If the variables that are wrong are locals in a function, we should be able to figure out what went wrong from just the .o file containing the "bad" debug info, and the preprocessed output of the source file. If they are globally visible entities, then the linker could have mucked with them, and we would need the whole project. Anyway, if it is the former case, then its certainly worth filing a bug.
Jim > On Jan 13, 2016, at 11:01 AM, Jens Alfke <[email protected]> wrote: > > I just had a weird problem in the debugger where it was showing nonsensical > values for some variables. I was able to convince myself that the actual > variables had correct values, by adding a printf call: > > body is at 0x7fff5fbfd090 > (lldb) p &body > (cbforest::slice *) $0 = 0x00007fff5fbfcad0 > > Turning off Address Sanitization made the problem go away. Looks as though it > was causing incorrect info in the dSYM file. > > (This is a 64-bit Mac OS target, a mix of Obj-C and C++. The function where > the problem occurs is C++. The variables having the trouble all appear to be > small structs. Xcode 7.2 7C68.) > > I can file a bug report, but I don’t think it’ll help much since (a) this is > a pretty big project, ~50kloc, and (b) I’ve never seen this happen before so > it’s probably specific to some particular thing I’m doing and would be > difficult to isolate. > > Has anyone else seen this? > > —Jens > _______________________________________________ > 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/jingham%40apple.com > > This email sent to [email protected] _______________________________________________ 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]
