This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
[PING^2] Re: [PATCH,contrib] Add option to patch tester to avoid caching baseline test results.
- From: Dave Korn <dave dot korn dot cygwin at gmail dot com>
- To: GCC Patches <gcc-patches at gcc dot gnu dot org>
- Date: Sat, 18 Dec 2010 12:04:26 +0000
- Subject: [PING^2] Re: [PATCH,contrib] Add option to patch tester to avoid caching baseline test results.
- References: <4CE15509.4090602@gmail.com> <4CF725C1.5080207@gmail.com>
On 02/12/2010 04:51, Dave Korn wrote:
> On 15/11/2010 15:43, Dave Korn wrote:
>
>> I discovered a problem with the patch_tester.sh script in the contrib/
>> directory. By design, the script caches the results of the clean test run it
>> performs against any given branch/revision combination, which saves a lot of
>> time when testing a whole bunch of patches. However, it can give false
>> results in the case when testing one patch against a variety of different
>> configure options. Different configuration options can alter the baseline set
>> of FAILs that the testsuite will produce, meaning that the cached results of
>> the clean test run may falsely fail to match the results from the test run
>> done against the patched build if they originally came from a run with
>> different configure options - the particular case that bit me was compiling
>> with --disable-shared, which causes a load of mudflap FAILs to appear(*).
>>
>> This patch addresses that problem by adding a command-line option that
>> prevents the use of cached results, forcing the baseline results to be
>> regenerated anew for each patch that is tested.
>>
>> contrib/ChangeLog:
>>
>> * patch_tester.sh (nopristinecache): New shell var, set according
>> to presence or absence of new -nopristinecache option.
>> (usage): Document new option.
>> (bootntest_pristine): Implement it.
>>
>> OK?
>
> Ping?
Re-ping?
cheers,
DaveK