do that though. But should be reasonably doable.

You'd want to look at https://github.com/Eeems-Org/oxide/blob/master/applications/system-service/wlan.h primarily as that is where most of the constant checking of things happens.

It seems though, that there was a bugreport submitted. I'm not sure what caused it. Seemed to be early on when testing the new build but I'm not sure exactly when.

I'm not seeing anything that came in. It might have just been finally uploading a previous instance.

* Add sentry * Change build location * Also load libld * Get building * Setup sentry contexts * Fix #239 and add more debug code * Fix #239 * Add more debug code * Add more logging and switch to shared library * Switch to dynamic library * Update package * Fix release build * Fix building with/without feature flag * Start getting package build working * Begin adding liboxide * Begin moving shared code to a shared library * Merge device settings and dbus settings into liboxide * Move more into liboxide * Switch settings handling to liboxide * Change namespace * Add shared setings * Add shared settings access to rot * Add some documentation about telemetry * Flush out FAQ more * Fix machine-id handling * Remove unused include * Fix load order for running on device * Fix compile * Error if deprecated APIs are used in the shared lib * More build fixes * Switch to Type=dbus * Start flushing out sentry API more * Make adding settings easier and fix wifi loading * Fix dynamic property loading * Make telemetry opt-in and update docs * Fix settings file creation * Fix string comparison * Fix telemetry state handling * Always set initial transaction * Minor cleanup * Fix #243 * Fix #242 * Move signalhandler to liboxide * Move eventfilter into liboxide * Add opt-in screen * Start trying to fix group logic * Fix group handling * Move logic into library instead of header * Upgrade sentry library * Add more spans to sleep handling * Get spans working * Fix application recording * Switch to using EPFrameBuffer for save/recall * Add more analytics * Move application usage to it's own opt-in flag * Update package * Fix #244 * Fix #246 * Add analytics around update timings * Add application name as tag * Fix corrupt name * Add liboxide package * Update package * Upgrade sentry * Fix bug with qt type not being declared * Include all symlinks in install * Fix the right package * Fix crash on first launch * Default telemetry off on first launch

Eeems added the bug Something isn't working label Jan 13, 2022

Eeems mentioned this issue Jan 13, 2022

QtPrivate::RefCount::deref #242

Closed

Eeems added a commit that referenced this issue Jan 16, 2022

Fix #243

fb5f936

Eeems linked a pull request Jan 16, 2022 that will close this issue

Add crash reporting #240

Merged

Eeems mentioned this issue Jan 16, 2022

Add crash reporting #240

Merged

Eeems added this to the v2.4 milestone Jan 16, 2022

Eeems closed this as completed in #240 May 15, 2022

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

msort_with_tmp.part.0 #243

msort_with_tmp.part.0 #243

eeems-sentry bot commented Jan 13, 2022

Eeems commented Jan 13, 2022

LinusCDE commented Jan 13, 2022

Eeems commented Jan 13, 2022

eeems-sentry bot commented Jan 13, 2022

Eeems commented Jan 16, 2022

LinusCDE commented Jan 17, 2022

LinusCDE commented Jan 17, 2022

Eeems commented Jan 17, 2022

LinusCDE commented Jan 17, 2022

Eeems commented Jan 17, 2022

msort_with_tmp.part.0 #243

msort_with_tmp.part.0 #243

Comments

eeems-sentry bot commented Jan 13, 2022

Eeems commented Jan 13, 2022

LinusCDE commented Jan 13, 2022

Eeems commented Jan 13, 2022

eeems-sentry bot commented Jan 13, 2022

Eeems commented Jan 16, 2022

LinusCDE commented Jan 17, 2022

LinusCDE commented Jan 17, 2022

Eeems commented Jan 17, 2022

LinusCDE commented Jan 17, 2022

Eeems commented Jan 17, 2022