Re: fakeroot a solution for multi-architecture building?
On Fri 26 Sep 1997, Guy Maor wrote:
> Excellent idea!
>
> I'm willing to work on master's end of the process. I think a pull
> process is easier than a push process.
ok, let's do it. i have written shell scripts before to do auto
compiling, so i will "only" have to rewrite in perl.
is there a machine somewhere read for testing ?
> 1. There would be directories for each architecture - COMPILE/{i386,m68k,..}.
> dinstall will drop the source packages into the appropriate place,
> using hard links between archs to save space. It will also touch a
> foo_vers.Ready file for each source's architecture signifying that the
> source is ready to be compiled.
will you :
a) fill a package to all architectures
b) only to architectures that are not yet compiled ?
with a) we still can drop the package, getting an ix86 isn't hard, and
we have an additional test.
> > 5. On success automatic upload to master
a special diretory ? only a normal changes file, or something special ?
the log file should also go to master, so we can collect them and make
them available via www for testing.
> > 6. On failure an automated bug report is filed against the package with
> > the output of the failed build.
lets say : upload the logfile and (if generated) all deb files to
master, and do there something with that. what about a seperate mailing
list for that ? debian bugs could grow a lot. we could also (for testing
purpose) create a mailing list for all logfiles/also successfull builds.
> I can implement the master end. Somebody else can write the script
> which runs on other machines. Anybody that has free cycles can then
> run it, only at night, or with a high nice value, or whatever.
ok. i can do that (i bought the perl book weeks ago and read 2/3 only
for that purpose :-).
andreas
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org .
Trouble? e-mail to templin@bucknell.edu .
Reply to: