PixInsight - V1.8.8-6 - Intermittent and unpredictable crashes

[dcarr]

Hi

PixInsight has been crashing on my Windows laptop since upgrading to v1.8.8-6. The crashes are intermittent and unpredictable, but in my case relate mainly (but not only) to the opening of image files. The symptoms do not occur when I revert to v1.8.8-5. I have repeatedly reverted to v1.8.8-5 to confirm that all is well with that version, then reinstalled v1.8.8-6 to see the problems reoccur.

The symptoms of each crash are: PixInsight reports a failure (see attachment) suggesting something may be in wrong in Qt, no other activities can be performed in the current PixInsight session, and the program cannot be shutdown using the close icon - forcing me to kill the session using the Windows Task Manager.

It is difficult for me to provide good diagnostic data because of the irregular and unpredictable nature of the errors, and because all I can do after a crash is to copy the error report from the Process Control and/or take screenshots – see attached.

Both example errors reported in the attachment relate to the opening of image files:

Example 1: Simple opening of a file using standard PixInsight menus. File opened, but the system then stopped working. PixInsight needed to be closed with TaskManager, and restarted.

Example 2: During a long-running script; failures occur often (only on v1.8.8-6), but at different locations in the script – i.e. not reliably repeatable, except the knowledge that an error will certainly occur at some time.

Again, none of these issues occur when I revert to v1.8.8-5.

Question: Can resource issues that build up, but do not manifest as errors, while running scripts create follow-on failures while working with standard PixInsight processes and menus (such as simply opening of files as in Example 1)?

I know you will rightfully ask for more data, but I feel I need to stay with the earlier version of PixInsight until I finish my current project.

Regards

Dean

 

[Juan Conejero]

Hi Dean,

None of these problems can be reproduced on any of our working and testing machines, on any platform. This does not happen under normal working conditions.

Besides the usual recommendations about installing the latest Windows updates, the latest version of your graphics drivers, checking for conflicting third-party utilities (bloatware), registry corruption, damaged system components, hardware issues, etc, I can only recommend a clean reinstall of your operating system. If this doesn't help or isn't possible, then I cannot help. These problems are originating from Qt, which is a proven stable and reliable library. If version 1.8.8-5 works for you, then you can stick with it unless you can use the latest version on a different machine or operating system.

Sorry for not being able to help you further on this problem, but I cannot reproduce it so it is out of my control.

 

[IanL]

Just scrolling back through previous bug reports to see if anyone else was having the same issue (Access violation: invalid memory read operation at address 0000000000000008) and found this - sorry for not spotting it earlier and starting a new thread (feel free to merge them if it helps), but I'm having exactly the same issue since upgrading to 1.8.8-6.

I understand the difficulty with debugging issue that can't be reproduced on demand, but it seems to have crept in from 1.8.8-5 to 1.8.8-6. Given that PI on Windows forcibly uninstalls previous versions now it doesn't seem possible to have two versions installed at once any more. Again I understand that it is probably the lesser of two evils to avoid conflicts between versions and preferences, but I'm now in the invidious position of having to save files and projects between every single step in the workflow as this bug strikes randomly three or four times a day.

 

[UrbainAstro]

Hello, I have a similar recurent crash since a week, I think it's the same. Take a look at the file attachement the ''backtrace informations''. I have this crash each time I use PixInsight, I'm really disapointed! To be honest it's the first time I see PI crashing like this, usually it's very robust and reliable.

 

[Juan Conejero]

As noted above, these issues are not reproducible and don't happen under normal working conditions. They are probably being caused by wrong/faulting graphics drivers, buggy system updates, system corruption, or some hardware/software errors or incompatibilities that we cannot recreate on our machines. See my previous post above for some recommendations.