RPM Community Forums

Mailing List Message of <rpm-devel>

Re: RPM_CHECK_LIB vs older autotools

From: Jay Soffian <jay-rpm@soffian.org>
Date: Tue 17 Jul 2007 - 21:55:36 CEST
Message-Id: <7B83A4E2-4C83-460D-B841-B7E0F2501292@soffian.org>
On Jul 17, 2007, at 2:15 PM, Ralf S. Engelschall wrote:

> IMHO mostly all of our actual users are mainly people from  
> distribution
> vendors who want to test-drive RPM 5 for _their_ distribution.  
> AFAIK the
> only way they are willing to do is by using the RPM 5 source tree and
> using the usual triple of "configure --prefix=/tmp/rpm [...]", "make",
> "make install". They will certainly _NOT_ use any SRPM out-of-the-box,
> except for simply unpacking it with rpm2cpio.

+1

> No, sorry, the _TARBALL_ itself has to carry the _FULL_ snapshot
> information to make sure people do not treat the tarball as a release
> version.

+1

On Jul 17, 2007, at 2:38 PM, Jeff Johnson wrote:

> If tarball (or srpm) only is distributed, then one gets build  
> failure reports, not
> functional failure reports, because of the choice of distribution  
> format.

A tarball (or srpm) is necessary, but not sufficient to get reports  
about both build failures and functional failures. It's gotta build  
first, of course...

A binary is necessary, but not sufficient for functional failure  
reports too, while eliminating the possibility of build failure reports.

Not trying to lecture here, but why not just offer up all of tarballs/ 
srpms/rpms? As long as a machine is doing the work of setting them  
all up, what's the difference?

$0.02. Feel free to ignore. :-)

j.
Received on Tue Jul 17 21:55:40 2007
Driven by Jeff Johnson and the RPM project team.
Hosted by OpenPKG and Ralf S. Engelschall.
Powered by FreeBSD and OpenPKG.