subject:"Re\: \[hugin\-ptx\] CRITICAL\: join us to nuke the pesky bug that is plaguing the fast preview"

2011/8/4 Gnome Nomad gnomeno...@gmail.com


 Hmm, why is a graphics app trying to use PulseAudio? I don't even have it
 installed on any of my computers here!



Exactly what I thought. Can you issue a ldd binary, like ldd
/usr/local/bin/Hugin to display all linked libraries for hugin and do the
same for all hugin libs in /usr/local/lib, something  like ldd
/usr/local/lib/libhugin*.so (or wherever you installed them).

Maybe pulseaudio is linked by one of the dependent libraries used by hugin.

Harry

-- 
You received this message because you are subscribed to the Google Groups 
Hugin and other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-04 Thread Terry Duell


Hullo Harry,

On Thu, 04 Aug 2011 20:05:18 +1000, Harry van der Wolf hvdw...@gmail.com  
wrote:



2011/8/4 Gnome Nomad gnomeno...@gmail.com



Hmm, why is a graphics app trying to use PulseAudio? I don't even have  
it

installed on any of my computers here!




Exactly what I thought. Can you issue a ldd binary, like ldd
/usr/local/bin/Hugin to display all linked libraries for hugin and do  
the

same for all hugin libs in /usr/local/lib, something  like ldd
/usr/local/lib/libhugin*.so (or wherever you installed them).

Maybe pulseaudio is linked by one of the dependent libraries used by  
hugin.


In Fedora, and probably all Linux, hugin is installed in /usr/bin and as  
my build is on an x86_64 the libs are installed in /usr/lib64/hugin.
I saved the output from 'ldd /usr/bin/hugin' and 'ldd  
/usr/lib64/hugin/lib*.so.*' in the file 'hugin-linked-libs.txt' which I  
have attached.

I hope I have done that correctly, if not please let me know.

Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
[terry@phenom ~]$ ldd /usr/bin/hugin
linux-vdso.so.1 =  (0x7fff3ae07000)
libhuginbase.so.0.0 = /usr/lib64/hugin/libhuginbase.so.0.0 
(0x00368800)
libboost_thread-mt.so.1.46.0 = /usr/lib64/libboost_thread-mt.so.1.46.0 
(0x00368760)
libboost_date_time-mt.so.1.46.0 = 
/usr/lib64/libboost_date_time-mt.so.1.46.0 (0x00368720)
libboost_regex-mt.so.1.46.0 = /usr/lib64/libboost_regex-mt.so.1.46.0 
(0x003688c0)
libboost_filesystem-mt.so.1.46.0 = 
/usr/lib64/libboost_filesystem-mt.so.1.46.0 (0x00368940)
libboost_iostreams-mt.so.1.46.0 = 
/usr/lib64/libboost_iostreams-mt.so.1.46.0 (0x003689c0)
libboost_system-mt.so.1.46.0 = /usr/lib64/libboost_system-mt.so.1.46.0 
(0x00368900)
libboost_signals-mt.so.1.46.0 = 
/usr/lib64/libboost_signals-mt.so.1.46.0 (0x003685c0)
libpano13.so.2 = /usr/lib64/libpano13.so.2 (0x003b7420)
libGLEW.so.1.5 = /usr/lib64/libGLEW.so.1.5 (0x00368680)
libhuginvigraimpex.so.0.0 = /usr/lib64/hugin/libhuginvigraimpex.so.0.0 
(0x00368640)
libImath.so.6 = /usr/lib64/libImath.so.6 (0x00368400)
libIlmImf.so.6 = /usr/lib64/libIlmImf.so.6 (0x00368580)
libIex.so.6 = /usr/lib64/libIex.so.6 (0x00368540)
libHalf.so.6 = /usr/lib64/libHalf.so.6 (0x00368480)
libIlmThread.so.6 = /usr/lib64/libIlmThread.so.6 (0x00368380)
libjpeg.so.62 = /usr/lib64/libjpeg.so.62 (0x003b74a0)
libtiff.so.3 = /usr/lib64/libtiff.so.3 (0x003b74e0)
libpng12.so.0 = /usr/lib64/libpng12.so.0 (0x003b71e0)
libz.so.1 = /lib64/libz.so.1 (0x003d3ac0)
libexiv2.so.10 = /usr/lib64/libexiv2.so.10 (0x003686c0)
libhuginbasewx.so.0.0 = /usr/lib64/hugin/libhuginbasewx.so.0.0 
(0x00368c40)
libceleste.so.0.0 = /usr/lib64/hugin/libceleste.so.0.0 
(0x00368cc0)
libicpfindlib.so.0.0 = /usr/lib64/hugin/libicpfindlib.so.0.0 
(0x00368d00)
libwx_baseu-2.8.so.0 = /usr/lib64/libwx_baseu-2.8.so.0 
(0x00368ac0)
libwx_gtk2u_core-2.8.so.0 = /usr/lib64/libwx_gtk2u_core-2.8.so.0 
(0x00368a00)
libwx_gtk2u_aui-2.8.so.0 = /usr/lib64/libwx_gtk2u_aui-2.8.so.0 
(0x00368b80)
libwx_gtk2u_xrc-2.8.so.0 = /usr/lib64/libwx_gtk2u_xrc-2.8.so.0 
(0x00368c00)
libwx_gtk2u_html-2.8.so.0 = /usr/lib64/libwx_gtk2u_html-2.8.so.0 
(0x00368b00)
libwx_baseu_xml-2.8.so.0 = /usr/lib64/libwx_baseu_xml-2.8.so.0 
(0x00368b40)
libwx_gtk2u_adv-2.8.so.0 = /usr/lib64/libwx_gtk2u_adv-2.8.so.0 
(0x00368bc0)
libwx_gtk2u_gl-2.8.so.0 = /usr/lib64/libwx_gtk2u_gl-2.8.so.0 
(0x00368c80)
libwx_baseu_net-2.8.so.0 = /usr/lib64/libwx_baseu_net-2.8.so.0 
(0x00368a80)
libGLU.so.1 = /usr/lib64/libGLU.so.1 (0x003683c0)
libGL.so.1 = /usr/lib64/nvidia/libGL.so.1 (0x003687c0)
libSM.so.6 = /usr/lib64/libSM.so.6 (0x003b73e0)
libICE.so.6 = /usr/lib64/libICE.so.6 (0x003d4380)
libX11.so.6 = /usr/lib64/libX11.so.6 (0x00367f00)
libXext.so.6 = /usr/lib64/libXext.so.6 (0x00367f40)
libglut.so.3 = /usr/lib64/libglut.so.3 (0x00368600)
libXmu.so.6 = /usr/lib64/libXmu.so.6 (0x00368200)
libXi.so.6 = /usr/lib64/libXi.so.6 (0x00367f80)

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-04 Thread Wirz

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

Moin,


 Hmm, why is a graphics app trying to use PulseAudio? I don't even have it
 installed on any of my computers here!

 Exactly what I thought. Can you issue a ldd binary, like ldd
 /usr/local/bin/Hugin to display all linked libraries for hugin and do the
 same for all hugin libs in /usr/local/lib, something  like ldd
 /usr/local/lib/libhugin*.so (or wherever you installed them).
 
 Maybe pulseaudio is linked by one of the dependent libraries used by hugin.


