I’m developing a custom allocator and a fatal exit I triggered via buggy code is triggering the raise() issue. Basically when it’s triggered it causes my entire login session to just suddenly quit on me, forcing me to re-login and re-open
my text editor, browser, etc.
My custom allocator project is here:
At first I thought it was valgrind not handling specific signals or tracee death or something, after running without gede or valgrind I found the issue to still be triggering. I removed the exit() call and saw the issue still trigger so I figure it must be signal related and raise(SIGINT) never used to trigger a full logout on previous manjaro installs so this must be a regression of some kind.
The docs don’t mention anything about logging out the user
It’s the responsibility of downstream to notify upstream if there’s an issue with the DE, not the user. The user doing it just creates confusion because it might not be an issue in upstream but rather an issue in downstream not updating some package. Can’t remember which vid it was but Brodie made a very good vid on the subject which convinced me to never report issues to upstream myself unless I happen to be using upstream packages directly.
And that is correct - I don’t know what desktop you are using - but you should probably try to reproduce using a clean Arch system - that can be done easily using a virtual machine.
This has been the worst “support” I have ever gotten, when I get round to it I will move from manjaro to another system because of your disgusting attempt to fob off your team’s responsibility to resolve issues in the packages you maintain. I will also never recommend manjaro again.
I don’t even know where or what you are doing - I believe you are shooting the messenger instead of the issue you have created
That would imply that it is in fact packages maintained by Manjaro Team - and that is very few - and it would imply it is known which package you are referring to.
If you know then you can find the packager
pamac info <pkgname>
Secondly - it is your code - why on earth would any member on Manjaro forum look at your source to help you troubleshoot it?
It’s in the title of the thread! Whatever package implements raise(), a standard function, should be looked at. Just because my own code is buggy doesn’t mean my session should quit when it triggers a signal that should terminate the process, not the user session
That is the fun with c - you are are charge - what ever happens in your code is on you.
I have been there - speculating myself nuts over an issue the stalled everything … until I found it … and it was always my own doing - I have blamed the system - I have blamed everybody but my self - in the end - it is always my self.
You’re not understanding the issue though, it is not my code that is causing the session to log out, it doesn’t even touch the system outside of semaphores, mapping memory and that one raise() call. All of which are wrapped to prevent my f**kups from reaching the system. I even have a dedicated function the prevents the use/reallocation/freeing of allocations that just aren’t in the buffers/mappings given. Even valgrind itself is not reporting any reads/writes to unmapped/unallocated memory.
Please review the forum rules and tutorials to help us help you.
To be straightforward, I have no idea what you’re talking about since I’m not a programmer. I imagine one would have the same issue no matter what distro or DE they’re using as neither should make a difference in that regard–except for the versions of software provided.
Keep in mind this is the official Manjaro support forum. There are numerous other websites specifically created for those seeking help with programming. I invite you to seek those out instead.