While going through http://db.test-smoke.org I bumped into a couple
of broken pages:
http://db.test-smoke.org/FAQ.html linked from the upper navigation bar
http://db.test-smoke.org/cpan.html linked from the 'stats' page
'smokers' leading to http://db.test-smoke.org/cgi-bin/smoke_db.cgi?smokers=1
has a section called 'Smokers in last 6 month' which is empty.
The 'Last failures' shows failures from more than 3 years ago. Is this
on purpose?
(e.g. aix 4.3.1.0 aix cc_r 5.0.2.0 smoke patch 18441 clicking on
'Original report')
Under 'Filter' the latest date one can select is 2003-12
Just in case someone cares to fix them I am about to show this to a few
QA managers on Sunday in hope they will implement something similar
in their organization.
Gabor
Hi Gabor,
> While going through http://db.test-smoke.org I bumped into a couple
> of broken pages:
I installed that some time ago with the software Alain Barbet provided. I've
never updated his software, as it was too geared for his situation, and it
took too much efford to keep his stuff up to date.
This is why I started TinySmokeDB and PerlCover on test-smoke.org:
http://gromit.test-smoke.org/cgi/tsdb
http://gromit.test-smoke.org/cgi/tsdb?mode=listlast
http://www.test-smoke.org/perlcover/
[snip]
> Just in case someone cares to fix them I am about to show this to a few
> QA managers on Sunday in hope they will implement something similar
> in their organization.
I *do* care, as I see perl-smoke-testing and reporting on it as my
contribution to the perl-comunity.
Please let me know what you want to show, and I'll try to help you, but I hope
the above links help you.
Good luck,
Abe
--
Nick> > Over to you, Jarkko ???
Jarkko> Urque.
Hmm... I can't seem to find a patch in there anywhere.
-- Nicholas Clark on p5p @ 2005-01-23
On 12/1/06, Abe Timmerman <a...@ztreet.demon.nl> wrote:
> Op een mooie herfstdag (Thursday 30 November 2006 18:46),schreef Gabor Szabo:
> > Hi,
>
> Hi Gabor,
>
> > While going through http://db.test-smoke.org I bumped into a couple
> > of broken pages:
>
> I installed that some time ago with the software Alain Barbet provided. I've
> never updated his software, as it was too geared for his situation, and it
> took too much efford to keep his stuff up to date.
>
> This is why I started TinySmokeDB and PerlCover on test-smoke.org:
>
> http://gromit.test-smoke.org/cgi/tsdb
>
> http://gromit.test-smoke.org/cgi/tsdb?mode=listlast
>
> http://www.test-smoke.org/perlcover/
Ah, when I wrote the e-mail I thought that http://db.test-smoke.org
is the official site as I saw it mentioned in a "recent" post on this list.
Then I found out that actually that post was from 3 years ago, just that the
daily...@perl.org list is so low in traffic that that message was still on
the first page of the archive.
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.
>
> [snip]
>
> > Just in case someone cares to fix them I am about to show this to a few
> > QA managers on Sunday in hope they will implement something similar
> > in their organization.
>
> I *do* care, as I see perl-smoke-testing and reporting on it as my
> contribution to the perl-comunity.
OK, it seems my wording again did not came through as a humble request
but as some rude comment. Sorry for that.
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?
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 :-)
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)
(C)opyright on the top is out of date.
Smokefarm: Nice pictures. I wonder are all these machines in one place?
Who owns them, who maintains them. 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.
Also how do you collect the status information
http://www.test-smoke.org/status.shtml about the smoke servers ?
>
> 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.
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.
Hi Gabor,
> On 12/1/06, Abe Timmerman <a...@ztreet.demon.nl> 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
Yes, that's it.
In addition, next to the FAILED lines it would be nice to have another
column with
the change number (and a link) of the last successful report.
Gabor