
jdhore at gentoo
Apr 29, 2012, 11:31 PM
Post #9 of 24
(341 views)
Permalink
|
|
Re: new virtual/pkgconfig to support lighter alternatives
[In reply to]
|
|
On 30 April 2012 02:16, Michał Górny <mgorny [at] gentoo> wrote: > On Mon, 30 Apr 2012 00:08:34 -0400 > Mike Frysinger <vapier [at] gentoo> wrote: > >> On Sunday 29 April 2012 18:40:00 Jeff Horelick wrote: >> > On 29 April 2012 18:11, Mike Frysinger <vapier [at] gentoo> wrote: >> > > the canonical pkg-config is getting fat.  it requires glib-2.  it >> > > runs pkg- config when building.  glib-2 requires pkg-config. >> > > whee. >> > > >> > > for our normal systems, this isn't a big deal.  but we'd like to >> > > enable a lighter alternative for embedded/alternative systems. >> > > as such, i'd like to introduce a virtual/pkgconfig that allows >> > > for selection of simpler (but compatible) implementations. >> > > >> > > we've got an implementation in perl (i'm not interested in), but >> > > there is also "pkg-config-lite" and "pkgconf".  they should be >> > > compatible with the canonical pkg-config.  they aren't yet in the >> > > tree, but will be once we agree on this topic. >> > > >> > > any comments ? >> > >> > I'd just like to say, i'm also an Atheme project member and I have >> > authorisation from nenolod (the primary pkgconf developer) to make >> > changes and stuff, so I can upstream any changes necessary to make >> > pkgconf work for us. >> >> that sounds really good.  i sent you some patches ;). >> >> however, it's missing pkg.m4.  any thoughts on that ? > > Maybe we should provide it independently in some other package. > Considering the implementations are supposed to be compatible, the .m4 > file should work fine with all of them. And we'll create same configure > files independently of which impl particular user uses. > > -- > Best regards, > Michał Górny Well since the 3 primary implementations (fd.o pkg-config, pkg-config-lite and pkgconf-0.2) now provide it, I don't see a huge use for a seperate package. Also, the pkg.m4 used by all 3 seems to be identical so...
|