Re: your build system is broken, and your bug reporting more so!
- To: Chris Waters <firstname.lastname@example.org>
- Cc: email@example.com, firstname.lastname@example.org
- Subject: Re: your build system is broken, and your bug reporting more so!
- From: Ingo Juergensmann <email@example.com>
- Date: Mon, 30 Apr 2001 21:14:23 +0200
- Message-id: <20010430211423.K1006@muaddib.localnet>
- In-reply-to: <20010430110647.B17098@starless.xtnet>; from firstname.lastname@example.org on Mon, Apr 30, 2001 at 11:06:47AM -0700
- References: <200104301231.f3UCV8903184@jeswick.caldera.de> <20010430110647.B17098@starless.xtnet>
On Mon, Apr 30, 2001 at 11:06:47AM -0700, Chris Waters wrote:
> > The autobuilder found some missing build dependencies for wmrack:
> > libxpm4-dev
> How did you come to that conclusion? There is no such package!!
> There wasn't the last time I closed a spurious bug report from you
> about this (#86585), and there still isn't *this* time that you
> submitted a spurious bug report! (#95826)
> > >From the logs:
> > > checking for XpmCreatePixmapFromBuffer in -lXpm... no
> > > configure: error: WMRack needs XPM support!!!
> > > make: *** [stamp-build] Error 1
> XPM support is provided by xlibs-dev. Which my package DOES
> build-depend on. So, first of all, you either don't have xlibs-dev
> installed, or you have a broken installation. But second, and more
> important, whatever mechanism you used to decide that this log message
> meant a missing dependency on a non-existent package is broken!
Grab your Browser, go to http://m68k.debian.org/cgi/build-info.pl und do a
search for xlibs-dev.
If xlibs-dev would be available and your build-depends would be valid,
sbuild would have installed and used it.
> I don't know if you came up with that on your own, or if you have some
> tool that suggests dependencies. If the former, please learn better!
> If the latter, please fix the tool. In any case, please stop filing
> erroneous bug reports.
Go to http://m68k.debian.org and read about how autobuilders work.
> I realize that porters are overworked, but
> submitting invalid bug reports helps no one!
Complaining in a very rough way about this, doesn't help either.
So, please calm down and request an account on one of the buildds to find
out by yourself what goes wrong.
> I'm cc'ing this message to the 68k porters list in the hopes that some
> of your fellow porters can help you sort this all out. I'm willing to
> help too, in any way that I can (except, of course, by adding a
> build-depends on a non-existent package).
m68k arch has now several buildds and there should be no problem to request
an account. On each machine you can use the same tools such as sbuild as the
buildd itself does.
>From my point of view:
Roman is the person who knows best about buildds and how they are working.
He wrote the buildd, but you can read this on above mentioned URL.
He surely has a long experience how to read error logs from buildd and knows
how to report bugs - no doubt he is only a human being, too, who can make
errors by his own.
Buildd is a very experienced and well tested system of programs to
automatically build packages from sources. If all build-depends are ok,
buildd usually works just fine and reports a successful build of your
package. But if the build failes it requires human intervention to detect
the error what actually went wrong - and in this case this resulted in a
bugreport concerning your maintained package.
Don't complain about resubmitting same error twice. Noone of the buildd
maintainers are able to keep all of their bugreports in mind. There a about
2200 packages currently installed and you cannot expect that one can know
Ciao... // PowerAnimator & Maya Operator
Ingo \X/ To boldly design where noone designed before!