|
[Valgrind-developers] Inconsistent Exit codes and error reporting
for failures in replaced functions
From: Matthias S. <zz...@ge...> - 2015-07-30 20:49:48
|
Hi! I think the error reporting out of replaced functions is a bit inconsistent. 1. They use different exit-codes in fatal cases: * vg_replace_malloc.c: malloc dies with exitcode of 1. * vg_replace_malloc.c: panic dies with exitcode of 99. * vg_replace_strmem.c: dies with an exitcode of 127. 2. Only VALGRIND_PRINTF_BACKTRACE is used so in xml <clientmsg> will appear instead of <error>. Is it possible to create an <error> via client requests? Regards Matthias |
|
From: Florian K. <fl...@ei...> - 2015-08-09 20:37:06
|
On 30.07.2015 22:49, Matthias Schwarzott wrote:
> Hi!
>
> I think the error reporting out of replaced functions is a bit inconsistent.
>
> 1. They use different exit-codes in fatal cases:
> * vg_replace_malloc.c: malloc dies with exitcode of 1.
> * vg_replace_malloc.c: panic dies with exitcode of 99.
> * vg_replace_strmem.c: dies with an exitcode of 127.
>
Fixed in r15515.
> 2. Only VALGRIND_PRINTF_BACKTRACE is used so in xml <clientmsg> will
> appear instead of <error>.
>
Parse error.. Perhaps file a bug report with an example. Otherwise, this
will fall through the cracks.
> Is it possible to create an <error> via client requests?
>
I don't think this is possible today but would not rule it out in
general (in the future).
Florian
|