bit tricky and
I'd have to worry about leaving corner cases in any rewritten code.
Reviewed-by: Keith Packard
--
keith.pack...@intel.com
pgpdUiBqksDv7.pgp
Description: PGP signature
___
xorg-devel mailing list
xorg-devel@lists.x.org
http://lis
On Tue, 15 Dec 2009 09:32:58 -0500, Adam Jackson wrote:
> Well, not devices, at least not for silken-enabled servers. But yes.
> And right now, it does; MaxClientFd ends up being getdtablesize()-1,
> which is all the file descriptors you can pass to select.
I was just thinking that 'MaxClientFd
; open below MAXCLIENTS that are not clients (input devices, mostly),
> which means you end up losing client slots.
>
> Rename MaxClients to MaxFd, to reflect its true nature, and don't clamp it
> to MAXCLIENTS.
>
> Signed-off-by: Adam Jackson
Reviewed-by: Keith Pa
#else clause.
Otherwise, it looks good to me.
Reviewed-by: Keith Packard
>
> Signed-off-by: Adam Jackson
> ---
> os/WaitFor.c|3 +--
> os/connection.c | 44
> os/io.c |2 +-
> os/osdep.h |6 +-
On Mon, 14 Dec 2009 15:19:49 -0500, Adam Jackson wrote:
> We can return from WaitForSomething with no clients ready for any number
> of reasons. There's no reason to set up the scheduler timer when this
> happens.
Looks like a fine plan.
Reviewed-by: Keith Packard
Bcc: kei...@keithp.comOn Thu, 17 Dec 2009 18:24:37 -0800, Alan Coopersmith
wrote:
> Default remains the same - on for most OS'es on i386 (except Solaris),
> off for everyone else. Can be manually toggled via --enable-pc98 or
> --disable-pc98.
This seems quite sensible.
A
Bcc: kei...@keithp.comOn Thu, 17 Dec 2009 18:24:38 -0800, Alan Coopersmith
wrote:
I thought these were already (effectively) reviewed?
--
keith.pack...@intel.com
pgpezjEk32hDe.pgp
Description: PGP signature
___
xorg-devel mailing list
xorg-devel@li
Bcc: kei...@keithp.comOn Thu, 17 Dec 2009 18:24:39 -0800, Alan Coopersmith
wrote:
> - [DEFAULT_FONT_PATH='catalogue:$(sysconfdir)/X11/fontpath.d'],
> + [DEFAULT_FONT_PATH='catalogue:${sysconfdir}/X11/fontpath.d'],
A classic confusion between make and shell s
On Fri, 11 Dec 2009 08:21:21 -0800, Alan Coopersmith
wrote:
> Convert all calls of CreateNewResourceType to pass name argument
I think all three of these patches are ready to be merged.
Reviewed-by: Keith Packard
--
keith.pack...@intel.com
pgpPMovLCjQ8z.pgp
Description: PGP signat
On Fri, 18 Dec 2009 15:16:27 -0800, Alan Coopersmith
wrote:
> That would require me to know, but I'm in the disinformed bunch. I
> vaguely remember it being a standard used in Japanese PC's that never
> caught on in the rest of the world - which of the choices in Wikipedia
> makes it sound mor
On Fri, 18 Dec 2009 13:32:36 -0800, Aaron Plattner wrote:
> This still seems like a silly reason to break the ABI to me, but
> whatev'.
This will make sure that existing and new code follow the rules; I
always appreciate it when the language provides mechanisms that let me
know about errors.
>
On Fri, 18 Dec 2009 16:54:35 -0800, Alan Coopersmith
wrote:
> Alan Coopersmith (3):
> Check for failures from CreateNewResourceType
> Ensure all resource types created have names registered
> Add type name argument to CreateNewResourceType
I also added an ABI bump to the video
o document this.
>
> Signed-off-by: Peter Hutterer
Acked-by: Keith Packard
--
keith.pack...@intel.com
pgpYhUrLCM47p.pgp
Description: PGP signature
___
xorg-devel mailing list
xorg-devel@lists.x.org
http://lists.x.org/mailman/listinfo/xorg-devel
On Mon, 21 Dec 2009 11:12:56 +1000, Peter Hutterer
wrote:
> This option isn't parsed by anything anymore.
