Bug 400624 - Frequent (data loss) crashes while saving with recent vector layer changes
Summary: Frequent (data loss) crashes while saving with recent vector layer changes
Status: RESOLVED FIXED
Alias: None
Product: krita
Classification: Applications
Component: Tools/Vector (show other bugs)
Version: 4.1.5
Platform: Other Microsoft Windows
: NOR crash
Target Milestone: ---
Assignee: Krita Bugs
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2018-11-03 15:09 UTC by jameswilddev
Modified: 2019-05-03 11:52 UTC (History)
2 users (show)

See Also:
Latest Commit:
Version Fixed In:


Attachments
kritacrash.log (173.15 KB, text/plain)
2018-11-03 15:09 UTC, jameswilddev
Details

Note You need to log in before you can comment on or make changes to this bug.
Description jameswilddev 2018-11-03 15:09:03 UTC
Created attachment 116064 [details]
kritacrash.log

SUMMARY
I'm seeing frequent crashes when saving immediately after making changes (changing a shape's fill color is sufficient) to a vector layer.

STEPS TO REPRODUCE
1. Open a document with a vector layer.
2. Change a fill color.
3. Save.

OBSERVED RESULT
Rougly 50% chance of complete Krita crash losing all unsaved changes.

EXPECTED RESULT
Document saved and Krita does not crash.

SOFTWARE VERSIONS
(available in About System)
Krita
  Version: 4.1.5 (git 1be40f6)
  NOTE: also observed in stable 4.1.5

Qt
  Version (compiled): 5.9.3
  Version (loaded): 5.9.3

OS Information
  Build ABI: x86_64-little_endian-llp64
  Build CPU: x86_64
  CPU: x86_64
  Kernel Type: winnt
  Kernel Version: 10.0.17134
  Pretty Productname: Windows 10 (10.0)
  Product Type: windows
  Product Version: 10

OpenGL Info
  **OpenGL not initialized**

ADDITIONAL INFORMATION
A couple of times this has happened I've been certain I'd saved along the way but lost ~20 minutes of work.  Might be imagining it.
Comment 1 wolthera 2018-11-16 14:58:25 UTC
There's a crashlog so I'll mark this as confirmed, but no idea yet where the crash is :)
Comment 2 Halla Rempt 2018-11-27 08:18:07 UTC
I cannot reproduce it -- it might need a fairly complex test file, but the backtrace is fairly clear, though it won't be easy to fix. It's also code that hasn't changed for ages. Do you perhaps have a test file with which it's easier to reproduce?
Comment 3 jameswilddev 2018-11-28 19:11:32 UTC
I can't reproduce this either now (with the same file) which is weird because I've otherwise not had any issues with Krita and a memory test passes.  My suspicion was some sort of threading bug with the save system (perhaps a lock which wasn't being respected by the vector tools, but was being respected by the raster tools).
Comment 4 Halla Rempt 2018-11-29 09:11:55 UTC
I suspect that it's indeed a timing issue, saving happens on a shallow copy in a background thread, and it looks like the copy is made while the shape is still being altered, or something like that.
Comment 5 Halla Rempt 2019-05-03 11:52:09 UTC
This should have been fixed in f984eab0aa72495867b67c51a68531093d086729