[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: MacOS X development and support

From: Graham Percival
Subject: Re: MacOS X development and support
Date: Mon, 5 Mar 2012 01:26:09 +0000
User-agent: Mutt/1.5.21 (2010-09-15)

On Mon, Mar 05, 2012 at 12:36:13AM +0000, Colin Hall wrote:
> Development of code specific to the MacOS X platforms, especially
> Lilypad, is slow.

Yes-ish.  We don't have a lot of osx-specific code, and unless
there's plans to make lilypad-osx fancier, we don't really need
any more code.

> We don't have devs or bug squadders with access to Mac hardware to
> verify new releases.

You're more in touch with that side than me.

> We don't need MacOS X developers. We just need someone to maintain the
> MacOS X aspects of GUB and direct those devs who can help with
> patches.

This gets my vote.

> Let's drop Lilypad. It's getting in the way of regular releases.


> Let's replace Lilypad with a simple app that is a drop target and log
> window.

> Some things I am doing or could do in the future:
> Mac hardware to test new releases, specifically MacOS X PPC
> 10.4.x. This is now a work in progress. I have an iMac Flat-Panel
> (10.4 PPC G4) and promise of a PPC G5 tower for testing new releases.

I think a single PPC machine is enough.

> Establishing a circle of Lilypond MacOS users that I can coordinate to
> verify specific aspects of new releases. I have started on this and
> have the following users signed up:

Sounds like a great plan!

> Resolving GUB build issues relating to Mac OS X targets. Not much
> progress here. I have cloned the GUB git on my Linux x86_32 and x86_64
> machines but I haven't got past make bootstrap. I wouldn't mind some
> help with that.

This is the single most important thing -- not just for osx, but
for all binary building.  However, AFAIK there is no help
available for GUB.

I can give quick answers to GUB questions that I know the answer
to, but I'm not willing to spend a great deal of time debugging
it.  I spent 100+ hours doing that in the fall of 2009; I consider
that I've paid my dues.

> Finally, some specific questions:
> Does GUB work?

Yes-ish.  It works on my desktop when I boot into lilydev 10.04.
It does not work on my desktop in my usual ubuntu 11.10, which
means that I can't use my desktop for anything important while GUB
is compiling.

Some developers can run GUB on lilydev.  Others can't.  No pattern
has emerged in this yet, nor has anybody wanted to get into it.
My initial guess is that it would take me 2-10 hours to get GUB to
compile on ubuntu 11.10.  To estimate how long it would take
somebody else, multiply those numbers by whatever factor you think
is appropriate.

That's not very encouraging, but *somebody* needs to do this at
some point.  I'm not going to keep ubuntu 10.04 around forever.

> Is GUB used to produce the Lilypond releases for all platforms, or is
> it only used to create some of them?

All platforms, plus the official regression test comparisons, plus
the official documentation.

> Which develops have access to MacOS X hardware?

Not me.

> Is it possible to verify Mac OS X x86 using a VM?

I doubt it; apple doesn't like their OSes inside VMs (IIRC it's
actually expressedly forbidden in their EULA), but even if that
was possible, it would likely only be osx-x86 running on top of
osx-x86 hardware.

> Finally, I can see me having a couple of Macs here to test releases, a
> working GUB install for resolving build issues, and a few keen Mac
> users who verify fixes on Mac plaforms that I don't have to
> hand. Getting to that point will be a chunk of work. I have an inkling
> that others have tried to do this too, and given up.

Focus your efforts on GUB.  And yes, everybody who's gotten the
latest make bootstrap error has given up.

> I don't want to set out in that direction unless it is something
> that the dev community support and that I have a reasonable
> chance of success.

As long as you don't give up, you *will* succeed eventually.  What
"eventually" means depends on your existing knowledge of unix
build systems and how much other aid you get.

But again: this problem isn't going to go away unless somebody
tackles it.  GUB is the weakest link in our development chain.
Bringing it under control would be a great step forward.

- Graham

reply via email to

[Prev in Thread] Current Thread [Next in Thread]