>
> Signed-off-by: Peter Hutterer
Reviewed-by: Keith Packard
--
keith.pack...@intel.com
pgpTsO3ollfWU.pgp
Description:
On Mon, 21 Dec 2009 12:26:26 +1000, Dave Airlie wrote:
> This rolls up the following commits and also keeps the image_from_pict
> API compatiblity. It introduces a new image_from_pict_18 API that is
> used by the server leaving the old API alone.
>
> I've rolled this up as I don't want to introdu
On Mon, 21 Dec 2009 15:56:35 +1000, Peter Hutterer
wrote:
> Any input device with this option will be automatically added to whichever
> server layout is selected at startup.
> Signed-off-by: Peter Hutterer
Reviewed-by: Keith Packard
--
keith.pack...@intel.com
pgpmmmAE
On Mon, 28 Dec 2009 12:32:28 +0200, Oliver McFadden
wrote:
>
> Signed-off-by: Oliver McFadden
Reviewed-by: Keith Packard
This is an allocation failure path; one or both of these will have
failed to be allocated.
--
keith.pack...@intel.com
pgpgVp3pklacr.pgp
Description: PGP sig
-- a/dix/resource.c
> +++ b/dix/resource.c
> @@ -209,7 +209,10 @@ CreateNewResourceType(DeleteType deleteFunc, char *name)
> if (!funcs)
> return 0;
> if (!dixRegisterPrivateOffset(next, -1))
> +{
> + xfree (funcs);
> return 0;
> +}
Re
On Mon, 28 Dec 2009 12:02:56 +0200, Oliver McFadden
wrote:
> freed_arg: Pointer "srcbits" and "mskbits" freed by function
> "AllocARGBCursor"
This one looks wrong to me.
--
keith.pack...@intel.com
pgpv4GTW2eSF5.pgp
Description: PGP signature
___
xo
On Mon, 28 Dec 2009 14:02:02 -0500, Matt Turner wrote:
> The fix here isn't to add back in a bunch of duplicated Alpha
> assembly.
Correct, the right fix is to solve the problem. However, I'd like to see
that solution come in a single patch instead of two separated by weeks
of time. Reverting th
On Sat, 2 Jan 2010 20:32:50 +0600, Mikhail Gusarov
wrote:
> More janitorial stuf, plus substantial changes:
Can you get a Reviewed-by: before Cc:'ing me? Whitespace changes are one
thing, but actual code changes need external review before I'm
comfortable applying them.
--
keith.pack...@inte
On Sat, 2 Jan 2010 11:58:40 +, Julien Cristau wrote:
> --- a/dix/dispatch.c
> +++ b/dix/dispatch.c
> @@ -3517,6 +3517,11 @@ int ProcForceScreenSaver(ClientPtr client)
> rc = dixSaveScreens(client, SCREEN_SAVER_FORCER, (int)stuff->mode);
> if (rc != Success)
> return rc;
> +
On Mon, 4 Jan 2010 09:26:06 +1000, Peter Hutterer
wrote:
> any particular reason this patch is needed? It touches a lot of files but
> doesn't actually improve the code.
I checked with git diff -b and saw that it changed nothing; as it
just eliminates a pile of trailing white space. Makes it ea
On Mon, 4 Jan 2010 15:32:49 -0500, Adam Jackson wrote:
> +static int
> +miScanLineWidth(
Is there any driver still using this function? I'd rather see it
disappear than live on without having any users.
Also, if it does need to survive, the copyright messages from mibank.c
will need to get mer
On Mon, 04 Jan 2010 17:21:49 -0500, Adam Jackson wrote:
> No driver uses it. But xf86ValidateModes (alone, hence static) does,
> which is why I moved it to xf86Mode.c. And does so in more than one
> place, so the factorization still makes sense.
Sigh. And we still have drivers using xf86Valdia
On Sun, 03 Jan 2010 17:37:48 +0100, Bernard Siaud alias Troumad
wrote:
> 3) I don(t know what is keystone shaping ! It's not easy for me to
> translate it in french !
