On Jun 17, 2015, at 1:41 PM, Doug Hill <[email protected]> wrote: > OK, now I try to go the manual route, using atos. I put the sym file in the > same directory as the executable and run: > > atos -o /Path/to/my/executable 0x00000001006341f8 > 0x00000001006341f8 (in Chartcube) > > Nothing. > > I try some other addresses for another thread and I get seemingly good > results: > > atos -o Chartcube 0x000000010047cbe0 > -[SChartDateFrequency setDefaults] (in Chartcube) + 32 > > So, it at least sort of works. > > Maybe atos is confused for some reason, so I’ll try lldb to lookup this > address:
This is probably because apps and frameworks are randomly relocated (slid) at app launch time. You need to find the value for the atos -l or -s options (these are two ways of supplying the same information). The load address and slide of each of the images in a process are listed at the end of the crash report, along with the build-UUID of each one, so you can find the value you need to pass to atos. (As I understand it, this is what Xcode and/or lldb are doing with the crash reports you give them as well. They also use Spotlight to find the correct dsyms file based on the build-UUID, which is what often fails for me.) Anyway, when Xcode/lldb fails, this is how you can do it “by hand”. You can also grub around at an even lower level using dwarfdump --lookup, if you really want to. > Looking at the address in the crash report I see: > > 16 Chartcube 0x0000000100634488 0x1000ec000 + 5538952 > 17 Chartcube 0x0000000100633fd8 0x1000ec000 + 5537752 > > Those are some suspiciously huge offsets which leads me to believe the stack > might be borked, but who knows. That’s also a possibility... _______________________________________________ Do not post admin requests to the list. They will be ignored. Xcode-users mailing list ([email protected]) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/xcode-users/archive%40mail-archive.com This email sent to [email protected]
