Front page | perl.perl5.porters |
Postings from August 2011
On 08/15/2011 12:22 AM, chromatic wrote: > On Sunday, August 14, 2011 at 03:08 am, Steffen Mueller wrote: > >> Just keep the cost of these tests in mind. We only have limited >> resources of people willing and able to do them. > > If the process (and necessary human interventions) were more clearly > documented, I can imagine an article on Perl.com could find a few more > volunteers. That may be useful at some point. Right now, somebody (me, *sigh*) needs to sit down and make it work in an automated way at all. I'm new to the current tools for CPAN smoking. My plan is to achieve all of the following: 1) Run full CPAN smokes on $N copies of perl that differ in some way (build flags and/or commit hash). 2) Get all FAIL/PASS/NA/etc results as text files in a directory. 3) Compare them one-by-one and create a report that has only the differing cases. This could be done with a bit of shiny by having a metacpan-alike server that shows such reports in a simple web interface. That would allow decoupling of some of the setup from volunteers willing to donate the actual CPU cycles. Let's see how far I get. Best regards, SteffenThread Previous