Xcode 5.1 no longer officially supports GDB, instead defaulting only to LLDB. The problem with LLDB is that it shows no useful debug information on app crashes. Furthermore, all Exception Breakpoints simply break on main.m. This makes debugging ridiculously tedious. I read here on SO that this is a common problem with LLDB and that GDB does a better job.
How do I enable GDB for xcode 5.1?
There is no way to use gdb with more recent Xcodes.
I don't know what you mean by "it shows no useful debug information on app crashes." Probably best to file a bug with bugreporter.apple.com with some more detail, there may be a way to get lldb to work correctly for you.
I am also not sure what you are seeing when you say "all Exception Breakpoints break on main.m". If you go to the lldb console and do:
(lldb) break list
are the breakpoints actually on main.m?
One thing to be a little careful about with the Xcode 5 Debugging UI (maybe this started with 4, I can't remember.) When your program stops due to a crash or exception in a stack frame that starts with frames that have no debug information, Xcode will actually select the first frame up the stack that has debug information. This is to avoid showing people screens full of disassembly, which some folks find frightening... So the source frame will show, say, main.m, though the actual bottom-most frame is something else.
Xcode also has a "stack compression" feature that will hide "uninteresting" frames. That can also make this kind of stop confusing - though it will generally show the bottom-most frame you might miss that and only see your source frame. The stack compression can be turned off if you don't like it.
Make sure that is not what you are seeing.
Related
In order to triage an issue I am trying to set a symbolic breakpoint in Xcode on a c++ function, but it never hits (I tried both live device and simulator, using a debug build).
I am using the function name as listed in a symbolicated stack trace. For example:
WebKit::WebLoaderStrategy::loadResourceSynchronously+
Is there any trick to get this to work?
While I think it is possible to build WebKit I don't think this would be of use since I can't rebuild iOS with my version of WebKit. But I assumed at least the symbols would match up with that is in stack traces of crash logs (and also with the source of WebKit on Git).
Update, if I just enter in "loadResourceSynchronously" it automatically found the following symbols: (cut and paste doesn't work for some reason, so will use an image)
This seems to indicate the symbols are understood by Xcode. However I never see any of them being called. Based on the stack traces I am seeing for crashes I am pretty sure they should be called.
It used to work fine but now I get this message when loading a certain scene and then the APP crashes.
2012-12-01 18:33:55.104 AppName[9561:707] cocos2d: CCSpriteFrameCache: Trying to use file 'art1-hd.png' as texture
I have no idea on where to start debugging this. Is there a way to get a more precise error message to understand why this happens?
Btw, I have changed my code signing identity to developer and XCode doesn't yet give me any trace of it on the console other than the plain message above. I checked as well the device logs on the Xcode-Organizer section but no trace.
I whish there was something like in Java + Eclipse, where I could get a precise STACKTRACE of where the crash happened.
Any suggestion on how to get a precise stacktrace (E.g. with function name generating the crash)?
PS: I have set strip debug symbols to "NO" in DEBUG, and to "YES" in Release, but I am builing on my own test iPod and hence I think it is automatically build for "Debug" (the section says build for running and testing which is confusing as the project settings allows only debug and release). But I think the problem is not here.. but in the issues above before the PS:
EDIT: If I look into the debug navigator I don't see any stacktrace of methods, as I want, but just threads.. I need to figure out how to use the global exceptions.
I have no idea on where to start debugging this.
You have the cocos2d source code. I would start there.
Search for "trying to use file" in CCSpriteFrameCache to understand what the cause of this message is (which I agree is confusing, why shouldn't you use that file?).
As for the stack trace, it's available as soon as you add a global exception breakpoint. If the breakpoint triggers in OpenAL code make sure to set it to catch only Objective-C exceptions.
My app was crashing only when not running using XCode debugger. It was hard to track because I can't debug but I finally figured it out. It was because of calling release on some object not owned by me. Before I corrected it I searched and found 2 related questions here (links below)
iOS App Crashes when running by itself on device, does not crash when running through Xcode using debugger, or in simulator
iPhone crash only when device not connected to xcode, how to understand the crash log?
None of the above question has answered why no crash when running via debugger.So my question is why it happens ? I know reasons for debug/release specific crashes but this is crazy. Is it just by chance although it happened more than 10 times.
What you describe is not atypical of obscure memory-related bugs. You might also want to use debug-malloc at such times. Although that is not guaranteed to find everything. The reason (and it's been happening probably as long as there've been source-level debuggers) is that memory is laid out at least somewhat differently in debuggable code, and when running under the debugger. So the error results in a different piece of memory being (harmlessly) corrupted when under the debugger. When not under the debugger the location corrupted is actually something that your code cares about, and it crashes.
The same could happen in reverse, but you'd never know - if it crashes when run debuggable, you'd find it before switching to running outside the debugging environment.
Reiterating #jyoung's answer since I didn't see it the first time I glanced through:
Try running with Zombie Objects turned off.
