search

houssamham / plcrashreporter

Automatically exported from code.google.com/p/plcrashreporter
Other
1 stars 0 forks source link

armv7s crash report from release 1.1rc1 can't be symbolicated. #52

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
What steps will reproduce the problem?
1. I just updated my app from the old 1.1 beta to the new 1.1rc1.
2. I added a line such as NSAssert(0, @"Some message");
3. I ran the app and hit the assertion. The app crashed and a crash report was 
generated.

What is the expected output? What do you see instead?

The crash report generated from plcrashreporter can't be symbolicated. I loaded 
the crash report in the Xcode organizer and I tried to use the symbolicatecrash 
script. Neither resulted in a symbolicated crash log.

What version of the product are you using? On what operating system?

This was with a copy of the CrashReport.framework from release 1.1rc1 (binary 
release).

This was run on an iPhone 5 running iOS 6.0.1. I using Xcode 4.5.2 on OS X 
10.7.5.

Please provide any additional information below.

Running the same app on an iPod touch (4th gen) running iOS 6.0.1 results in a 
crash report that can be symbolicated just fine.

So the issue only seems to occur with an armv7s device.

Original issue reported on code.google.com by rma...@gmail.com on 28 Nov 2012 at 7:28

GoogleCodeExporter commented 9 years ago 
I should point out that i generate the crash log within the iOS app and then 
send the report via email. I use the following code to create the crash report:

if ([[PLCrashReporter sharedReporter] hasPendingCrashReport]) {
        NSData *data = [[PLCrashReporter sharedReporter] loadPendingCrashReportData];
        if (data) {
            PLCrashReport *report = [[PLCrashReport alloc] initWithData:data error:nil];
            NSString *crash = [PLCrashReportTextFormatter stringValueForCrashReport:report withTextFormat:PLCrashReportTextFormatiOS];

The "crash" text is added to an email as an attachment. This all works great on 
an armv7 device but with an armv7s device, the resulting report can't be 
symbolicated.

Original comment by rma...@gmail.com on 28 Nov 2012 at 7:57

GoogleCodeExporter commented 9 years ago 
Is anyone else having this problem or is it working for everyone but me? I'd 
like to know if I did something wrong or there really is an issue with this 
update. Thanks.

Original comment by rma...@gmail.com on 30 Nov 2012 at 6:28

GoogleCodeExporter commented 9 years ago 
Could be more specific about what the symbolication error looks like? Is there 
any major difference in the generated crash reports? I am not aware of a 
general issue and we have millions of crash reports successfully created and 
symbolicated, with armv7 and armv7s. So without further details, it is nearly 
impossible to say what the problem is.

Did you check if the dSYM is the correct one? Do "dwarfdump --uuid my.app.dSYM" 
and compare the UUID with the one in the first line under "Binary Images" in 
the crash report.

Original comment by linde.andreas on 30 Nov 2012 at 9:23

GoogleCodeExporter commented 9 years ago 
After some digging, I have found a clue. When I get the crash report from my 
app (running on an armv7s device), the Binary Images section of the report 
shows arm-unknown after every library/framework. The exact same app (adhoc 
build) run on a device with an armv7, properly shows armv7 after every 
library/framework in the Binary Images section.

Looking at the source code, this is coming from the PLCrashReportTextFormatter 
class which I'm using in the app like shown above in my earlier comment.

Does anyone have any pointers to resolve this?

Thanks

Original comment by rma...@gmail.com on 30 Nov 2012 at 11:30

GoogleCodeExporter commented 9 years ago 
I found the bug. There is a missing "break" statement in 
PLCrashReportTextFormatter.m at line 355 (from 1.1-rc1 tag). I'll open a new 
bug specific to this issue.

Original comment by rma...@gmail.com on 30 Nov 2012 at 11:43

GoogleCodeExporter commented 9 years ago 
This issue can be closed.

Original comment by rma...@gmail.com on 30 Nov 2012 at 11:44

GoogleCodeExporter commented 9 years ago

Original comment by landon.j.fuller@gmail.com on 1 Dec 2012 at 2:24