I can't provide specifics yet, but I have a feeling our old friends,
random segfaults and "glibc detected" memory corruption, are back with
a vengeance.
Anybody else noticed it?
--
Csaba/Jester
--
All the data continuous
On Monday 17 December 2007 09:15:23 am Tim Moore wrote:
> Robert Black wrote:
> ...
>
> > I did a CVS update Sunday and couldn't compile it. Simgear updated and I
> > upgraded from OSG 2.0 to 2.2 with the same results.
>
> This looks like a problem with the way you upgraded OpenSceneGraph.
> OpenT
On Monday 17 December 2007 22:47, Georg Vollnhals wrote:
> Tim Moore schrieb:
> > Since we're talking about OSG now... this may be related to the
> > pager changes I checked in over the weekend. Could you post
> > some details on this problem of yours?
> >
> > Thanks,
> > Tim
>
> Hi Tim,
>
> as no
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Georg Vollnhals wrote:
> Tim Moore schrieb:
>> Since we're talking about OSG now... this may be related to the pager
>> changes I checked
>> in over the weekend. Could you post some details on this problem of yours?
...
> Hi Tim,
>
> as no other user
Tim Moore schrieb:
> Since we're talking about OSG now... this may be related to the pager changes
> I checked
> in over the weekend. Could you post some details on this problem of yours?
>
> Thanks,
> Tim
>
Hi Tim,
as no other user is complaining about this new "starting sub
ground-level" pro
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Robert Black wrote:
...
> I did a CVS update Sunday and couldn't compile it. Simgear updated and I
> upgraded from OSG 2.0 to 2.2 with the same results.
>
This looks like a problem with the way you upgraded OpenSceneGraph. OpenThreads
is now
an int
On Monday 17 December 2007 12:10:35 am Tim Moore wrote:
> Georg Vollnhals wrote:
> > the NEW problem I have here is that I start under the terrain level with
> > some a/c since a recent new CVS update (OSG and data).
> > Sometimes FlightGear just goes into an endless loop and hangs, sometimes
> > a
On Mon, 17 Dec 2007 07:10:35 +0100
Tim Moore <[EMAIL PROTECTED]> wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> Georg Vollnhals wrote:
>
> >
> > the NEW problem I have here is that I start under the terrain level with
> > some a/c since a recent new CVS update (OSG and data).
> >
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Georg Vollnhals wrote:
>
> the NEW problem I have here is that I start under the terrain level with
> some a/c since a recent new CVS update (OSG and data).
> Sometimes FlightGear just goes into an endless loop and hangs, sometimes
> a FlightGear RES
On Sun, 16 Dec 2007 17:19:18 -0800
Syd&Sandy <[EMAIL PROTECTED]> wrote:
> On Sun, 16 Dec 2007 16:02:02 -0800
> Syd&Sandy <[EMAIL PROTECTED]> wrote:
>
> > On Sun, 16 Dec 2007 15:50:06 -0800
> > Syd&Sandy <[EMAIL PROTECTED]> wrote:
> >
> > > On Mon, 17 Dec 2007 00:40:58 +0100
> > > Melchior FRANZ
On Sun, 16 Dec 2007 16:02:02 -0800
Syd&Sandy <[EMAIL PROTECTED]> wrote:
> On Sun, 16 Dec 2007 15:50:06 -0800
> Syd&Sandy <[EMAIL PROTECTED]> wrote:
>
> > On Mon, 17 Dec 2007 00:40:58 +0100
> > Melchior FRANZ <[EMAIL PROTECTED]> wrote:
> >
> > > m.
> > >
> >
> > probably the latter , I disabled
On Sun, 16 Dec 2007 15:50:06 -0800
Syd&Sandy <[EMAIL PROTECTED]> wrote:
> On Mon, 17 Dec 2007 00:40:58 +0100
> Melchior FRANZ <[EMAIL PROTECTED]> wrote:
>
> > m.
> >
>
> probably the latter , I disabled it to see if the Bravo would load , and it
> did in the water , but I must have reset it wr
On Mon, 17 Dec 2007 00:40:58 +0100
Melchior FRANZ <[EMAIL PROTECTED]> wrote:
> m.
>
probably the latter , I disabled it to see if the Bravo would load , and it did
in the water , but I must have reset it wrong ...
