Re: GUB unofficial release still relevant?

[Graham Percival]

On Sat, Jun 02, 2012 at 08:33:24PM +0100, Phil Holmes wrote:
> ----- Original Message ----- From: "Graham Percival"
> <address@hidden>
> To: "Lilypond Dev" <address@hidden>
> >If anybody else wants to chime in and offer hosting, of course
> >that would be quite appreciated.
> 
> I'm personally not convinced this is worth doing.

As far as "a service for users", then no, it's not worth doing.
But as far as "a service for developer", it'll be extremely
valuable.  Work on GUB basically stopped a few years ago.

The main point right now is just to test whether the binaries
Colin's produced are valid -- particularly on OSX.  I recall that
he has an OSX-ppc machine; if he doesn't have an OSX-x86 machine
then that's where the web server becomes useful.

Once his unmodified (other than the lilypond-headers fix) GUB is
known to produce good binaries, then he can start fixing other
stuff.  For example, IIRC all osx binaries call themselves
2.14.2-1 ?  If he thinks he has a fix for it, he can whip up some
new darwin binaries, upload them, and ask for testing -- without
involving me (or you).  If his untested patches need to wait for a
new devel release, then that'll add a really annoying 2-week delay
to any effort to improve the OSX releases.
(or windows releases, or anything else)


This is all part of my sneaky plan to diversify and broaden our
developer community.  Colin is now (or soon will be) able to
debug+add new features to GUB without relying on me.  That's a
huge step forward, and both our (yours and mine) lives will be
easier for it.

> However, if we really want to do it, I could host it off
> philholmes.net/lilypond, or a new domain. I can set Colin or
> others up with ftp access if needed.

Sounds good!  Again, I'm not imagining a lot of downloads.  In
fact, when it comes to testing experimental OSX releases, probably
only the first 2 or 3 people would really matter.

- Graham