I believe the easiest way to check is saying

ldd $(which hugin)| awk '{print $3}'| xargs ldd | grep Audio

In case of my setup (debian squeeze, hugin 2011.3.0.5460 compiled from
trunk, proprietary NV driver) there is no result ...

cheers, lukas
- -- 
Do not believe the naysayers who say it cannot be done.
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk46i+8ACgkQ1m2kmG8CIy16ygCgthXLZBnr/NaJBY6T1OjvCLaK
QrkAn1UT9P9Th5AKjBx5AXU0iUUcHWS8
=pZw+
-END PGP SIGNATURE-

-- 
You received this message because you are subscribed to the Google Groups 
Hugin and other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011/8/4 Terry Duell tdu...@iinet.net.au

 Hullo Harry,


 On Thu, 04 Aug 2011 20:05:18 +1000, Harry van der Wolf hvdw...@gmail.com
 wrote:

  2011/8/4 Gnome Nomad gnomeno...@gmail.com


 Hmm, why is a graphics app trying to use PulseAudio? I don't even have it
 installed on any of my computers here!



  Exactly what I thought. Can you issue a ldd binary, like ldd
 /usr/local/bin/Hugin to display all linked libraries for hugin and do the
 same for all hugin libs in /usr/local/lib, something  like ldd
 /usr/local/lib/libhugin*.so (or wherever you installed them).

 Maybe pulseaudio is linked by one of the dependent libraries used by
 hugin.


 In Fedora, and probably all Linux, hugin is installed in /usr/bin and as my
 build is on an x86_64 the libs are installed in /usr/lib64/hugin.
 I saved the output from 'ldd /usr/bin/hugin' and 'ldd
 /usr/lib64/hugin/lib*.so.*' in the file 'hugin-linked-libs.txt' which I have
 attached.
 I hope I have done that correctly, if not please let me know.

 Cheers,
 --
 Regards,
 Terry Duell



You did correctly. It shows that pulseaudio is not directly linked in any
way.
I will elaborate on lukas wirz example for a more thorough search.

Harry

-- 
You received this message because you are subscribed to the Google Groups 
Hugin and other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011/8/4 Wirz s...@lukas-wirz.de

 -BEGIN PGP SIGNED MESSAGE-
 Hash: SHA1

 Moin,


  Hmm, why is a graphics app trying to use PulseAudio? I don't even have
 it
  installed on any of my computers here!
 
  Exactly what I thought. Can you issue a ldd binary, like ldd
  /usr/local/bin/Hugin to display all linked libraries for hugin and do
 the
  same for all hugin libs in /usr/local/lib, something  like ldd
  /usr/local/lib/libhugin*.so (or wherever you installed them).
 
  Maybe pulseaudio is linked by one of the dependent libraries used by
 hugin.


 I believe the easiest way to check is saying

 ldd $(which hugin)| awk '{print $3}'| xargs ldd | grep Audio

 In case of my setup (debian squeeze, hugin 2011.3.0.5460 compiled from
 trunk, proprietary NV driver) there is no result ...

 cheers, lukas
 - --


Much better indeed. My suggestion was only a quick search of direct use of
pulseaudio but it isn't used directly. Your command will do a full search
but grep is case sensitive where I would always use a case insensitive
version (grep -i).

In other words: I would change your command to:
ldd $(which hugin)| awk '{print $3}'| xargs ldd | grep -i Audio

to make grep case insensitive as pulseaudio is not in capital letters (but
maybe sometimes it is?).

@Terry: Could you try the command as mentioned above? It should directly
show if and where audio libraries are used.

Note: My linux skills get a bit rusty since OSX is my primary platform since
2007.

Harry

-- 
You received this message because you are subscribed to the Google Groups 
Hugin and other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-04 Thread Tommaso Massimi

Hi all,

I have dug a bit more on the valgrind output,
and I figured out that there are 2 kinds of problem:

1) main thread #1 and one other thread are fighting for some audio vars

2) main thread #1 and one other thread are fighting for some allocation vars
 (managed by function g_slice_alloc, g_slice_free...
http://developer.gnome.org/glib/2.29/glib-Memory-Slices.html )

in the first case the audio library is invoked by library libcanberra-gtk,
which is an audio library;
it seems that gtk+ widgets are able to play a sound when receiving
some events, i.e. each widget can play a sound.
so it seems that both the fighting threads are using a graphic widget
which are playing a sound at the same time.

in the second case the lower level gtk memory functions are used from
two threads
at the same time.

