Re: summary of the last weeks of testing.
i move this thread to the debian-edu list, since it is in
english for my sake (thanks!) and development related.
* Morten Arnesen (mort1arn@online.no) [040415 13:34]:
> this started allready on the slx gathering, when the gathering-cd could
> not be made before i left, but jim sent me a pr47 cd, not a daily build.
> this gave me a 80 Mb download on isdn (!) wow, not strange that i first
> of all made me a snapshot of the server before testing.
how come? what did those 80 Mbyte consist of?
> when i finaly got around started with the tests things started to
> happends. a lot of testers and developers worked day and night. GREAT
> JOB ALL !
yes, great. i am impressed with the diligence and endurance
people have shown.
> then suddenly Andreas started to implement new functions in wlus (?) hey
> man!
which new features do you refere to?
allow me to explain a bit background here. i assume you mean the
attic, which i described as "the biggest change" in my last
release mail. this attic related patch is a wooping +53 -19
lines. of those most are cosmetic. the changes happen in the last
three sections of
http://developer.skolelinux.no/cgi-bin/viewcvs.cgi/skolelinux/src/webmin-ldap-skolelinux/ldap-users.pl.diff?r1=1.90.2.37&r2=1.90.2.38
no rocket science there. And no breakage at all.
The changes that triggered the breakage that was there for a
longer time allready (better hidden, but there non-the-less) were
in the second and third section of
http://developer.skolelinux.no/cgi-bin/viewcvs.cgi/skolelinux/src/webmin-ldap-skolelinux/WebminLdapSkolelinux.pm.diff?r1=1.39.2.55&r2=1.39.2.56
those two lines beak my neck and webmins http headers. those were
no new features but merely a new attempt to fix the
smbpasswd/sambasync problems finnarne and i saw for some time.
> you don't add new functions when the users test is started, that
> we'll take in the next version. result was a major brake in our work,
> and some days off, or at least some days without any usefull testing.
i understand and regret that loss of productive time. I did not
intent or knowingly put up with the breakage. i did not waste
time hanging out in the sun, either. i felt a certain discompfort
of having a bunch of people sitting and waiting and searched
franticly for a bug that eluded me for two days and I got more
and more itchy.
> wlus is of course all testers favourite, because it's wery important to
> get this running, most everything can wait but wlus. i suggest that you
> don't throw such comments to the list even though you might be tired. :)
huh? was anything i wrote offensive? i felt i mocked myself, at
best. sorry if it was percived disrespectfull in any way. and
yes, it was 03.00 and my son had just made a mess in his dipers.
(c:
> finaly i'll mention that i feel the programmers test was not done, just
> because of that we spent the first time in testing with programmers
> test.
with programmer test you mean "does this thing run, at all?"? i
understand that the first time you fought a few problems that
were unrelated to wlus (like no running ldap server and other
issues).
well, in retrospective it is hard to belive (even for me) but i
went through the normal quick user adding, searching deleting
routin before i inflicted that fatal release upon the world. and
it worked for me. I cant really explain how it did that, though. (c:
> just correct me if i'm wrong, but this is just my opinion. i guess
> we could have been close to the finish line by now if those things have
> not been so.
i agree 100%. i got other private mails pointing out similar
issues. i had tried to finish wlus in january and had released
three betas before then, and announced those on the mailinglist.
i had asked for feedback during the process and knuty even wrote
a mail encouraging people to test it, after i got only little
response. feedback has not been overwhelming after his mail,
either, but i got a few bug reports.
i had not expected that wlus was done 100%, but i felt pretty
confident about it. things i expected to change were wording and
minor layout changes. pretty naive from todays perspective, i
might add.
a few weeks back people found the jnadmin and generic age group
in the root.ldif which were a result of my effords and it became
aparent that agegroups confused people more then they helped. so
before the developer gathering i had fixed all but 2 of the bugs
in bugzilla and also removed the agegroups, again. that was a
pretty invasiv change with 3 or 4 issues discovered by the testing
crew. my testing infrastructure (consisting of a notebook,
provided graciously by InOut) had decayed somewhat, segfaulting a
lot. The tests that i did were on my debian sid development
system and lacking for that reason.
Reply to: