Re: Getting libboost into testing
On Fri, Feb 01, 2002 at 11:52:53AM +1000, Anthony Towns wrote:
> On Thu, Jan 31, 2002 at 05:46:52PM -0500, Steve M. Robbins wrote:
> > One thing worries me: the old boost package built a whole bunch
> > of shared libs. The new one does not. Is the testing script going
> > to throw a fit that those packages will no longer be buildable from
> > source?
> Not unless other packages depended on those libs, and haven't been updated
> to work without them.
> What is bad, though, is:
> * 1.26.0-2 (alpha) (latest build at Jan 23 01:36: failed)
> (-3 hasn't been tried, maybe it fixes the problem)
Yeah. I have no clue what is causing that (the *assembler* says
"string too big"). Alpha is the only architecture with that problem.
I was hoping that it is just a transient bug in the toolchain or "sed"
(the file in question is generated using a template file and sed) and
that the buildd would succeed on -3. But it hasn't tried it yet.
I get the impression that the buildd's work on a LIFO basis: for a day
or two after an upload, the package tends to get built on several
architectures, but then the rate of rebuilding falls off ...
> + libboost-examples/mips unsatisfiable Depends: libboost-dev 
> + libboost-examples/mipsel unsatisfiable Depends: libboost-dev 
> Arch: all packages should *only* depend on packages that are available for
> *all* architectures.
That's tricky. In principle, "libboost-dev" should be available on
all architectures -- it is architecture "any". In practice, I am
held hostage to the abilities of the buildd machines to get it built.
For this package it is easy to just remove the dependency, since it is
just example code. Arguably that is the more correct approach anyway:
one could want to install it just to have a look at the code, not
meaning to compile it.
by Rocket to the Moon,
by Airplane to the Rocket,
by Taxi to the Airport,
by Frontdoor to the Taxi,
by throwing back the blanket and laying down the legs ...
- They Might Be Giants