In debug mode if you have it turned on it is handling memory allocation differently. Try running it without.
Go to Edit Scheme... > Run > Diagnostics. Then make sure zombie objects is turned off:
Then run through your code path again.
I had this same issue while working on a project modularised with Xcode Frameworks. Even after removing all the logic in AppDelegate and only returning true inside application:didFinishLaunchingWithOptions, I was still getting the crash. Then I switched to my project settings, in the Frameworks, Libraries, and Embedded Content section and changed the embed option for the frameworks I added to Embed & Sign. This was what fixed the issue for me. I hope someone finds this helpful.
I was having this problem as well and was fortunate to figure out the cause quickly, hopefully by posting here I can save someone else some wasted time. To clarify, my app would run with no issues when launched directly from XCode, but would crash immediately when launched manually on the iPad.
The app in question is written in Obj-C but relies on some 3rd party code written in Swift. The Swift code is included in the app as an embedded framework. I had to set "Embedded Content Contains Swift Code" to Yes in the Build Settings for the app (under Build Options), then the problem went away.
I experienced this symptom when I made a NSString, sent a UTF8String from it to another object, and assigned it to a char pointer. Well, it turns out that I forgot to retain the original NSString, which wouldn't have mattered anyway, since I also failed to realize that the UTF8String method (which is presumably an object that gives access to the pointer itself) operates in the autorelease pool. That is, retaining the NSString itself did not fix the problem.
I suppose this appeared to work just fine when attached under the debugger only because I had zombies enabled, so the pointer I had was still valid. I should see if this is the reason it worked; if so, this is a good reason to test with and without NSZombie enabled.
At any rate, this was probably poor design to begin with, and a pretty obvious newbie memory management mistake once I found it. Luckily the console in the Organizer window gave me some hints on where to start looking, and debugging ultimately showed me where my pointer's value was changing. Hope this helps anyone who finds the way here.
I had this issue when accessing SQLite databases from outside the [[NSBundle mainBundle] resourcePath] directory, which caused iCloud errors.
I discovered the error only by installing a Console app onto my iPhone which logged the errors.
Once I accessed the databases from the correct directory, the errors disappeared and the application booted correctly.
When I crash (for reasons I understand; that's not the problem) when I try to do something Cocoa isn't okay with, such as calling a method that doesn't exist or attempting to insert nil into a set, the debugger shows the stack from main() to __pthread_kill, without any of the frames that were present when the actual crashing code ran. There is a frame (9th from main) called objc_exception_rethrow. This leads me to believe that Cocoa Touch is trying to do something or other to recover all exceptions and die gracefully or something. However, it is very irritating when debugging to not have the ability to actually use Xcode4's debugging tools to investigate the calling stack frames, or even see where in my code I crashed.
Is there some way to make the objc_exception_rethrow behavior not happen, and just crash as soon as an exception is raised? Perhaps there's a debug setting that makes it crash earlier (at the right time)? (I haven't messed with any of the build settings in this project yet.)
I don't know any Xcode setting that could disable re-throwing exceptions. To my knowledge they are re-thrown by the runtime. You could try running the app without the debugger attached and let it crash. The crash report should contain a section "Last Exception Backtrace" which will give you exactly what you need in this case.
I found the answer: set a breakpoint on Obj-C Exceptions. It will go to the debugger when objc_exception_throw is hit, which is good. Unfortunately, this happens before the exception is printed, but we can make that happen anyway (most of the time) by setting the breakpoint's action to be (Debugger Action) po *(id *)($ebp + 8).
I'm trying to optimize the memory usage of an iOS app, and I'd like to see what the app's total memory usage is at specific points in the code. I was thinking I should be able to set breakpoints, profile the app with Activity Monitor, and just look at the memory use when each breakpoint catches. But when I run Instruments, it seems breakpoints no longer stop execution, so it's hard to know exactly when memory usage is changing.
Is it possible to use breakpoints and Instruments at the same time? If not, is there a clever way of writing some code to insert a marker into the Instruments timeline when specific events occur?
I also ran into this issue today, and after a bit of searching I found this solution. Text below is a quote from the post:
Breakpoints Do Not Break. Instruments utilizes debug information from
your debug build, however, it does not stop at break points you set.
This is because while you are loading your application into
Instruments from the menu in XCode, Instruments simply uses the path
of the current executable as its start path and loads it externally
from XCode. The menu in XCode is really there as a convenience
mechanism. This is not a big deal as you can always run again in Debug
mode after your instruments session should you want your application
to break. It’s just something to make a note of.
NSLog Statements Do Not Show In The Debugger Console. If you want to
see your NSLog statements, you will need to load the system Console
application (/Applications/Utilities/Console).
Reference: http://www.cimgf.com/2008/04/02/cocoa-tutorial-fixing-memory-leaks-with-instruments/
Well, you aren't running under control of the debugger.
One approach might be to add alerts at the key points, and take a heapshot then (manually).
Or there may be some dtrace wizardry.