Both situation are against this rule:

GTK+ is “thread aware” but not thread safe — it provides a global
lock controlled by gdk_threads_enter() and gdk_threads_leave() which
protects all use of GTK+. That is, only one thread can use GTK+ at any
given time.

http://www.gtk.org/api/2.6/gdk/gdk-Threads.html
http://blogs.operationaldynamics.com/andrew/software/gnome-desktop/gtk-thread-awareness.html

so it seems that hugin's threads are not well synchronized with the
other part of the task;
adding some call to gdk_threads_enter() and gdk_threads_leave() could
solve the problem.

Tom

-- 
You received this message because you are subscribed to the Google Groups 
Hugin and other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-04 Thread Lukáš Jirkovský

On 4 August 2011 20:09, Tommaso Massimi tmass...@gmail.com wrote:
Hi all,

I have dug a bit more on the valgrind output,
and I figured out that there are 2 kinds of problem:

1) main thread #1 and one other thread are fighting for some audio vars

2) main thread #1 and one other thread are fighting for some allocation vars
(managed by function g_slice_alloc, g_slice_free...
http://developer.gnome.org/glib/2.29/glib-Memory-Slices.html )

in the first case the audio library is invoked by library libcanberra-gtk,
which is an audio library;
it seems that gtk+ widgets are able to play a sound when receiving
some events, i.e. each widget can play a sound.
so it seems that both the fighting threads are using a graphic widget
which are playing a sound at the same time.

in the second case the lower level gtk memory functions are used from
two threads
at the same time.

Both situation are against this rule:

GTK+ is “thread aware” but not thread safe — it provides a global
lock controlled by gdk_threads_enter() and gdk_threads_leave() which
protects all use of GTK+. That is, only one thread can use GTK+ at any
given time.

http://www.gtk.org/api/2.6/gdk/gdk-Threads.html
http://blogs.operationaldynamics.com/andrew/software/gnome-desktop/gtk-thread-awareness.html

so it seems that hugin's threads are not well synchronized with the
other part of the task;
adding some call to gdk_threads_enter() and gdk_threads_leave() could
solve the problem.

Tom

I'm pretty sure PA has nothing to do with this crash. It's just
coincidence that it's linked to some of the libs (probably wxGTK). The
fact the bug is present on systems without PA (Windows, OS X and my
linux box too) can serve as a proof it's not pulse audio related.

Take a look at the last lines of the heldrinds output posted by Terry.
It shows a nice backtrace that the crash occurs in the Fast Preview
callback PreviewIdentifyTool::ShowImageNumber. Also there is a problem
in HuginBase::Panorama::changeFinished(bool) but it might be
unrelated. Unfortunately I didn't have time (or mood, to be honest) to
look at the code in detail.

Lukas

--
You received this message because you are subscribed to the Google Groups
Hugin and other free panoramic software group.
A list of frequently asked questions is available at:
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011/8/4 Lukáš Jirkovský l.jirkov...@gmail.com

On 4 August 2011 20:09, Tommaso Massimi tmass...@gmail.com wrote:
Hi all,

I have dug a bit more on the valgrind output,
and I figured out that there are 2 kinds of problem:

1) main thread #1 and one other thread are fighting for some audio vars

2) main thread #1 and one other thread are fighting for some allocation
vars
(managed by function g_slice_alloc, g_slice_free...
http://developer.gnome.org/glib/2.29/glib-Memory-Slices.html )

in the first case the audio library is invoked by library
libcanberra-gtk,
which is an audio library;
it seems that gtk+ widgets are able to play a sound when receiving
some events, i.e. each widget can play a sound.
so it seems that both the fighting threads are using a graphic widget
which are playing a sound at the same time.

in the second case the lower level gtk memory functions are used from
two threads
at the same time.

