Re: What happened to libgtkhtml19 in unstable?
On Tue, Dec 11, 2001 at 12:51:54 +0000, Adam Olsen wrote:
> On Tue, Dec 11, 2001 at 10:57:28AM +0100, J.H.M. Dassen (Ray) wrote:
> > That's not a guarantee at all. As a libgal developer explained to me:
> > "GAL is not meant to be api stable. It role is to provide a place to
> > share code and to develop new interfaces. When things stabilize they
> > move down into gnome or gtk.".
>
> If there's no stable api *at all*,
Well, there is API stability in the sense that source that can be compiled
against one version of gal will usually compile against a newer version as
well.
> then wouldn't it make sense for apps that want to use GAL stuff to just
> rip it out, and plonk it into their own source trees?
Linking against gal statically is much more straightforward.
On Tue, Dec 11, 2001 at 06:33:57 -0600, Anthony Towns wrote:
> The gal so-name nonsense is a pain in the neck for management of "testing"
> too.
That's the reason I started linking against gal statically. It simply
changed to fast for the gnumeric packages to have a reasonable chance at
percolating into testing.
On Tue, Dec 11, 2001 at 12:45:33 +0100, Adrian Bunk wrote:
> Should we stop to ship shared versions of libgal and link everything
> statically with it? This would resolve the problems in other programs the
> same way gnumeric already does.
I'm in favour of this. There is however the problem of policy 11.2 which
says "All libraries must have a shared version in the `lib*' package [...]".
I vaguely recall a proposal to lessen this requirement, but can't find it in
the BTS.
Ray
--
"a infinite number of monkeys typing into GNU emacs would never make a good
program"
.../linux/Documentation/CodingStyle
Reply to: