develooper Front page | perl.perl5.porters | Postings from September 2011

Re: Some preliminary cpan smoking reports

Thread Previous | Thread Next
Steffen Mueller
September 4, 2011 07:20
Re: Some preliminary cpan smoking reports
Message ID:
On 09/04/2011 11:54 AM, Nicholas Clark wrote:
> On Sat, Sep 03, 2011 at 08:23:08PM -0700, Reverend Chip wrote:
>> Information!  Wonderful, delicious facts!  Thank you muchly.
>> Now to dive into the specifics, which will also take time.
>> The number of modules that went from fail or unknown to pass seems odd,
>> doesn't it?
> I happened to look at Coro. The fail is because of Steffen's (intermediate)
> XS changes. The pass would be because the branch being smoked is either
> based on Steffen's final XS version, or from before Steffen's changes.
> Which is unfortunate. The test really needs to be between a branch, and the
> point on blead where the test diverged. Does Steffen's smoker script
> instead use "current blead" as its base point?

Turns out I accidentally built a reference perl that is considerably 
newer than the base of chip's branch. Doh! So much for working on trains.

So yes. This can cause both false-positives and false-negatives. Should 
I restart the smoke? In a few days(?) we might actually have stronger 
hardware available.

In related news, if any frequent CPAN tester is reading this, I'd like 
to hear about distributions that you commonly blacklist because they do 
stupid things like reading from STDIN in Makefile.PL (which means 
they'll be killed by Chris' timeout logic and waste a whole lot of time).


Thread Previous | Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About