Seeing this (100% repeatable) in web project when attempting to call MySQLCommunityServer’s ‘SQLExecute’ method.
I tried previous versions of the Xojo-provided MySQL plugin, no difference. I tried clearing caches, no difference.
Crash report is always identical. This is on MacOS 14.7, Intel.
The same project works fine in 2024r3, so I think this is a regression (or possibly a deprecation I missed??)
Crash report references threads, but it’s a web project so I’m not sure if it’s really a thread-related issue or just has “thread noise” in the crash log because Web Projects use threads.
Crash report full text attached below:
crash report.txt.zip (13.4 KB)
1 Like
Following up on this item.
This is a 100% repeatable crash in 2024r3.1
Same code works without issue in 2024r3; fails 100% of the time in 2024r3.1.
Crash report attached in initial post.
It would be hard to for me to submit this, even as a private case as I’d need to setup an additional database server and schema at AWS just for the case. (due to contractual obligations)
Hoping there’s a large enough “breadcrumb” in the crash report to point at the likely culprit.
I have “feels”/“spidey sense” the issue may be related to how threads are used to manage sessions in Web apps.
FYI. @Ricardo_Cruz @William_Yu
Hi @Anthony_Dellos, even if you can’t share the project, could you please create the Issue so we can track it?
Unless something exceptional happens, we only include bug fixes in point releases. If it worked fine in 2024r3, it sounds like a regression.
@Ricardo_Cruz - thanks for following up. Truly appreciated.
Issue is created. It’s here:
https://tracker.xojo.com/xojoinc/xojo/-/issues/77712
I can’t attach the real project, but if you need me to try and make a simpler example and attach it, I’ll make an effort to do so.
(You’d need to provide your own MySQL server though or coordinate with me to setup a separate database just for Xojo engineer’s to test with)
Thanks a lot @Anthony_Dellos, we can reproduce the issue.
1 Like
@Ricardo_Cruz - appreciate the effort and glad you were able to reproduce.
Any workaround? 3.1 has some nice improvements added by… checks release notes… YOU
If not, I’ll stick with 3.0 for now and wait for the fix.
Again, many thanks for helping track this down.
1 Like
In your case, yes, use 3.0 for the moment. There is no workaround yet.
@Ricardo_Cruz - thanks. I will do. Appreciate all that you do.