[gentoo-dev] [RFC] Common build directory variable (for cmake-utils & autotools-utils)

[Michał Górny]

Hello,

The new python-r1 eclass provides a python_foreach_impl() function
which runs given command for each python implementation, in its
respective build directory.

This function collides a bit with the build-dir functionality provided
in cmake-utils and autotools-utils. If used to spawn the phase
functions of those eclasses, it first enters the correct (python
implementation-specific) build directory, then the eclass re-enters its
own, chosen build directory.

In order to work-around that, users either have to:

a) wrap all phase functions in custom wrapper which set CMAKE_BUILD_DIR
or AUTOTOOLS_BUILD_DIR,

b) use cheap tricks like setting CMAKE_BUILD_DIR=. globally (untested).

Both solutions being not really that good, I think it would be useful
to use a common BUILD_DIR variable for all eclasses. In that case,
python-r1 would just set BUILD_DIR, and cmake-utils or autotools-utils
phase functions would just use that instead of FOO_BUILD_DIR.

What are your thoughts?

-- 
Best regards,
Michał Górny


signature.asc
Description: PGP signature

Re: [gentoo-dev] CIA.VC down for the count?

[Patrick Lauer]

On 08/23/12 19:34, Anthony G. Basile wrote:
> Hi everyone,
> 
> With cia.vc no longer working, its hard to keep track of one another's
> commits in real time.  I used to use the web page and the IRC channel
> like a gitweb log to see what was going on.  Any suggestions on how we
> can get the visibility back?

I've reanimated #gentoo-commits. It's displaying all commits from the
gentoo-comits@lists.g.o emails. The output might randomly break, so if
you notice any glitches please tell me.

> 
> I think its safe to "break the silence" in #gentoo-commits if we want to
> discuss it there.
> 

I've broken the silence ;)

Take care,

Patrick