Here's the English wikipedia article on the subject; I'm afraid the only
other language linked from that page is German though.
On Tue, 5 Jan 2010 13:46:07 +1000, Peter Hutterer
wrote:
> @@ -630,9 +630,12 @@ updateFromMaster(EventListPtr events, DeviceIntPtr dev,
> int type, int *num_event
> if (master && master->last.slave != dev)
> {
> CreateClassesChangedEvent(events, master, dev, type);
> -
On Tue, 5 Jan 2010 16:15:20 +1000, Peter Hutterer
wrote:
> yes, you're right, thanks. This one addresses this, moves the last.slave out
> of the condition. The numValuators assignment is not needed afaict, in fact
> it doesn't seem to be used anywhere anyway.
If it isn't used, can we eliminate
On Tue, 05 Jan 2010 14:20:31 -0500, Adam Jackson wrote:
> Two months is long enough to wait for objections, I think. Keith,
> please merge?
Thanks for the ping.
--
keith.pack...@intel.com
pgphCvqlAH8jA.pgp
Description: PGP signature
___
xorg-devel
On Wed, 6 Jan 2010 07:41:41 +1000, Peter Hutterer
wrote:
> I think so though I have to check my notes again to be sure. since the
> DeviceIntRec is exported ABI, I'll do that in a separate patch though.
Sure. ABI isn't an issue for 1.8 as we're already bumping all of the
module ABI versions. Bu
On Wed, 06 Jan 2010 16:12:15 -0500, Adam Jackson wrote:
> a) Enabling support for the Composite stack in Xinerama mode
Sure. Just bug fixes for fixes/render/composite. From my brief glance,
these all look like reasonable patches.
> b) Exposing animated cursors to the DDX for acceleration
Agree
On Wed, 06 Jan 2010 17:39:15 -0500, Adam Jackson wrote:
> I described this wrong. It corrects the RANDR geometry to account for
> the Xin muxer's idea of the root window size, but XRRGetScreenResources
> still only gives you the RANDR objects on screen 0. Not useful
> enough.
Right, RandR gets
On Thu, 07 Jan 2010 10:04:45 -0500, Adam Jackson wrote:
> It doesn't _have_ to work that way, but that's what it looks like DMX2
> does.
Ok, so just something we should work on fixing then.
> If we consider a non-transformative, non-GL compositor, then I think
> it's pretty straightforward. Pi
On Sun, 27 Sep 2009 01:08:42 +1000, Peter Hutterer
wrote:
>I think 3 + 2 + 1 months should be approporiate for the various stages.
Following Peter's proposed schedule, we're now in the 'bugfix' phase for
server version 1.8. Any major feature changes not incorporated in the
server at this po
On Thu, 07 Jan 2010 23:29:22 +0100, Rémi Cardona wrote:
> I think you have enough Acked-bys and Reviewed-bys here. Maybe you
> should put those in a branch and ask Keith to pull all three of them?
I'm good with either [PATCH] or [PULL] requests, although I'm trying to
encourage people to flag th
On Fri, 8 Jan 2010 15:27:27 +1000, Peter Hutterer
wrote:
> Eventually the callbacks should be updated to work with InternalEvents (only
> RECORD actually uses the callbacks). Currently, WriteEventsToClient deals
> with wire events and changing it to keep the InternalEvent constitutes an
> ABI br
On Fri, 8 Jan 2010 15:18:41 +1000, Peter Hutterer
wrote:
> This value isn't actually set for normal events but it saves us some work
> for the record extension support.
>
> Signed-off-by: Peter Hutterer
Reviewed-by: Keith Packard
--
keith.pack...@intel.com
pgpw85Xte2m2
On Fri, 8 Jan 2010 12:31:35 -0800, Jeremy Huddleston
wrote:
> > fbGetDrawablePixmap (pict->pDrawable, pixmap, *xoff, *yoff);
fbGetDrawablePixmap is a macro which stores through *xoff and *yoff.
