On Fri, 2025-02-07 at 09:43 +0100, Tomaz Canabrava wrote: > This is a crash I'm experiencing as well, > I tried to fix but the backtrace tells me that it's a iterator > invalidation, meaning that it's probably it happens somewhere else on > a thread call.
Note that you can record the application execution with `rr record …`, and then use launch a gdb with the record with `rr replay`, and use tricks like setting watchpoints in the past. Like, you `run` the execution till the bug happens, and then you can set a watchpoint to the invalidated pointer, and use `rc` (reverse-execute) gdb command to execute the application in the opposite direction. So when watchpoint triggered, you will know what accessed it the last time.
