Subject: Re: Webfrontend on pbulks
To: None <email@example.com>
From: Ulrich Habel <firstname.lastname@example.org>
Date: 11/12/2007 11:42:33
> This is very cool.
Thanks, it's just a start and I think there is a use for it. I'll
continue to setup things and import pbulks to gather more data.
> It would be good if the "done" packages could be links to the binary
> packages, and if the "failed" ones would link to the build logs.
good point - however, I see a problem there. Not all builders are
uploading the packages/are developers. Usually I delete the bulk reports
from to time to free some diskspace (around 200-300mb for a SunOS 5.9
This would just lead to broken links - maybe I should add them and check
them from time to time.
> What does the "apache" listing mean? Is it apache13, apache2 or
> apache22? And which version of them?
huh, this is just an error with my algorithm. Right now I am using
from the pbulk report file. This, however, doesn't work with all the
packages. Maybe we should leave this out for now, make some more
improvement and see if I can get this tool to something usuable. Then we
should start of thinking what to change in the pbulk report :)
> When this tool has grown up, it should probably be integrated into the
> README.html generation, and maybe it can even replace it. I don't like
> the current practice that the README.html are in the pkgsrc/ directory,
Would be a worthful addition, yes. However, I'll leave this out for now
and get more work on the functionality for bulks done. Remember - I only
use the pbulk report file and it just doesn't provide data.
some things are missing like: the os, architecture, builddate
So right now the "importtool" uses some commandline options to feed in
additional data. I try to avoid to end in a config file for the import