Both situation are against this rule:

http://www.gtk.org/api/2.6/gdk/gdk-Threads.html

http://blogs.operationaldynamics.com/andrew/software/gnome-desktop/gtk-thread-awareness.html

so it seems that hugin's threads are not well synchronized with the
other part of the task;
adding some call to gdk_threads_enter() and gdk_threads_leave() could
solve the problem.

Tom

Agreed. Overlooked that fact.

Lukas

It's a bit of a shame that this discussion is now more or less split over 2
threads. Did you see my response in the other thread (1).
It can be total rubbish and maybe it's not.

Harry

(1): http://groups.google.com/group/hugin-ptx/msg/02812f290aa20f3e

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-04 Thread Terry Duell


Hullo Harry,

On Fri, 05 Aug 2011 03:41:06 +1000, Harry van der Wolf hvdw...@gmail.com  
wrote:



2011/8/4 Wirz s...@lukas-wirz.de

In other words: I would change your command to:
ldd $(which hugin)| awk '{print $3}'| xargs ldd | grep -i Audio

to make grep case insensitive as pulseaudio is not in capital letters  
(but

maybe sometimes it is?).

@Terry: Could you try the command as mentioned above? It should directly
show if and where audio libraries are used.


I tried both forms of the command, as follows,

[terry@phenom ~]$ ldd $(which hugin)| awk '{print $3}'| xargs ldd | grep  
-i Audio

ldd: ./(0x7fff6a9de000): No such file or directory

[terry@phenom ~]$ ldd $(which hugin)| awk '{print $3}'| xargs ldd | grep  
Audio

ldd: ./(0x7fff03bff000): No such file or directory


Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-04 Thread Terry Duell


Hullo Tommaso,

On Fri, 05 Aug 2011 04:09:36 +1000, Tommaso Massimi tmass...@gmail.com  
wrote:



Hi all,

I have dug a bit more on the valgrind output,
and I figured out that there are 2 kinds of problem:

1) main thread #1 and one other thread are fighting for some audio vars

2) main thread #1 and one other thread are fighting for some allocation  
vars

 (managed by function g_slice_alloc, g_slice_free...
http://developer.gnome.org/glib/2.29/glib-Memory-Slices.html )

in the first case the audio library is invoked by library  
libcanberra-gtk,

which is an audio library;
it seems that gtk+ widgets are able to play a sound when receiving
some events, i.e. each widget can play a sound.
so it seems that both the fighting threads are using a graphic widget
which are playing a sound at the same time.

in the second case the lower level gtk memory functions are used from
two threads
at the same time.

[snip]

so it seems that hugin's threads are not well synchronized with the
other part of the task;
adding some call to gdk_threads_enter() and gdk_threads_leave() could
solve the problem.


That's good work.
Do you understand the valgrind output well enough to be more specific  
about where in the hugin code these problems are occurring?
Being able to home in one where the fixes are needed would help, if that's  
possible.


Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-03 Thread Terry Duell


Hullo,

On Wed, 03 Aug 2011 15:55:05 +1000, Gnome Nomad gnomeno...@gmail.com  
wrote:




I suppose you could attach them to the bug report?


If you haven't already seen it, Greg Lehey's recent post gives the URL  
where my test images can be downloaded.
My experience with my current Fedora 15 and all the test builds of Hugin,  
since the merge of the overview branch, will quickly show the FPW problem  
with this image set. Other hardware/OS may not react the same way, which  
is one of the aspects that makes this bug difficult to be sure about.


Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-03 Thread Terry Duell


Hullo,

On Wed, 03 Aug 2011 15:50:18 +1000, Gnome Nomad gnomeno...@gmail.com  
wrote:



If Hugin only works properly with NV and has problems with Nouveau,  
Hugin needs to fix them.


For some hugin works OK with Nvidia driver.
I don't think we know enough about the problem yet to know what to fix!

Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-03 Thread Greg 'groggy' Lehey

On Tuesday,  2 August 2011 at 19:50:18 -1000, Gnome Nomad wrote:

 If Hugin only works properly with NV and has problems with Nouveau,
 Hugin needs to fix them.