--
keith.pack...@intel.com
pgpyTi1iWTvm6.pgp
Description: PGP signature
_
On Fri, 08 Jan 2010 19:14:16 -0500, Gaetan Nadon wrote:
> Using git send-email, the CC to "Keith Packard " did
> not work. I used the same command to send a patch to the list and CC
> Alan and it did work. Maybe someone had this problem before.
Hrm. I got the patch and pus
On Sun, 10 Jan 2010 16:06:51 -0500, Gaetan Nadon wrote:
> The dmx and scaled text files are generated from sgml files.
> The linuxdoc program is used to generated them.
We've generally been distributing the .txt files with the .tar.gz
package so that people don't have to figure out how to get li
On Mon, 11 Jan 2010 11:40:53 -0500, Gaetan Nadon wrote:
> That will help distcheck but not dist and only if you have doxygen
> installed (some platforms don't). I first propose to solve the root
> cause of the problem, and if that's not feasible, then patch some of the
> symptoms.
I'd be fine wi
On Tue, 12 Jan 2010 14:53:42 -0500, Eamon Walsh wrote:
> Ping?
Sorry, I didn't see this before; might well have been spamassassin
before I got the 20XX bug fixed.
They'll need a bit of review before merging; I'll try to get that done
tomorrow.
--
keith.pack...@intel.com
pgpxn60nUMyrz.pgp
De
d drivers in
> DRI2ScreenInit.
>
> Signed-off-by: Aaron Plattner
This looks good to me, but I'd like krh to also take a look, then I'll
merge it in.
Reviewed-by: Keith Packard
--
keith.pack...@intel.com
pgpoF2RjQIBgT.pgp
Description: PGP signature
___
On Mon, 14 Jun 2010 18:00:23 +0300, Tiago Vignatti
wrote:
> FYI, we have also the privates implementation in Xorg DDX (privates field of
> ScrnInfoRec). I don't see any particular reason to have another privates
> mechanism inside Xorg given we have already the one in dix, so we might want
> to
On Thu, 17 Jun 2010 15:22:41 +0300, Tiago Vignatti
wrote:
> - DevToConfig[i].GDev.vendor = (char *)VendorName;
> - DevToConfig[i].GDev.board = (char *)CardName;
> -
You're removing the assignment here
> -/*ptr->dev_identifier = DevToConfig[screennum].GDev.identifier;*/
> ptr-
On Thu, 17 Jun 2010 22:19:04 -0400, Adam Goode wrote:
> https://bugs.freedesktop.org/show_bug.cgi?id=20334
Disjoint areas have always been the issue with this bug; not sure this
solves it?
--
keith.pack...@intel.com
pgpnfXWFdOKB3.pgp
Description: PGP signature
___
On Fri, 18 Jun 2010 11:03:49 -0400, Adam Goode wrote:
> On 06/18/2010 03:08 AM, Keith Packard wrote:
> > On Thu, 17 Jun 2010 22:19:04 -0400, Adam Goode wrote:
> >
> >> https://bugs.freedesktop.org/show_bug.cgi?id=20334
> >
> > Disjoint areas have always be
On Tue, 15 Jun 2010 23:25:34 +0300, ville.syrj...@nokia.com wrote:
> Fix the problem by sending port notifications only when
> SetPortAttribute succeeds.
Reviewed-by: Keith Packard
--
keith.pack...@intel.com
pgpnxjjENkWg2.pgp
Description: PGP sig
On Mon, 14 Jun 2010 10:44:38 -0700, Jamey Sharp wrote:
> These cleanups and warning fixes have been reviewed and don't, as far as
> I can tell, affect API or ABI. Can they be merged to master?
I don't know. We're past RC1 and are supposed to be working on
regressions and other bug fixing.
What
On Mon, 14 Jun 2010 09:25:22 -0400, Kristian Høgsberg
wrote:
> - pos = next, next = __container_of(next->member.next, next, member))
> + pos = tmp, tmp = __container_of(pos->member.next, tmp, member))
Any reason this uses a ',' instead of a ';'?
On Sat, 19 Jun 2010 15:22:50 -0700, Jamey Sharp wrote:
> Three months does feel like a
> long time to stall development on master, though, given how
> attention-starved this code already is. I'd prefer a three-month total
> release cycle, assuming a willing release manager.
Note that 1.9 is curr
On Mon, 21 Jun 2010 14:56:45 +0100, Daniel Stone wrote:
> As I said before, I'm pretty concerned about this. We have enough
> trouble supporting our stable releases as it is (though Peter's doing a
> fairly heroic job single-handedly holding up 1.8.x), and changing to
> three months just makes t
On Mon, 21 Jun 2010 07:53:30 -0700, Alan Coopersmith
wrote:
> Keith Packard wrote:
> > Beyond that, I'd like for people to consider the tree closed to
> > large-scale changes as we prepare for the 1.9 release. There are a
> > couple of additional cleanup patch seque
me as our coordinate system has integers in the
upper left corner of each pixel, so to get from one upper left location
to another, you need to get to the middle of the pixel *before* the
transform instead of afterwards.
Reviewed-by: Keith Packard
--
keith.pack...@intel.com
pgpxkDRK3snKh.pgp
Descrip
On Sun, 20 Jun 2010 20:39:59 -0400, Matt Turner wrote:
> Really, the headers need a massive clean-up.
It would be good to start with some idea of how things 'should'
work. Afaict, there's no rhyme or reason to how and where various system
include files are referenced from our current sources.
-
On Mon, 21 Jun 2010 14:08:50 +1000, Dave Airlie wrote:
> I built latest server and copied over a mesa 7.8 dri2 driver to use with it,
> and it crashed in here starting compiz. Looks like we need to validate the
> flush extension is registered before calling invalidate.
This is the only place scr
ys hit one of the freeing
> paths or else its a bug in the DIX.
This looks right to me -- pulling the piece of the old
MoveWindow/ResizeWindow/ChangeBorderWidth functions into this new
compFreeOldPixmap function. I'd say using these hooks is fine;
ConfigureWindow will always call one of t
On Mon, 21 Jun 2010 17:38:07 +0100, Simon Farnsworth
wrote:
>
> +if (!((OsCommPtr)client->osPrivate)->trans_conn)
> +return FALSE;
> +
This looks almost right to me -- I reviewed the os code to see when
_XSERVTransClose is called (which is what frees the trans_conn data) and
found
On Mon, 21 Jun 2010 16:04:51 +0300, Tiago Vignatti
wrote:
> Yes, we have significant work happening right now: clean-up. It's a feature as
> desirable as any other.
I'm not saying there isn't significant and useful work going on, but
that we aren't seeing long-term development and testing happe
On Tue, 22 Jun 2010 10:13:30 +0100, Simon Farnsworth
wrote:
> Keith, does this look better to you?
Yup, seems fine now.
Reviewed-by: Keith Packard
> +if (!client->osPrivate)
> +return FALSE;
> +if (!((OsCommPtr)client->osPrivate)->trans_conn)
&g
ching any invalid pointers.
Krh may yet convince me that the generic list functions might well be a
good plan for things like this :-)
Reviewed-by: Keith Packard
--
keith.pack...@intel.com
pgpdpeAw4iXh7.pgp
Description: PGP signature
___
xorg-deve
On Sat, 19 Jun 2010 13:05:22 +0300, Tiago Vignatti
wrote:
> Although vendor and board naming are used to create the configure file, the
> server doesn't actually use it when fetching such file and probing devices.
>
> Reported-by: Richard Barnette
> Signed-off-by: Tiago Vignatti
> ---
> from v
On Sat, 19 Jun 2010 14:02:02 +0300, "Vignatti Tiago (Nokia-D/Helsinki)"
wrote:
> ping?
Yeah, this all seems reasonable. But, as I'm trying to get release
candidates made frequently, and those can't depend on unreleased
modules. It doesn't look like libXfont 1.4.2 has been released yet;
there's
On Sat, 19 Jun 2010 12:58:45 -0400, James Willcox wrote:
> Ok, is there anything about these patches that you don't like? What can I
> do to get this in to $XORG_NEXT?
