Weekly report week 31, plans for week 32
> next week i plan to
> - struggled with the swedisch Försäkringskassan and tax
> authority, for money for parental leave
no results here, yet. they still refuse to talk to each other
directly.
> - revise the cerebrum package and upload it
i spent the major part of the week on this, but the still early
state the cerebrum code is in does not make things easy. see
below.
> - finish the fix_ldif rewrite and upload a new version of wlus
done; tested by ragnar. old samba installations suffer
considerably. this needs more work. I need to talk to finnarne
about how he did his upgrade.
> - look for a new notebook
one T41p looked nice.
> - refine the "small jobs list for german feature localisation"
did that. it was announced on debian-edu@l.d.o and the
user@s.d list and translated to german immediatly by Peter.
(thanks, again!). The run of prospective developers is on its way
but has not reached me yet. :-)
> - check out new mobile phone contracts
> - check out Max and Arkturs Kerberos configurations
> - send in the phone and travel bills to markus and vidar
i did not get to this.
Instead i got involved in
- creating a job description for me, togethether with Knut, who
did most of the work.
- rewriting the cerebrum meeting minutes i mentioned allready
last week and circulating them (plus some introductory cerebrum
docs) to lots of involved people. No feedback yet.
- reading lots of email about the SLX Debian labs. because most
was in norwegian, it took a bit longer to read.
- writing cooperative emails to other german schoolserver
developers. Some do not like skolelinux a whole lot for obscure
reasons.
- asking a kind lady at UiO to write/put together urgently
needed documentation for cerebrum.
- informing myself more about bootstrapping a kerberos secured
network. this is not trivial, but several people have done this
and some would be willing to share their code once it is in
presentable form.
Concluding i must say that cerebrum might be stable to run, but a
unpleasent thing to get running. hardly any (are there any at
all?) errors or missconfigurations (like missing options in
config files) lead to scary python tracebacks instead of
informative error messages. Understanding those usually requires
understanding the surrounding code. Since the Configuration files
are in python, too (and are sourced by the code), the config file
options can be in non obvious formats (list, single strings, ...)
which are not apparent from the name of the option. this together
with a general lack of documentation makes this work interesting
and helps me learn more python. It is furthermore hard to be sure
that things are working as intended because i dont know how to
operate the program. there is not self-test for this, either.
We will certainly manage to get this beast to work for us, but it
is not a package that i would like to be confronted with after an
unsuspecting "apt-get install cerebrum-server". I want to rework
the packages to use dpatch and be able to put all the patches
into cvs. I am not sure in how far upstream intents to polish the
package to catch all those errors and make the program nicer to
work with. they seem to be on a tight budget and schedule and
vacation. (c:
In the next week i plan to
- follow up on the tax and social security authorities.
- mail all those bills i have to norway. vidar, markus, could you
please mail me your postal addresses again?
- work more on the job specification
- massaging the debian part of the cerebrum package into shape
- expand the "small programming jobs" list with more details once
i find out what people want to know
- try to come up with a clever plan on when to do all the listed
tasks in my job description (planning)
- solve the fix_ldif issue with samba and the lost passwords
(ragnar) and release an even better wlus.
- look at other schoolservers user admin system and ldap
stucture.
Reply to: