fixed I'm getting a lot or crashes with PI 1.9.0 on macOS

robyx said:
Do you remember how many frames did you add to the wbpp session? And at which step it crashed?

If you have a wbpp log file or the crash log it may be helpful to investigate, thanks!
Click to expand...
3589 x10s frames but here's a very strange and interesting update. Just before downgrading for the 4th time to 1.8.9-3 I decided to try again but enabled Local Normalization in WBPP, something I haven't enabled in months, and it actually completed the stack!! This was the only difference between the last attempt which failed during image integration. Unfortunately I did not save the log file.
 
robyx said:
new version 1.9.2 is available now, you can download it form the website, and we worked hard to solve the problems what were causing frequent crashes, especially on MacOS.

We expect that now no more crashes will occur, when you have time please update and feel free to confirm.

Robyx
Click to expand...
Version 1.9.2 is a definite improvement and completes tasks more reliably. However, I’m still encountering fatal crashes that remain somewhat unpredictable. For example, WBPP might crash during one run and complete successfully on another, even with the same data and settings.

That said, I’ve managed to identify a crash that occurs consistently. It reproduces every time on two different Macs (one Intel-based and the other ARM-based, using an M3 processor). I’ve uploaded a dataset of 168 subs along with error logs from two crashes that occurred when running them through SubframeSelector in 'measure' mode.

Here a link to the data in dropbox.
 
lgibbons said:
Version 1.9.2 is a definite improvement and completes tasks more reliably. However, I’m still encountering fatal crashes that remain somewhat unpredictable. For example, WBPP might crash during one run and complete successfully on another, even with the same data and settings.

That said, I’ve managed to identify a crash that occurs consistently. It reproduces every time on two different Macs (one Intel-based and the other ARM-based, using an M3 processor). I’ve uploaded a dataset of 168 subs along with error logs from two crashes that occurred when running them through SubframeSelector in 'measure' mode.

Here a link to the data in dropbox.
Click to expand...
Currently running WBPP with your data and it made it through the initial measurement phase without throwing up any of the errors from the log file you posted.

M4 Mac Mini 32GB
Sequoia 15.2

edit: I've stacked it successfully, I think. This is a mono image right?
 

Attachments

  • Screenshot 2024-12-29 at 11.22.55 AM.jpg
    Screenshot 2024-12-29 at 11.22.55 AM.jpg
    289.5 KB · Views: 33
  • Screenshot 2024-12-29 at 11.42.32 AM.jpg
    Screenshot 2024-12-29 at 11.42.32 AM.jpg
    219 KB · Views: 31
  • ProcessLogger.txt
    ProcessLogger.txt
    7 KB · Views: 30
Last edited:
clipper99 said:
Currently running WBPP with your data and it made it through the initial measurement phase without throwing up any of the errors from the log file you posted.

M4 Mac Mini 32GB
Sequoia 15.2

edit: I've stacked it successfully, I think. This is a mono image right?
Click to expand...
Great. Were you able too discover anything by looking at the code where my crash logs indicate a problem occurred?
 
lgibbons said:
Version 1.9.2 is a definite improvement and completes tasks more reliably. However, I’m still encountering fatal crashes that remain somewhat unpredictable. For example, WBPP might crash during one run and complete successfully on another, even with the same data and settings.

That said, I’ve managed to identify a crash that occurs consistently. It reproduces every time on two different Macs (one Intel-based and the other ARM-based, using an M3 processor). I’ve uploaded a dataset of 168 subs along with error logs from two crashes that occurred when running them through SubframeSelector in 'measure' mode.

Here a link to the data in dropbox.
Click to expand...
Hi,

did you cancel WBPP while it was measuring frames? or using the mouse and clicking somewhere? I see there is a mouse button event involved in the crash stack trace...
 
i have just reported the same issue (should have checked the other threads first i guess) but for me on OSX 13.7.1 on all 1.9 versions i get regular intermittent crashes to desktop when starting scripts. I restart, the plugins and scripts are no longer in the menu's i have to reset all updates and reinstall. starxterminator, imagesolver, graxpert etc.
 
i think osx13 is not explicitly supported anymore, so you might be on your own with that. the official release notes for 1.9 says it might work with osx12 and osx13 but is not guaranteed to work.
 
robyx said:
Hi,

did you cancel WBPP while it was measuring frames? or using the mouse and clicking somewhere? I see there is a mouse button event involved in the crash stack trace...
Click to expand...
No, not as far as I remember. I spent hours testing this and ran numerous trials—some while I was in the room and others while I let it run unattended. Based on that, I highly doubt mouse movement plays a significant role overall.
 
I've just installed 1.9.2 build 1633 and ran SubframeSelector on that dataset I provided earlier. PI crashed immediately. I've uploaded the PI log, the crash dumps, and a video I took holding my phone up to the screen showing the crash happening in real time. Hope this helps.

The files are in this dropbox folder: LINK
 
You must log in or register to reply here.
Back
Top