We're already past 1.9 RC1, so we'd be looking to X server 1.10 in any
case.
If you want a constraint rectangle, then I'd sugg
On Wed, 23 Jun 2010 16:24:29 +0300, Tiago Vignatti
wrote:
> ABI break.
Doesn't this also break anhy config files that include vendorname or
boardname elements?
--
keith.pack...@intel.com
pgp8bG6SYgdb1.pgp
Description: PGP signature
___
xorg-devel@
itionally.
> XTS test results for XGrabButton are identical with and without this
> patch.
Sure would be nice to have a test case which checks this behaviour...
> This reverts commit 48585bd1e3e98db0f3df1ecc68022510216e00cc.
>
> Conflicts:
>
> dix/events.c
>
>
On Fri, 25 Jun 2010 10:37:47 +1000, Peter Hutterer
wrote:
> fwiw, with this patch, XGrabButton test 5/39 now PASSes instead of FAIL.
> Of course, you have to run all 39 manually, one-by-one to notice, because
> earlier tests don't always reset the modifier state and later tests then
> fail when
On Mon, 28 Jun 2010 14:52:21 +0300, "Vignatti Tiago (Nokia-D/Helsinki)"
wrote:
> On Thu, Jun 24, 2010 at 08:07:07AM +0200, ext Keith Packard wrote:
> > On Wed, 23 Jun 2010 16:24:29 +0300, Tiago Vignatti
> > wrote:
> >
> > > ABI break.
> >
>
On Mon, 28 Jun 2010 15:58:46 -0700, Jesse Barnes
wrote:
> When ResetCurrentRequest is called, or IgnoreClient is called when a
> client has input pending, IgnoredClientsWithInput will be set. However,
> a subsequent IgnoreClient request will clear the client fd from that fd
> set, potentially c
On Mon, 28 Jun 2010 16:59:03 -0700, Jesse Barnes
wrote:
> On Mon, 28 Jun 2010 16:55:30 -0700
> Keith Packard wrote:
>
> > On Mon, 28 Jun 2010 15:58:46 -0700, Jesse Barnes
> > wrote:
> >
> > > When ResetCurrentRequest is called, or IgnoreClient is called w
On Tue, 29 Jun 2010 02:18:31 +0200, Mario Kleiner
wrote:
> What i assume but didn't check is that xlib doesn't have a problem
> with getting replies out of sequence
The X protocol is purely sequential -- you can't execute requests out of
order (and queuing replies is part of the request execu
hat the count starts out at zero?
Reviewed-by: Keith Packard
--
keith.pack...@intel.com
pgpDZK7cQXEjL.pgp
Description: PGP signature
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: http://lists.x.org/mailman/listinfo/xorg-devel
On Mon, 28 Jun 2010 19:14:41 -0700, Jesse Barnes
wrote:
> Thanks. Do you want to apply it with that change or shall I send an
> update?
Update with Rb line attached would be most excellent. We may want to see
if others have comments before I apply it to master.
--
keith.pack...@intel.com
p
On Tue, 29 Jun 2010 14:05:52 +1000, Peter Hutterer
wrote:
> +
> +/* Merge the lockedPtrButtons from all attached SDs for the given master
> + * device into the MD's state.
> + */
> +void
> +XkbMergeLockedPtrBtns(DeviceIntPtr master)
> +{
> +DeviceIntPtr d = inputInfo.devices;
> +XkbSrvIn
On Tue, 29 Jun 2010 16:08:03 +0300, Tiago Vignatti
wrote:
> diff --git a/hw/xfree86/doc/man/xorg.conf.man.pre
> b/hw/xfree86/doc/man/xorg.conf.man.pre
> index 6b3636f..3443e9b 100644
> --- a/hw/xfree86/doc/man/xorg.conf.man.pre
> +++ b/hw/xfree86/doc/man/xorg.conf.man.pre
> @@ -1455,9 +1455,6 @
used by software that was used prior to
> starting X.
This looks good to me -- initial_rotation will be set to zero when the
output is created (via calloc) and isn't set otherwise in the X server.