... assuming the problem is with Hugin.  So far we can't be sure,
though the appearance of the problem with Microsoft as well suggests
that it *is* a Hugin problem.

Greg
--
Sent from my real computer
Finger g...@freebsd.org for PGP public key.
See complete headers for address and phone numbers.
This message is digitally signed.  If your Microsoft MUA reports
problems, please read http://tinyurl.com/broken-mua


pgpeQR4nqgvRn.pgp
Description: PGP signature

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-03 Thread Greg 'groggy' Lehey

On Tuesday,  2 August 2011 at 19:55:05 -1000, Gnome Nomad wrote:
 Terry Duell wrote:
 Hullo Greg,

 On Wed, 03 Aug 2011 09:12:42 +1000, Greg 'groggy' Lehey
 groog...@gmail.com wrote:

 [snip]

 And FWIW, I can't reproduce this with FreeBSD and the proprietary
 nVidia driver.

 OK. Does this indicate that it might be affected by different versions
 of some libs, or somesuch?

 Might be effected by Xorg versions, too?

Quite possibly.  Until we have a better feeling for what the problem
is, we can't exclude anything, though my bet is that it's not
dependent on the phase of the moon.

Greg
--
Sent from my real computer
Finger g...@freebsd.org for PGP public key.
See complete headers for address and phone numbers.
This message is digitally signed.  If your Microsoft MUA reports
problems, please read http://tinyurl.com/broken-mua


pgpQslYHhNtol.pgp
Description: PGP signature

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview


Terry Duell wrote:

Hullo,

On Wed, 03 Aug 2011 15:55:05 +1000, Gnome Nomad gnomeno...@gmail.com 
wrote:



I suppose you could attach them to the bug report?


If you haven't already seen it, Greg Lehey's recent post gives the URL 
where my test images can be downloaded.


Yah, I saw that after I sent mine.

My experience with my current Fedora 15 and all the test builds of 
Hugin, since the merge of the overview branch, will quickly show the FPW 
problem with this image set. Other hardware/OS may not react the same 
way, which is one of the aspects that makes this bug difficult to be 
sure about.


Might also make it easier, if one's able to confirm that it happens on 
this hardware/OS combination, then run it under debug to pin down just 
where the program was when it started happening. I've had developers of 
other products tell me they couldn't help me with issues I was having 
because they couldn't duplicate the problem on their hardware/OS combo, 
and I couldn't run it under debug on my set up and send them the results.


If it's a race condition in the preview, that could possibly be spotted 
by code inspection.


--
Gnome Nomad
gnomeno...@gmail.com
wandering the landscape of god
http://www.cafepress.com/otherend/

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview


Terry Duell wrote:

Hullo,

On Wed, 03 Aug 2011 15:50:18 +1000, Gnome Nomad gnomeno...@gmail.com 
wrote:



If Hugin only works properly with NV and has problems with Nouveau, 
Hugin needs to fix them.


For some hugin works OK with Nvidia driver.
I don't think we know enough about the problem yet to know what to fix!


I agree with that!

--
Gnome Nomad
gnomeno...@gmail.com
wandering the landscape of god
http://www.cafepress.com/otherend/

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview


Greg 'groggy' Lehey wrote:

On Tuesday,  2 August 2011 at 19:50:18 -1000, Gnome Nomad wrote:

If Hugin only works properly with NV and has problems with Nouveau,
Hugin needs to fix them.


... assuming the problem is with Hugin.  So far we can't be sure,
though the appearance of the problem with Microsoft as well suggests
that it *is* a Hugin problem.


Sounds likely to me.

--
Gnome Nomad
gnomeno...@gmail.com
wandering the landscape of god
http://www.cafepress.com/otherend/

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview


Greg 'groggy' Lehey wrote:

On Tuesday,  2 August 2011 at 19:55:05 -1000, Gnome Nomad wrote:

Terry Duell wrote:

Hullo Greg,

On Wed, 03 Aug 2011 09:12:42 +1000, Greg 'groggy' Lehey
groog...@gmail.com wrote:

[snip]


And FWIW, I can't reproduce this with FreeBSD and the proprietary
nVidia driver.