--
Syd&Sandy <[EMAIL PROTECTED]>
--
On Sun, 16 Dec 2007 17:38:42 -0600
"Curtis Olson" <[EMAIL PROTECTED]> wrote:
> On the surface, it looks like the problem involves the scenery loader so it
> may also help to know your location when the segfault happens.
>
> On the other hand, line #98 of FGTileLoader.cxx is this:
>
>
* Syd&Sandy -- Monday 17 December 2007:
> #0 0xb7b49b25 in __cxa_throw () from /usr/lib/libstdc++.so.6
> #1 0x084a5126 in FGTileLoader::add (this=0xadde67c, tile=0x117c4878)
> at FGTileLoader.cxx:98
OK, that's this code:
tile_path = globals->get_fg_scenery();
if (!tile_path.
On the surface, it looks like the problem involves the scenery loader so it
may also help to know your location when the segfault happens.
On the other hand, line #98 of FGTileLoader.cxx is this:
throw sg_throwable(string("No valid scenery path defined!"));
So you may want to double
On Sun, 16 Dec 2007 23:01:01 +0100
Melchior FRANZ <[EMAIL PROTECTED]> wrote:
> * Syd&Sandy -- Sunday 16 December 2007:
> > fgfd --Aircraft=Bravo --log-level=debug
>
> Works here. No segfault. (Just dead slow. :-)
>
>
>
> > Uncaught Exception: you should see a meaningful error message
>
> Try
Syd&Sandy schrieb:
> O
>
> Everything was fine until I did a CVS update about 10:00 am today...
>
>
Hi all,
the NEW problem I have here is that I start under the terrain level with
some a/c since a recent new CVS update (OSG and data).
Sometimes FlightGear just goes into an endless loop and ha
On dim 16 décembre 2007, Heiko Schulz wrote:
> > Here, with FG 1.0 both, Bravo and Blackbird are
> > right,
> > I get mainly problems with OSG.
> >
> > Regards
> >
> >
> >
> > --
> > Gérard
> > http://pagesperso-orange.fr/GRTux/
>
> FGFS 12/13/2007 windows precompiled
>
May be something differen
>
> Here, with FG 1.0 both, Bravo and Blackbird are
> right,
> I get mainly problems with OSG.
>
> Regards
>
>
>
> --
> Gérard
> http://pagesperso-orange.fr/GRTux/
>
>
FGFS 12/13/2007 windows precompiled
still in work: http://www.hoerbird.net/galerie.html
But already done: http://www.
* Syd&Sandy -- Sunday 16 December 2007:
> fgfd --Aircraft=Bravo --log-level=debug
Works here. No segfault. (Just dead slow. :-)
> Uncaught Exception: you should see a meaningful error message
Try this:
$ gdb --args fgfs --aircraft=Bravo --log-level=debug
(gdb) break main
(gdb) run
(g
On dim 16 décembre 2007, Heiko Schulz wrote:
> --- Syd&Sandy <[EMAIL PROTECTED]> schrieb:
> > Hi all ,
> > Just a note to say I'm getting a lot of segfaults in
> > the last few days ...
> > Maybe the Bravo is getting to big , but I get it
> > randomly with other aircraft too.
> > Im recompiling the
On Sun, 16 Dec 2007 22:24:41 +0100
Melchior FRANZ <[EMAIL PROTECTED]> wrote:
> * Syd&Sandy -- Sunday 16 December 2007:
> > [...] the Bravo also loads a bunch of models from the
> > Instruments-3d/primus folder , so I might have pushed it to the limit
>
> PLIB doesn't have problems with lots
* Syd&Sandy -- Sunday 16 December 2007:
> [...] the Bravo also loads a bunch of models from the
> Instruments-3d/primus folder , so I might have pushed it to the limit
PLIB doesn't have problems with lots of vertices/faces etc., but
only with too many per object, so this could only be trigge
On Sun, 16 Dec 2007 21:54:41 +0100
Melchior FRANZ <[EMAIL PROTECTED]> wrote:
> * Melchior FRANZ -- Sunday 16 December 2007:
> > No segfaults here. But PLIB has some hard-coded limits:
>
> Yeah, yeah ... "hard coded limits". Of course, it's only badly
> chosen data types, which are now kept for c
--- Melchior FRANZ <[EMAIL PROTECTED]> schrieb:
> * Melchior FRANZ -- Sunday 16 December 2007:
> > No segfaults here. But PLIB has some hard-coded
> limits:
>
> Yeah, yeah ... "hard coded limits". Of course, it's
> only badly
> chosen data types, which are now kept for
> compatibility reasons.
