Re: Building against testing [was Re: "Recompile with libc6 from testing" <-- doesn't help i think!]
On Mon, 16 Sep 2002, Anthony Towns wrote:
> On Mon, Sep 16, 2002 at 12:24:11PM +0200, Santiago Vila wrote:
> > I believe build daemons should ideally compile against "the most stable
> > development packages which produce right dependencies for sarge".
> ^^^^^
>
> YM sid, here, by the looks.
"[...] right dependencies for what we want testing to be" would be
more accurate.
If we have libfoo1 in stable and testing and libfoo2 in unstable, but we
have decided that libfoo2 will not enter testing, then Depends: libfoo2
would be the wrong dependency and Depends: libfoo1 would be the right one,
for a package which is uploaded for unstable with the intention that it
propagates to testing after ten days.
> You'd need to have some clever hackery involving the shlibs to work
> out when you should do it, let alone getting the buildds to produce
> a chroot constructed from three different suites. [...]
Of course, but I don't think such clever hackery would be much more
clever or much more hackery than the scripts that currently create the
testing disrtibution.
I'm not asking anybody to do this, I'm just saying that if someone had
the time and motivation to write the required code, we should
seriously consider using it.
(Isn't this, more or less, what happened with the testing distribution?).
Reply to: