Bug 383630

Summary: Number of errors detected are changing over use of num_callers
Product: [Developer tools] valgrind Reporter: Santosh <pleasedolook>
Component: memcheckAssignee: Julian Seward <jseward>
Status: RESOLVED NOT A BUG    
Severity: major CC: philippe.waroquiers
Priority: NOR    
Version First Reported In: 3.13.0   
Target Milestone: ---   
Platform: Ubuntu   
OS: Linux   
Latest Commit: Version Fixed/Implemented In:
Sentry Crash Report:
Attachments: zipped file contains the xmls and logs corresponding to rung with num_caller set to 4 and 25

Description Santosh 2017-08-17 13:35:55 UTC
Created attachment 107330 [details]
zipped file contains the xmls and logs corresponding to rung with num_caller set to 4 and 25

While using memcheck and logging it into xml file, number of errors detected are varying based on num_caller value 
Example 
if num_caller = 4 in the argument the number of errors detected is 4
if num_caller = 25 in the argument the number of errors detected is 2

There is no change in the exe file and the error is recurring
Comment 1 Philippe Waroquiers 2017-08-24 08:36:51 UTC
It is not abnormal to have the number of leaks/leak loss records/errors
changing with --num-callers:
errors are 'glued' together according to the captured stack trace 
(as specified with --num-callers).
Suppression entries might also suppress a different number of errors
when --num-callers is changed.

If you believe there is really a problem, please re-open a bug, and
give all the details needed to reproduce the problem:
* a small standalone compilable c program
* the suppression files used
 * Also the xml output contains less detail than the textual output.
   So preferrably give the textual output, using also a high verbosity (-v -v)

I am closing the bug based on the above, but if needed, please re-open
Thanks