How does universe works?

[Daniel Stone]

On Thu, Oct 14, 2004 at 06:02:05PM +0200, Stefan Kluth wrote:
> On Thu, 14 Oct 2004, Daniel Stone wrote:
> > Er, our buildds don't usually suffer from transient failures like that.
> > :)  If there's a power outage, then it'll get rebuild when the daemon
> > comes up (unlikely as we're in a proper hosting facility with redundancy
> > up the hizzle).  As for system crashes -- that happens for most packages
> > (copies of logs are easily available, and automatically sent to the
> > people responsible for that architecture), which will cause a retry, but
> > I don't believe this happens for universe.
> >
> > I certainly wouldn't like 10,000 build logs to sift through in my inbox.
> 
> I understand that ... but isn't there a smarter way of doing this?  Like a
> clever script which automatically analyses the logs and builds webpages
> summarising the results in a nutshell: "build-failed", build-ok",
> "build-in-progress" and with links to the logs.  With this we users could
> get a clear picture of the status of a package in (multi)universe
> ourselves without having to bother you.  We might even be able to figure
> out ourselves why a package didn't build from the logs!  And having that
> information we could perhaps even work out fixes.  At the moment it is
> pure guesswork and trial-and-error.

http://buildd.debian.org does this, and as I understand it, having one
of these for Ubuntu is on the TODO list.  It gives you status such as
dep-wait (waiting on a package it depends on to be built, IIRC), failed,
given-back (transient error, try again), compiled, etc, etc.

Also, logs are now available at
http://people.ubuntu.com/~lamont/buildLogs.

-- 
Daniel Stone                                        <daniel.stone at canonical.com>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <https://lists.ubuntu.com/archives/ubuntu-users/attachments/20041014/ff842c4e/attachment.sig>