[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Outdated GNU config (config.{sub,guess}) and autotools-dev



On Mon, Jul 23, 2001 at 11:35:49AM +0200, Robert Bihlmeyer wrote:
> Autobuilders will not upload source packages. They will upload a
> binary built from a modified source that is available nowhere. 
> What's worse, after the build log is purged, you won't even know for
> sure which version of autotools-dev was used, so there is NO sure
> way to reconstruct the source used.
> 
> My proposition is to empower the maintainer (or a person doing a
> NMU) to check for up-to-date config.* scripts, and update them
> easily. The check part could be made in different parts of the
> infrastructure, depending on how obnoxious you want to get. 
> Examples:
> 
> * lintian warning or error
> * dpgk-buildpackage refuses to work with old scripts
> * dpkg-source -b refuses to bundle up old scripts

I think _in_general_ this is the right direction to go. However, I
dislike the "refuses" - there will always be reasons to break the
rules, one ("s/linux-gnu/linux/") even mentioned in this thread. 
Anyway, how exactly are you going to determine whether the script is
"old"?

> And we could standardize on a "update-config-guess" target in
> debian/rules, that would copy the files to the right locations of
> the package.

That's what I've also been thinking of - but, more flexible would be a
"source-prep" target in debian/rules, called by dpkg-buildpackage
whenever a build will result in source being uploaded. People could
then use this hook to their liking.

> P.S. Why does libtool use this dreaded scripts at all? The concept
> of determining the system type to base some other decision on that
> seems severily broken to me, and goes against the spirit of
> autoconfiguration.

I think it's useful mainly in case of cross-compilation. I don't like
it, either...

Cheers,

  Richard

-- 
  __   _
  |_) /|  Richard Atterer     |  CS student at the Technische  |  GnuPG key:
  | \/¯|  http://atterer.net  |  Universität München, Germany  |  0x888354F7
  ¯ ´` ¯



Reply to: