Front page | perl.perl5.porters |
Postings from June 2007
Steve Hay wrote:
> demerphq wrote:
>> On 5/2/07, Nicholas Clark <nick@ccl4.org> wrote:
>>> On Wed, May 02, 2007 at 12:57:33PM +0100, Steve Hay wrote:
>>> > Nicholas Clark wrote:
>>>
>>> > Oh. And I assume that smoking PERL_OLD_COPY_ON_WRITE wouldn't be
>>> useful
>>> > either (judging by its name)?
>>>
>>> I'm not convinced. Right now I suspect that it will fail an assertion
>>> somewhere in the regexp code (not tried, and typing faster than
>>> thinking or
>>> trying, it seems - will take time to try)
>>>
>>> It's not un-useful. But it's not that useful either.
>>> And it's sort of nice if it doesn't break, because it validates that
>>> all the
>>> current "this is a write so you need to make a copy" logic is working.
>>
>> I agree it would be cool to know if it works...
>>
>> Especially as I'm likely to have been the person who broke it. :-)
>
> Okay, I'll look at setting up a smoke to test COW once a week instead of
> ithreads.
>
> I currently have VC6 on Mon-Thu, BCC on Fri, GCC on Sat and VC7 on Sun.
> I'd like to replace one of those VC6s with a VC8 (when all the VC8
> changes have gone into maint), and I'll then make one of the remaining
> three do COW instead of ithreads.
You may have seen that I've now slotted in a smoke to test with
-DPERL_OLD_COPY_ON_WRITE using VC6 on Wednesdays. (I arranged for it to
test that option instead of the uselargefiles option, rather than
instead of the ithreads option as I had originally suggested.)
And it's immediately found a problem in the Compress::Raw::Zlib tests:
http://www.nntp.perl.org/group/perl.daily-build.reports/2007/06/msg48735.html
Verbose output of the failure follows. Any ideas?
ok 1 - use Compress::Raw::Zlib;
ok 2
#
#Length 1
#
ok 3
ok 4
ok 5 - status is Z_OK
ok 6 - no msg
ok 7 - total_in == 0
ok 8 - total_out == 0
ok 9 - status is Z_OK
ok 10 - flush returned Z_OK
ok 11 - no msg
ok 12 - length total_in
ok 13 - length total_out
ok 14
ok 15
ok 16 - status is Z_OK
ok 17 - no msg
ok 18 - total_in == 0
ok 19 - total_out == 0
ok 20 - status is Z_STREAM_END
ok 21 - got expected output
ok 22 - no msg
ok 23 - length total_in ok
ok 24 - length total_out ok
#
#Length 2
#
ok 25
ok 26
ok 27 - status is Z_OK
ok 28 - no msg
ok 29 - total_in == 0
ok 30 - total_out == 0
ok 31 - status is Z_OK
ok 32 - flush returned Z_OK
ok 33 - no msg
ok 34 - length total_in
ok 35 - length total_out
ok 36
ok 37
ok 38 - status is Z_OK
ok 39 - no msg
Compress::Raw::Zlib::Inflate::inflate input parameter cannot be
read-only when ConsumeInput is specified at
../ext/Compress/Raw/Zlib/t/07bufsize.t line 96.
ok 40 - total_in == 0
ok 41 - total_out == 0
# Looks like you planned 288 tests but only ran 41.
# Looks like your test died just after 41.
dubious
Test returned status 255 (wstat 65280, 0xff00)
DIED. FAILED tests 42-288
--
Thread Previous
|
Thread Next