Bug#584125: RFH: piuparts.d.o is maintainance hungry
Hi Scott,
please keep bugs cc:ed! :)
On Montag, 30. Mai 2011, Scott Schaefer wrote:
> > I can explain both tasks in more detail should you be interested.
> Yes. I am interested. Please provide greater detail when you are able.
yay! keep on asking/commenting!
> Since it appears that most important bug is 526045
well... it has the highest severity for sure... (as its a rather high impact
blocker...) - but then there are other things which are "blocking"/rather very
importing too, eg aiding bug filing or better reporting. Those things are
harder to file in the BTS (and probably would be severity wishlist then),
while they are also very much needed.
(ie piuparts-report now runs for several hours each day. if there would be
caching of the logfile-analysises this could probably go down to a few
minutes, thus taking away load _and_ making development in this area much
easier.
aiding bugfiling (=sub-automatic bug filing) has obvious benefits to me. bugs
not filed are bugs not likely to be fixed, whether they are shown on
piuparts.d.o or not...))
but really, pick what interests you! :-)
> , and from your
> earlier reply, I will need to setup master/slave to attempt to resolve,
> I will set this as my goal for next few days. Since README is missing
> the section on how to do that, I will probably try to produce a document.
see svn+ssh://holger@svn.debian.org/svn/piuparts/piatti/README.txt
have you looked at http://wiki.debian.org/piuparts/ ? It has a "Howto setup a
piuparts test-instance for development" which says: *to be written* - maybe
you can enhance that and eg. need the link to the above mentioned
README.txt?! :)
hm, but then this url _is_ mentioned in
http://piuparts.debian.org/doc/README.html#_running_piuparts_in_master_slave_mode_piuparts_report_and_the_setup_on_piuparts_debian_org
but still - I'm the old fart here and you didnt find the information, so
probably _you_ have more useful insight how to improve the docs than I do! ;-)
> I have limited my work so far to "direct (??)" mode; i.e. not
> master-slave. Still in the introductory stage of understanding. Please
> do not hesitate to tell me if I am headed in wrong direction.
"direct mode" works reasonable well, there is IMO "more room for improvement"
in the parts that deal with processing 30k logs per distro, IOW, large scale
testing.
but it's also (always) the little things that matter much!
cheers,
Holger
Reply to: