Re: libegg revisited

[Jonathan Blandford <jrb redhat com>]

Mark McLoughlin <mark skynet ie> writes:

> 	1) The multiple copies of the code are not generally not being kept up
> to date in the modules where they are being used. I think we need some
> mechanism whereby the maintainer of the code in libegg notifies the
> holders of copies of any changes. Either that or libegg should be
> installing loads of little static libraries. Actually, I think the
> latter way could work well.

There's a couple of update scripts in CVS that should get wider use.  I
don't feel strongly that this is really a huge problem, insomuch as it's
a symptom of problem two.  If we have nice documentation on how to use
the upgrade scripts.  Perhaps putting a sample directory in cvs would be
good.

> 	2) There's no doubt - various bits are languishing in libegg. They
> either don't have a clear destination library or the path to get it into
> the destination library isn't clear. What's supposed to happen here is
> that the author and maintainer of the destination library are
> responsible for making sure things are progressing, but that's not
> working very well. Perhaps we can have some overall libegg maintainers
> who can get nazi on people's asses and kick out things that aren't
> progressing.

Yeah.  This is a serious problem.  The biggest offenders here are
probably the EggStatusIcon and the recently-used files stuff.  I'm still
not sure what the upstream target for them is.  I don't think we really
need a maintainer here, in as much as being a lot more disciplined about
getting things upstream in a timely fashion.  

Perhaps a file in every directory indicating the target module and
version would be a step in the right direction?

-Jonathan