Op een mooie herfstdag (Friday 01 December 2006 08:40),schreef Gabor Szabo: > Hi Abe,
Hi Gabor, > On 12/1/06, Abe Timmerman <[EMAIL PROTECTED]> wrote: > > Op een mooie herfstdag (Thursday 30 November 2006 18:46),schreef Gabor Szabo: > > > Hi, > > > > Hi Gabor, [snip] > I cross post it back to p5p I think we should drop the daily-build list > from this discussion. Actually maybe it should be dropped from the > perlhack.pod as well. http://www.test-smoke.org/ should be mentioned in > that pod. Well, the daily-build list could be "merged" with perl-qa; but I can imagine most porters do not want all of this detail discussed on p5p. I'm not the judge of that. [more snip] > Thanks for the other data collectors you provide. > > A few comments and questions: > > http://www.test-smoke.org/perlcover/ seem to show an old and actually > failed build. (gcov run for 29204 while the latest seem to be 29427) Well, > that if I understand what I see there. Is this on purpose? Could it be > updated? It's running (@29431), but apeared to be hanging, I'll monitor more closely, and update asap. > BTW is the test coverage of perl really only 71.7% ? > > > The HTML and CSS seem to be, well, not the niceset > (but who am I to talk about such issues :-) There's no accounting for taste ;-) > and it is jumping (e.g. on http://www.test-smoke.org/ when the mouse is > over 'my perl stuff' using Firefox. (and it is not valid XHTML any more) Hmmm..., I can only see a problem with the smoquel page.. > (C)opyright on the top is out of date. fixed, thanks! > Smokefarm: Nice pictures. I wonder are all these machines in one place? > Who owns them, who maintains them. Those pictures are all of my own machines, and most of them run perl-smokes. > I though many of the smoke reports come > in from volunteers on this list (and maybe outside of it) who run perl > smoke testing > on their own machine. A description of the situation would be nice. Most of the smoke reports come from other -very nice- people, that donate there cycles to perl smoke testing. > Also how do you collect the status information > http://www.test-smoke.org/status.shtml about the smoke servers ? It's a little script that goes out to the smoke-host and calls the smokestatus.pl script and collects the output. http://www.test-smoke.org/svn/SmokeStatus/ > > Please let me know what you want to show, and I'll try to help you, but I > > hope the above links help you. > > I am preparing a presentation to a bunch of QA managers. > I would like to show them several large open source projects and what > and how do they do to improve quality. There are a number of nice > systems out there with > automatic builds and test runners with systems collecting the test results. > > I would like to show them the 'business value' in these systems and > how do they work. That is the quick feedback cycle. The easy way to > pinpoint locations of issues. > (e.g. how do I find out at which point did the build break on a > certain platform ?) > > BTW it might be useful to show the latest patch level. > I am not sure about others I would like to see a table of platforms with > one line for the latest test report on that platform with red/green colors > on success failure. That is what http://gromit.test-smoke.org/cgi/tsdb?mode=listlast does, doesn't it? > I started to work on something similar for the CPAN Testers: > http://www.szabgab.com/cpantested/ > http://www.szabgab.com/cpantested/dist/Test-Smoke.html > I hope we can fold it in the main cpantesters.perl.org site. > > Something like this (just nicer :-) would - I guess - be useful for P5P > too. > > Look at what PostgreSQL have: > http://www.pgbuildfarm.org/cgi-bin/show_status.pl > > or the Tinderbox of Mozilla: > http://tinderbox.mozilla.org/showbuilds.cgi?tree=Firefox > > actually both written in Perl. > > regards > Gabor > ps. Just to make it clear, I am not trying to push anyone to do any of > these things, > (nor can I currently volunteer to do it myself) I am just sharing my > thoughts. Good luck, Abe -- Anton Tagunov> P.S. People do like gory details :) I don't like telling them. -- Jarkko Hietaniemi on p5p @ 2002-03-06