Reviewed-by: Keith Packard
--
keith.pack...@intel.com
pgpy2s
Initializing the dev privates code after allocating the server client
dev privates would cause the memory leak check to fire at server
startup or reset.
Signed-off-by: Keith Packard
---
This eliminates the memory leak warning which prompted Tiago's change
to print those only in debug b
On Fri, 25 Jun 2010 11:34:44 +0100, Julien Cristau wrote:
> Signed-off-by: Julien Cristau
Wow. What did that code do before? It looks like it was completely
broken.
This appears to do something sensible at least.
Reviewed-by: Keith Packard
--
keith.pack...@intel.com
pgp4o0QJzsV8f.
On Wed, 30 Jun 2010 19:30:51 +0300, Tiago Vignatti
wrote:
> On Wed, Jun 30, 2010 at 05:39:34PM +0200, ext Keith Packard wrote:
> > Initializing the dev privates code after allocating the server client
> > dev privates would cause the memory leak check to fire at server
>
MiDbeScreenPrivPrivRec is not used in the server. Remove it, along
with the MI_DBE_SCREEN_PRIV_PRIV macro that tried to use it.
Signed-off-by: Keith Packard
---
dbe/midbestr.h | 16
1 files changed, 0 insertions(+), 16 deletions(-)
diff --git a/dbe/midbestr.h b/dbe
x
that, if not, here it is:
From 4f3397774a48a214755740f846e1eb8995a744bb Mon Sep 17 00:00:00 2001
From: Keith Packard
Date: Fri, 30 Apr 2010 13:33:14 -0700
Subject: [PATCH] Destroy screen pixmap on screen close.
This avoids a memory leak on server reset.
Signed-off-by: Keith Packard
---
uxa/uxa.c
, and (in the new
dixPrivate world) causes an assert failure.
Signed-off-by: Keith Packard
---
dbe/midbe.c |6 --
1 files changed, 0 insertions(+), 6 deletions(-)
diff --git a/dbe/midbe.c b/dbe/midbe.c
index 954f46e..b43ac1b 100644
--- a/dbe/midbe.c
+++ b/dbe/midbe.c
@@ -202,12 +
On Fri, 2 Jul 2010 09:00:43 +1000, Peter Hutterer
wrote:
> Keith,
>
> please pull the fixes below into master. They fix three different cases
> where a button or grab may get stuck, leading to an unresponsive
> desktop.
Argh. I messed up and merged some of the patches to get them into
RC4. Mea
On Fri, 2 Jul 2010 13:38:53 +1000, Peter Hutterer
wrote:
> Unless I specifically tell you, simply ignore any patches that are from me
> or input-related. I do try to send you pull requests, even if it's just one
> or two that need to go on master.
Ok, cool. Please feel free to chime in on patch
On Fri, 2 Jul 2010 14:51:11 +1000, Dave Airlie wrote:
> From: Dave Airlie
>
> If you didn't call the temp variable for the macro, next, it wouldn't
> actually work.
This was fixed in 2c0159d21788b16607e6ebe4571b57fe78545c2a (one of the
two patches in RC3).
--
keith.pack...@intel.com
pgpqUX
On Fri, 2 Jul 2010 15:31:19 +1000, Peter Hutterer
wrote:
> Having miPointerMove and miPointerMoved is confusing, especially since both
> do the same thing bar the event delivery. Also, miPointerMove calls
> miPointerMoved which indicates some confusion in the temporal alignment of
> cause and eff
On Fri, 2 Jul 2010 09:43:18 +0100, Chris Wilson
wrote:
> As pixman does not guard against a fill request outside of the buffer,
> we must be be careful and trim oversized fills.
Why isn't the clipping in fbPolyFillRect sufficient? fbFill should never
be given values beyond the clip region.
