Bug 68763 - Patch to bsd.port.mk and tree to support X.org
Summary: Patch to bsd.port.mk and tree to support X.org
Status: Closed FIXED
Alias: None
Product: Ports & Packages
Classification: Unclassified
Component: Individual Port(s) (show other bugs)
Version: Latest
Hardware: Any Any
: Normal Affects Only Me
Assignee: Port Management Team
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2004-07-07 11:10 UTC by Eric Anholt
Modified: 2004-07-23 20:27 UTC (History)
0 users

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Eric Anholt freebsd_committer freebsd_triage 2004-07-07 11:10:25 UTC
	At USENIX I talked with re (scottl) and others about making xorg the
	default for 5.3, and it was agreed that the switch would be fine, even
	with pkg_delete -f and reinstalling being necessary, as long as it
	happened before -STABLE.  Of course some solution involving MOVED so it
	was all automatic would be better, but I'm not aware of any without
	repocopying XFree86 away and making X.org the default for all.

	Here's a (link to a) patch to make the switch, and make the XFree86 and
	X.org ports be supported equally by bsd.port.mk and the ports tree.  I
	haven't actually completed a "make index" -- it died on
	x11-toolkits/open-motif last time and since then I cvsupped and other
	ports broke it it seems.

	Could this get queued for bento testing?

Fix: 

http://people.freebsd.org/~anholt/X/xorg-conv-2.diff
Comment 1 Mark Linimon 2004-07-08 21:08:30 UTC
Depending on timing, the patch to x11-fonts/p5-type1inst might or might
not apply -- I have approved a pending change submitted by hrs to take
over maintainership, and that would be within the space of the context diff.
The Makefile does not change, otherwise, except for a PORTREVISION bump.
Comment 2 Pav Lucistnik freebsd_committer freebsd_triage 2004-07-08 22:33:02 UTC
Responsible Changed
From-To: freebsd-ports-bugs->portmgr

Over to portmgr
Comment 4 Kris Kennaway freebsd_committer freebsd_triage 2004-07-14 12:02:14 UTC
State Changed
From-To: open->analyzed

Currently testing on pointyhat
Comment 5 Kris Kennaway 2004-07-14 12:18:03 UTC
On Sun, Jul 11, 2004 at 09:00:41PM +0000, Eric Anholt wrote:
> The following reply was made to PR ports/68763; it has been noted by GNATS.
> 
> From: Eric Anholt <eta@lclark.edu>
> To: gnats <freebsd-gnats-submit@FreeBSD.ORG>
> Cc:  
> Subject: Re: ports/68763: Patch to bsd.port.mk and tree to support X.org
> Date: Sun, 11 Jul 2004 13:54:36 -0700
> 
>  New diff is at:
>  
>  http://people.freebsd.org/~anholt/X/xorg-conv-3.diff

I'd like to retain support for XFREE86_VERSION for now, i.e. if that
variable is set then give a warning and set X_WINDOW_SYSTEM
appropriately.  After some time we can make the warning fatal.

Come to think of it, should we take the opportunity to change to using
WITH_X11=(x-org|...) (cf existing WITH_X11=foo/WITHOUT_X11=bar)
instead of X_WINDOW_SYSTEM, to bring it into the WITH_* namespace?
OTOH most (all?) other WITH_* variables don't care about the value as
long as they're set, so that may be a POLA violation.

I didn't have time to review the patch in more detail.

Kris
Comment 6 Eric Anholt 2004-07-15 03:10:10 UTC
On Wed, 2004-07-14 at 04:18, Kris Kennaway wrote:
> I'd like to retain support for XFREE86_VERSION for now, i.e. if that
> variable is set then give a warning and set X_WINDOW_SYSTEM
> appropriately.  After some time we can make the warning fatal.

Oh, you mean in the XFREE86_VERSION=3 case?  Yeah, that can be fixed.  I
think we're pretty sure that nobody uses XFree86-3 any more (how long
was INDEX broken for XFREE86_VERSION=3?).  Seriously, XFree86 3 is dead,
and I have a hard time understanding why we support it when we don't
force ourselves to support ancient versions of other pieces of
software.  I think it would be interesting to see how much of the ports
tree goes broken (at build-time, at least) if X_WINDOW_SYSTEM=3 gets
set.

> Come to think of it, should we take the opportunity to change to using
> WITH_X11=(x-org|...) (cf existing WITH_X11=foo/WITHOUT_X11=bar)
> instead of X_WINDOW_SYSTEM, to bring it into the WITH_* namespace?
> OTOH most (all?) other WITH_* variables don't care about the value as
> long as they're set, so that may be a POLA violation.

That's about what I was thinking, then I got Dejan's patch with the
X_WINDOW_SYSTEM name, so I went with it.  It's an ugly name, but
WITH_SOMETHING didn't seem to be very nice either.

I was also talking with linimon, and I really like the idea now of
moving this stuff into a bsd.x11.mk so we can do more work without
having to necessarily iterate through portmgr every time.  Notably, at
some point we'll probably see xorg-snap ports, which would be another
X_WINDOW_SYSTEM= switch.  And then at some point in the next year we're
going to end up with modularization and need a USE_X11= type solution
like gnome does.

-- 
Eric Anholt                                eta@lclark.edu          
http://people.freebsd.org/~anholt/         anholt@FreeBSD.org
Comment 7 Eric Anholt 2004-07-19 10:56:34 UTC
http://people.freebsd.org/~anholt/X/xorg-conv-4.diff

Here's an updated version after fixing most of the issues reported in
the test build.

Breakdown of ports that are still broken:
larbin: doesn't fetch (checksum).
xbl, ja-chimera, xdtm, ko-hanterm, pixmap, xzx: Imake issue with X
vendor name define.
xstroke: Some issue with Freetype2 directory define in imake.
dri: IGNORE that's probably not strictly necessary

I would like to commit this soon, once I get the DRI sorted out.  I need
to get the xorg dri tested on 4.3 -- if compatible with the older libGL,
it'll become the new stable DRI.  This will inconvenience XFree86 4.3
users, but newer DRI is more of an advantage than downloading new
distfiles is a disadvantage.

I'm really having a hard time understanding the imake issues,
particularly the vendor name define.  I'm absolutely stumped by that
one.

Any opposition to committing once the DRI is sorted out?

-- 
Eric Anholt                                eta@lclark.edu          
http://people.freebsd.org/~anholt/         anholt@FreeBSD.org
Comment 8 Eric Anholt 2004-07-23 08:15:35 UTC
OK, the DRI and the imake issues have been sorted out.  I think I've
been approved, so I'll commit this tomorrow unless there are other
issues.

-- 
Eric Anholt                                eta@lclark.edu          
http://people.freebsd.org/~anholt/         anholt@FreeBSD.org
Comment 9 Eric Anholt freebsd_committer freebsd_triage 2004-07-23 20:26:48 UTC
State Changed
From-To: analyzed->closed

Committed.  Thanks kris, marcus, linimon, Dejan, and everyone else for all 
their help in getting this update done!