OK. Does this indicate that it might be affected by different versions
of some libs, or somesuch?

Might be effected by Xorg versions, too?


Quite possibly.  Until we have a better feeling for what the problem
is, we can't exclude anything, though my bet is that it's not
dependent on the phase of the moon.


Might be if you were in orbit around the moon; your orbital position 
might have some influence on the strength of magnetic fields and 
radiation just in case ... ;-)


--
Gnome Nomad
gnomeno...@gmail.com
wandering the landscape of god
http://www.cafepress.com/otherend/

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-03 Thread Tommaso Massimi

Hi all,

I'm a hugin user and, after read this thread, I'd like to help you as I can:

zthreads project seems not mantained, but the community is working:
here you can found the zthreads forum and in the first mail there is a fix
for a memory leak suggested by some user:

http://sourceforge.net/projects/zthread/forums/forum/22488

that patch was accepted by debian:
http://packages.debian.org/changelogs/pool/main/z/zthreads/zthreads_2.3.2-3/changelog



as zthreads is a wrapper for pthreads, probably helgrind works with it.
If somebody able to reproduce the problem (and with a lot of patience..) runs

valgrind --tool=helgrind hugin

and report the output we could see some error

my 2 cent

Tom

-- 
You received this message because you are subscribed to the Google Groups 
Hugin and other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-03 Thread Terry Duell


Hullo Tommaso,

On Wed, 03 Aug 2011 20:00:16 +1000, Tommaso Massimi tmass...@gmail.com  
wrote:





as zthreads is a wrapper for pthreads, probably helgrind works with it.
If somebody able to reproduce the problem (and with a lot of patience..)  
runs


valgrind --tool=helgrind hugin

and report the output we could see some error

my 2 cent


Thanks for that advice. I have not used valgrind in anger, just  
experimented a bit.
I ran the above command, and the FPW problem did occur, but initially it  
wasn't easy to be sure, as everything happens So Slow, but eventually when  
the pano wasn't being displayed in the FPW I tried closing the FPW and the  
system reported that it wasn't responding, which is one of the ways a  
failure mainfests itself here.
There was no valgrind output from about the point when the images were  
selected to be loaded, so I don't think it provided anything useful.

It is possible that I haven't gone about this correctly.
Should I have a debug version of hugin installed (on Fedora that would be  
the hugin-xxx-debuginfo.rpm)?
I will check if I have the debug installed and try again. Perhaps also  
check valgrind options to see if there are any that might help.


Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-03 Thread Greg 'groggy' Lehey

On Tuesday,  2 August 2011 at 21:50:39 -1000, Gnome Nomad wrote:

 If it's a race condition in the preview, that could possibly be
 spotted by code inspection.

If you're *very* lucky.  Ultimately all problems are solved by code
inspection, of course, but there's lots of code to inspect.  It would
be nice to catch the thing in action and point a debugger at it.  That
would at least narrow things down.

Greg
--
Sent from my real computer
Finger g...@freebsd.org for PGP public key.
See complete headers for address and phone numbers.
This message is digitally signed.  If your Microsoft MUA reports
problems, please read http://tinyurl.com/broken-mua


pgpaMK4kKNQ9w.pgp
Description: PGP signature

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-02 Thread Harry van der Wolf

My 2 cents.

2011/7/29 David Haberthür david.haberth...@gmail.com


 On 26.07.2011, at 23:22, Carl von Einem wrote:

  Yuval Levy schrieb am 26.07.11 18:27:
  On July 26, 2011 08:10:28 AM Lukáš Jirkovský wrote:
  I think this is one of the most serious bugs hugin ever had.
 
  Mainly because of the number of users affected by this issue.
 