* Melchior FRANZ -- Sunday 16 December 2007:
> No segfaults here. But PLIB has some hard-coded limits:
Yeah, yeah ... "hard coded limits". Of course, it's only badly
chosen data types, which are now kept for compatibility reasons.
(And because Steve says that ought to be enough for anybody. :-)
* Syd&Sandy -- Sunday 16 December 2007:
> Just a note to say I'm getting a lot of segfaults in the last few days ...
> Maybe the Bravo is getting to big
No segfaults here. But PLIB has some hard-coded limits: number
of vertices per object, number of triangles, etc. (2^15 for each,
IIRC). Some air
--- Syd&Sandy <[EMAIL PROTECTED]> schrieb:
> Hi all ,
> Just a note to say I'm getting a lot of segfaults in
> the last few days ...
> Maybe the Bravo is getting to big , but I get it
> randomly with other aircraft too.
> Im recompiling the PLIB version to test again ,it's
> probably my fault.
Hi all ,
Just a note to say I'm getting a lot of segfaults in the last few days ...
Maybe the Bravo is getting to big , but I get it randomly with other aircraft
too.
Im recompiling the PLIB version to test again ,it's probably my fault.
Just wondering if anyone else is having this problem .
Che
On Wednesday 31 January 2007 07:35:15 Frederic Bouvier wrote:
> Nick,
>
> Selon Nick Warne :
> > Fred, Mathias
> >
> > I got this again tonight :-(((
> >
> > Chat [mpserve] GFS is now online, using
> >
> > Chat [mpserve] Aircraft/B-2/Models/b2-spirit.xml
> >
> > Chat [mpserve] GFS is now online, u
Nick,
Selon Nick Warne :
> Fred, Mathias
>
> I got this again tonight :-(((
>
> Chat [mpserve] GFS is now online, using
>
> Chat [mpserve] Aircraft/B-2/Models/b2-spirit.xml
>
> Chat [mpserve] GFS is now online, using
>
> Chat [mpserve] Aircraft/B-2/Models/b2-spirit.xml
>
> Segmentation fault
>
>
On Friday 26 January 2007 20:45:55 Nick Warne wrote:
> On Friday 26 January 2007 20:14, Maik Justus wrote:
> > Hi Fred,
> >
> > your patch works fine for me.
> >
> >
> > Big thanks!
> > Maik
>
> Fred,
>
> I tested with Maik while I was building. I just tested with ndim, and NO
> CRASH - it looks v
On Friday 26 January 2007 20:14, Maik Justus wrote:
> Hi Fred,
>
> your patch works fine for me.
>
>
> Big thanks!
> Maik
Fred,
I tested with Maik while I was building. I just tested with ndim, and NO
CRASH - it looks very, very good.
Thank you!
Nick
>
> Frederic Bouvier schrieb am 26.01.200
Hi Fred,
your patch works fine for me.
Big thanks!
Maik
Frederic Bouvier schrieb am 26.01.2007 20:41:
> Selon Nick Warne :
>
>
>> Reports from IRC say this is fixed in plib, but I can still get a 100%
>> reliable crash with OSG build.
>>
>> Enter MP server using Harrier. Get someone to join
Selon Nick Warne :
> Reports from IRC say this is fixed in plib, but I can still get a 100%
> reliable crash with OSG build.
>
> Enter MP server using Harrier. Get someone to join/leave/join with the
> dhc2F.
> Immediate segfault on the second join.
>
> Below is my earlier mail with trace etc.
On Friday 26 January 2007 19:08, Maik Justus wrote:
> Hi all,
>
> here is a workaround for osg. It disables the chrome shader, but as long
> as this causes the fatal crashes we need to commit.
>
> Unfortunately I am not compiling osg myself; I am using Olafs
> precompiled library. Therefore I can n
Hi all,
here is a workaround for osg. It disables the chrome shader, but as long
as this causes the fatal crashes we need to commit.
Unfortunately I am not compiling osg myself; I am using Olafs
precompiled library. Therefore I can not dig into the code :-(
Maik
Maik Justus schrieb am 26
On Friday 26 January 2007 09:02, Frederic Bouvier wrote:
> Hi Maik,
>
> Selon Maik Justus :
> > Hi,
> > the same thing with _condition.