--
On Fri, 2 Jul 2010 12:26:34 -0700, Jesse Barnes
wrote:
> If a pixmap header is modified or the drawable serial changes, some
> aspects of the drawable are likely to have changed so we should
> re-allocate the corresponding DRI2 drawable in that case. This is one
> way of catching when the root w
On Fri, 2 Jul 2010 12:26:34 -0700, Jesse Barnes
wrote:
> If a pixmap header is modified or the drawable serial changes, some
> aspects of the drawable are likely to have changed so we should
> re-allocate the corresponding DRI2 drawable in that case. This is one
> way of catching when the root
On Sun, 04 Jul 2010 22:16:55 +1000, Peter Hutterer
wrote:
> I think it's fine as it is. moving the pointer always generates an
> event, the exception being miPointerMoveNoEvent, explicitly labeled
> so.
Sounds good to me.
Reviewed-by: Keith Packard
I will wait to see this in
On Mon, 5 Jul 2010 15:30:44 +0300, "Vignatti Tiago (Nokia-MS/Helsinki)"
wrote:
> I don't have time now to keep digging on it but seems that the fix should be
> inside XAA, isn't?
That would be ideal, but I don't think it will work -- as discussed in
the thread on the intel driver patch, there m
On Mon, 5 Jul 2010 23:44:17 -0700, Jamey Sharp wrote:
> Those changes were released in xserver 1.7, apparently. This patch
> looks to me like a low-risk fix to a longstanding bug, so should it
> perhaps be applied to 1.9?
We may find some client bugs this way, but I think it's appropriate to
hav
On Wed, 7 Jul 2010 08:09:31 +1000, Peter Hutterer
wrote:
> miPointerMoved already has the same code.
It's got a bit more -- it only does SIGIO-time pointer motion on the
VCP. Seems sensible enough to me though.
Reviewed-by: Keith Packard
--
keith.pack...@intel.com
pgpcRHadi
On Wed, 7 Jul 2010 12:38:10 +0200, Michel Dänzer wrote:
> From: Michel Dänzer
>
> Previously we assumed every pixmap destroyed during a software fallback was
> also created during a software fallback and had access prepared, but that's
> not always true.
>
> Fixes a server abort reported by 邓逸
On Wed, 7 Jul 2010 16:11:34 -0700, Aaron Plattner wrote:
> Sadly, moving it to the end doesn't look like it'll be enough:
>
> dix.h:extern _X_EXPORT ClientPtr clients[MAXCLIENTS];
That's just an array of pointers.
--
keith.pack...@intel.com
pgp8FvTodHl5I.pgp
Description: PGP signature
__
On Wed, 7 Jul 2010 12:59:39 +0300, Tiago Vignatti
wrote:
> So what's the idea?
I think Peter is working on fixing some stuck-key issues, and that these
patches are related to that work. Peter? Am I merging stuff I shouldn't be?
--
keith.pack...@intel.com
pgpxcIaD9LFK0.pgp
Description: PGP s
On Thu, 8 Jul 2010 15:00:33 +0100, Daniel Stone wrote:
> +if (strncmp(s, "XF86_", 5) == 0) {
> +KeySym ret;
> +char *tmp = strdup(s);
Please check malloc returns in Xlib...
Otherwise,
Reviewed-by: Keith Packard
--
keith.pack...@intel.com
pgp8ndxs3A
On Thu, 8 Jul 2010 16:50:48 +0100, Daniel Stone wrote:
> Jesus, how embarassment. Is the following OK?
Yup, looks good.
> @@ -159,6 +159,8 @@ XStringToKeysym(_Xconst char *s)
> if (strncmp(s, "XF86_", 5) == 0) {
> KeySym ret;
> char *tmp = strdup(s);
> +if (!tmp)
On Thu, 08 Jul 2010 08:52:42 +0200, Michel Dänzer wrote:
> I considered that but decided to keep the fix minimal for now.
I'd rather see the code cleaned up; it's not like someone is going to
come back later to fix it, and leaving it as is risks having a bug fixed
in one place at not the other t
On Thu, 8 Jul 2010 11:39:01 -0700, Jesse Barnes
wrote:
> If the root window changes size, we need to re-allocate any DRI2
> buffers that share the same pixmap so that they'll pick up the new root
> pixmap (if the underlying driver ended up creating a new one).
Why is this only relevant at the s
601 - 700 of 2003 matches
Mail list logo
