Front page | perl.perl5.porters |
Postings from July 2017
On 06/29/2017 01:29 AM, Dave Mitchell wrote: > On Tue, Jun 27, 2017 at 01:52:19PM -0600, Karl Williamson wrote: >> # Failed test 29 - croak: cachegrind output format at porting/bench.t line >> 349 >> # got 'Error: call::sub::empty/A empty/short loop gave return status >> 0x7f00\nwith the following output\n:valgrind: Boz: command not found\n\n' >> # expected /(?^:Error: while executing call::sub::empty/A empty/short >> loop:\nunexpected code or cachegrind output:\n)/ > > Hmm, strange. It's probably harmless, and I can easily fix it by expanding > what the test expects the failing output to match, but I don't quite > understand why it's happening, and would prefer to understand rather than > just brush it under the carpet. > > I can't reproduce the failure at all. Perhaps it depends on the shell? > If you can reproduce it, can you give me the output of the following: > > $ ./perl -Ilib Porting/bench.pl --grindargs=Boz --tests=call::sub::empty --debug ./perl=A ./perl=B > > for a run where the output contains > > Error: call::sub::empty/A empty/short loop gave return status .... > > rather than > > Error: while executing call::sub::empty/A empty/short loop: ... > > I just got it again: # Failed test 29 - croak: cachegrind output format at porting/bench.t line 349 # got 'Error: call::sub::empty/A empty/short loop gave return status 0x7f0 0\nwith the following output\n:valgrind: Boz: command not found\n\n' # expected /(?^:Error: while executing call::sub::empty/A empty/short loop:\nun expected code or cachegrind output:\n)/ but this is only the 2nd or 3rd time in hundreds of runs. Is there a way to change the default on this so when it does crop up again we'll have the data you need?Thread Previous | Thread Next