@spotup
Quote:
Kas1e, in that case, it didn't workfor me, it crashed on the breakpoint.
In which situation it didn't work for you ? It those ones which i describe, or when you do something else ? And what breakpoint, when, where ?:)
Quote:
I didn't know if it was supposed to do that or not, as you can click the attach to gdb button.
If attach button present, then by idea you can attach to crashed process, but not in all the cases (for example if crash heavy one, and overwrite important areas of kernel, or anything which can cause problems for gdb itself).
As i understand for you gdb crashes, when you have crash in some programm, then you have GR, press "attach GDB" button, and then GDB crashes ? If so, then what programm is it , have it debug information or not, use it shared objects or not, and can you upload (or point out) on the programm which crash it.
Quote:
But when i did that i got no useful info either.
What kind of info you want to see in gdb ?:) Even if you will attach gdb to crashed process, then by idea you can only check (by yourself) what happens and where, but i do not know if it good idea to use that "attach gdb button" at all.
Better, its run gdb, and load your crashed programm to it, and then run it (inside of gdb).
To be honest, that "attach gdb" button on the crashed process, can help only to show something which GR already do not show for you (because GR already show registers, disassembly, memory dump, stacktrace). I.e. it mostlu statistic info for you, to know where crash is happens. But as crash _already_ happen, you can't debug it futher, or set breakpoitns, or anything - its already crashed.
So, you then run gdb, and run programm in the GDB, and then, you by youtself (as programmer) think where you need to put breakpoints, what you want to do, where and how execution should stops, so you can understand why it crashes.
Right now, i do not know for what need "attach gdb" button at all. My only "imho" currently its for taking more info than in GR (just for collect information).
Edited by kas1e on 2012/1/5 19:31:19