> >
> > It's systematically done wrong.
> >
> > _textureData could be done correctly :-)
>
> _condition, _effectTexture and _textureData are not SGShaderAnimation
Hi Fred,
Frederic Bouvier schrieb am 26.01.2007 10:02:
> Hi Maik,
>
> Selon Maik Justus :
>
> _condition, _effectTexture and _textureData are not SGShaderAnimation members
> in
> the OSG code.
>
> In the PLIB branch, _contition is not shared but initialized with new in
> sgReadCondition, so un
Hi Maik,
Selon Maik Justus :
> Hi,
> the same thing with _condition.
>
> It's systematically done wrong.
>
> _textureData could be done correctly :-)
_condition, _effectTexture and _textureData are not SGShaderAnimation members in
the OSG code.
In the PLIB branch, _contition is not shared but i
Hi,
the same thing with _condition.
It's systematically done wrong.
_textureData could be done correctly :-)
Maik
Maik Justus schrieb am 25.01.2007 22:58:
> Hi all,
>
> I am not sure, but maybe we have the cause for the segfaults/failed
> assert (big thanks at Joacim for finding the point).
Hi all,
I am not sure, but maybe we have the cause for the segfaults/failed
assert (big thanks at Joacim for finding the point).
The _effectTexture is "loaded" by ssgTexture*
ssgLoaderOptions::createTexture. If the texture is already loaded, a
pointer to this texture is returned.
But SGShade
On Wednesday 24 January 2007 17:05, Joacim Persson wrote:
> On Fri, 5 Jan 2007, Maik Justus wrote:
> > Hello,
> >
> > sometimes I get a segfault in function
> > SGShaderAnimation::~SGShaderAnimation(), file
> > simgear/scene/model/shadanim.cxx.
>
> I never get a SIGSEGV from there, but rather a SIG
On Fri, 5 Jan 2007, Maik Justus wrote:
> Hello,
>
> sometimes I get a segfault in function
> SGShaderAnimation::~SGShaderAnimation(), file
> simgear/scene/model/shadanim.cxx.
I never get a SIGSEGV from there, but rather a SIGABORT due to a failed
assert. (This with plib, nota bene) And from the
Hi Fred,
no, it only helped to comment out the whole delete statement. I have to
check, if it was due to a local bug.
Maik
Frederic Bouvier schrieb am 15.01.2007 20:40:
> Mathias Fröhlich a écrit :
>
>> Hi Maik,
>>
>> On Friday 05 January 2007 11:16, Maik Justus wrote:
>>
>>
>>> sometim
Mathias Fröhlich a écrit :
> Hi Maik,
>
> On Friday 05 January 2007 11:16, Maik Justus wrote:
>
>> sometimes I get a segfault in function
>> SGShaderAnimation::~SGShaderAnimation(), file
>> simgear/scene/model/shadanim.cxx.
>>
>> Please commit the attached patch.
>>
> The patch you have att
Hi Mathias,
you are right. The problem were not zero pointers, but invalid pointers.
I have commented out the delete statements and got no further crash of
flightgear up to now (of course this is not the way to go, but to check
if my local crashes are to this delete statement). But I am not sur
Hi Maik,
On Friday 05 January 2007 11:16, Maik Justus wrote:
> sometimes I get a segfault in function
> SGShaderAnimation::~SGShaderAnimation(), file
> simgear/scene/model/shadanim.cxx.
>
> Please commit the attached patch.
The patch you have attached is a noop. you can safely delete a zero point
When I first looked at your patch, my first thought was that the original
code should have worked fine and your patch shouldn't really make any
difference. Maybe something else is corrupting the some structures ... an
out of bounds array write or something??? These are tough to track down,
but w
Hi,
I still get segfaults at this position (in the plib-branch). Commenting
the delete calls seem to stop them, but we get a memory leak this way.
Anyone else having segfaults?
Maik
Maik Justus schrieb am 05.01.2007 11:16:
> Hello,
>
> sometimes I get a segfault in function
> SGShaderAnimati
Hello,
sometimes I get a segfault in function
SGShaderAnimation::~SGShaderAnimation(), file
simgear/scene/model/shadanim.cxx.
Please commit the attached patch.
Maik
Index: simgear/scene/model/shadanim.cxx
===
RCS file: /var/cvs/
52 matches
Mail list logo
