[gentoo-dev] Remove your modular X unmask

[gentoo-dev] Remove your modular X unmask

[Donnie Berkholz]

Hi all,

This is a reminder to everybody that now modular X is in ~arch, you need 
to REMOVE YOUR MODULAR X UNMASK.

We're getting lots of issues because people aren't doing this and are 
getting package.mask'd versions of programs because they forgot about it.

In particular, xorg-server 1.0.99 has an ABI break, which breaks all 
Nvidia and ATI binary drivers.

Thanks,
Donnie
-- 
gentoo-dev@gentoo.org mailing list

[gentoo-dev] Re: Remove your modular X unmask

[Duncan]

Donnie Berkholz posted <44471766.7000605@gentoo.org>, excerpted below,  on
Wed, 19 Apr 2006 22:08:54 -0700:

> This is a reminder to everybody that now modular X is in ~arch, you need
> to REMOVE YOUR MODULAR X UNMASK.
> 
> We're getting lots of issues because people aren't doing this and are
> getting package.mask'd versions of programs because they forgot about it.
> 
> In particular, xorg-server 1.0.99 has an ABI break, which breaks all
> Nvidia and ATI binary drivers.

FWIW, I played with xorg-server 1.0.99 and the new ATI drivers (Radeon
9250, dual monitor in merged framebuffer mode, on amd64, compiled with
gcc-4.1.0) for about 12 hours yesterday. There's some really nice
improvements coming (with Option AccelMethod "EXA", the CPU cost of
composite transparency/shadows/fading is virtually zero, and EXA is far
more stable than in current ~arch), but as might be expected with an early
release candidate, there's still enough issues that I eventually reverted
back to ~arch, which now seems slow and CPU hungry with
composite/transparency on, dull and stuffy without! <g>

In particular:

* The zoom hotkeys (CTRL-ALT-PLUS/MINUS) didn't work at all -- I had to
use krandrtray which isn't the same since it shrinks the virtual desktop,
not just the viewport.

* krandrtray failed to restore a full sized desktop in some cases (this
may be due to an error in the log I saw -- from memory so isn't exact,
failed to allocate backup/restore memory, display may get corrupted)

* Overlay was doing strange thing.  In this case it was SDL overlay as
used by dosbox -- I tried rebuilding both libsdl and dosbox, but the
overlay still failed.  Maybe related to the ABI change?

* General instability, with both EXA and XAA.  Not entirely unexpected at
this point, and EXA is /vastly/ improved over  both the current ~arch
version and XAA (new and ~arch) efficiency-wise.

Is it worth posting bugs on these yet?  Gentoo or upstream?  Any tracker
bug to point me at?  (I have the binpkgs so can remerge them for
additional testing with little trouble, if necessary.)

It's impressive enough to have me eagerly awaiting the next rc.  Good work
both here and upstream!  Is the proposed schedule on the wiki at
desktop.org still valid? A couple more RCs, and release in about a month
if all goes well?

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman in
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html


-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Remove your modular X unmask

[Mart Raudsepp]

On Thu, 2006-04-20 at 07:05 -0700, Duncan wrote:
> * Overlay was doing strange thing.  In this case it was SDL overlay as
> used by dosbox -- I tried rebuilding both libsdl and dosbox, but the
> overlay still failed.  Maybe related to the ABI change?

SDL is being dumb with alpha visuals here.
Need XLIB_SKIP_ARGB_VISUALS=1 env var for SDL apps if a composite
manager is running. That's for SDL apps, giving that to everything would
defeat the purpose of the composite manager quite a bit, I think.

I'll have to pass commenting on anything else. Not sure about the other
points and don't want to express guesses :)

> It's impressive enough to have me eagerly awaiting the next rc. Good work
> both here and upstream!

Indeed!

> Is the proposed schedule on the wiki at desktop.org still valid?

I think ajax is getting out his whip soon, if he hasn't already :P
The schedule should be pretty accurate on xorg.fd.org wiki.

> A couple more RCs, and release in about a month
> if all goes well?

-- 
With regards,
Mart Raudsepp

Project manager of wxMUD      - http://wxmud.sourceforge.net/
Developer of wxWidgets        - http://www.wxwidgets.org/
GTK+ port maintainer of OMGUI - http://www.omgui.org/

-- 
gentoo-dev@gentoo.org mailing list

[gentoo-dev] Re: Re: Remove your modular X unmask

[Duncan]

Mart Raudsepp posted <1145542992.21629.19.camel@localhost>, excerpted
below,  on Thu, 20 Apr 2006 17:23:12 +0300:

> on xorg.fd.org wiki.

Doh...  I always call it desktop.org, it's /free/desktop.org!  One
would think with a sig such as mine I'd remember that!  Thanks!

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman in
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html


-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Re: Remove your modular X unmask

[Donnie Berkholz]

Duncan wrote:
> Is it worth posting bugs on these yet?  Gentoo or upstream?  Any tracker
> bug to point me at?  (I have the binpkgs so can remerge them for
> additional testing with little trouble, if necessary.)

Yes, and they are most likely upstream bugs. Two things to do before 
posting a bug:
1) Make sure you're using drivers known to be compatible with 1.0.99.
2) Check out your X log for errors.
3) For the RandR bugs, see whether running xrandr allows you to change 
resolutions, and make sure your XKB is working (setxkbmap -print).

> It's impressive enough to have me eagerly awaiting the next rc.  Good work
> both here and upstream!  Is the proposed schedule on the wiki at
> desktop.org still valid? A couple more RCs, and release in about a month
> if all goes well?

Sounds pretty reasonable.

Thanks,
Donnie
-- 
gentoo-dev@gentoo.org mailing list

[gentoo-dev] Re: Re: Remove your modular X unmask

[Duncan]

Donnie Berkholz posted <4447BA3F.50103@gentoo.org>, excerpted below,  on
Thu, 20 Apr 2006 09:43:43 -0700:

> Duncan wrote:
>> Is it worth posting bugs on these yet?  Gentoo or upstream?  Any tracker
>> bug to point me at?  (I have the binpkgs so can remerge them for
>> additional testing with little trouble, if necessary.)
> 
> Yes, and they are most likely upstream bugs. Two things to do before 
> posting a bug:
> 1) Make sure you're using drivers known to be compatible with 1.0.99.
> 2) Check out your X log for errors.
> 3) For the RandR bugs, see whether running xrandr allows you to change 
> resolutions, and make sure your XKB is working (setxkbmap -print).

Thanks.  1.  Done (I noted the posted warning and heeded it).  2.  Done 
(I'd attach it upstream, as I don't claim to understand it all, myself).
3. XKB, check.  Have to try xrandr.

My setup is probably uncommon enough now, what with 8 gig memory,
SMP amd64, and dual 1600x1200 monitors in merged framebuffer, that the bug
info could be quite useful, so I should get filin'!  =8^)

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman in
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html


-- 
gentoo-dev@gentoo.org mailing list