  Yes, it is growing.  And I see two drivers of growth that would confirm
 your
  hypothesis that this is a threading issue:
 
  1) as time progresses, people buy newer, better PCs with more threads.
  If
  single-thread machines are not affected; dual-thread machines are barely
  affected; and machines with 4 threads and more are significantly
 affected, it
  seems to indicate that this is a threading issue.
 
  I didn't see this issue so far on both of my Macs: the oldish one is a
 dual processor PowerPC (2x 2.3 GHz, 6.5 GB RAM), the newer one is a dual
 hexacore Intel Mac with about 24 or 26 GB of RAM.
 
  I haven't noticed this bug is mentioned by Mac users at all, or did I
 miss some messages?
 
  Carl

 I haven's seen any weird thing happening with the fast preview in Harrys
 (newest) build on my MacBook Pro 2.33 GHz Intel Core 2 Duo with 2 GB of RAM.
 Habi


I have seen it very occasionally in the last builds. Never from the
Assistant but 2-3 times from the normal manual workflow. It makes hugin hang
and after some time (2-3 minutes?) it gives a crash message. I can't
reproduce it though and I don't see errors in the error log or the system
log.
I'm back this weekend and will give it some more attention.

Harry

-- 
You received this message because you are subscribed to the Google Groups 
Hugin and other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-02 Thread Terry Duell


Hullo All,

On Mon, 01 Aug 2011 08:23:23 +1000, Gnome Nomad gnomeno...@gmail.com  
wrote:


[snip]
Whatever version is currently in 64-bit Debian Sid gave me the hang  
yesterday after I (mistakenly) did a distribution-upgrade. The upgrade  
(among other things) got rid of the NV (open source video driver for  
NVidia hardware) and forced me to use the proprietary (binary) NVidia  
video drivers. That changed the GLX library used. The hang happened on a  
uniprocessor machine.


If I understand this correctly, hugin worked OK for you with the NV driver  
but failed when you used the proprietary nvidia driver.
My experience with Fedora 15 is that I have had the same FPW failures with  
the proprietary nvidia driver and the with the open source nouveau driver.


On the wider question, can anyone who knows more about this business than  
I, say if we are getting closer to homing in on this bug?
I would like to help more, if I can, but that stops well short of looking  
for a coding error.
If there is more testing that needs to be done to help refine the search,  
I will do what I can.


Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-02 Thread Greg 'groggy' Lehey

On Wednesday,  3 August 2011 at  8:51:37 +1000, Terry Duell wrote:

 If I understand this correctly, hugin worked OK for you with the NV
 driver but failed when you used the proprietary nvidia driver.  My
 experience with Fedora 15 is that I have had the same FPW failures
 with the proprietary nvidia driver and the with the open source
 nouveau driver.

And FWIW, I can't reproduce this with FreeBSD and the proprietary
nVidia driver.

 On the wider question, can anyone who knows more about this business
 than I, say if we are getting closer to homing in on this bug?

I was wondering that too.

 I would like to help more, if I can, but that stops well short of
 looking for a coding error.

I don't think this is a traditional coding error.

I made two suggestions that don't seem to have been followed up on:

- Find a set of photos which provoke the problem.
- Try limiting the number of concurrent threads.

One thing that I undertook to do was to try my photos out on
Microsoft.  That didn't get off the ground.  My old Microsoft laptop
overheated and shut down, and running it in a VM didn't work because I
didn't have the right support software installed, and my Microsoft-fu
isn't up to installing it.  I also suspect that it would have run out
of memory (I can't give a VM more than about 800 MB memory).

Greg
--
Sent from my real computer
Finger g...@freebsd.org for PGP public key.
See complete headers for address and phone numbers.
This message is digitally signed.  If your Microsoft MUA reports
problems, please read http://tinyurl.com/broken-mua


pgpXVTkLA8pSK.pgp
Description: PGP signature

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-02 Thread Terry Duell


Hullo Greg,

On Wed, 03 Aug 2011 09:12:42 +1000, Greg 'groggy' Lehey  
groog...@gmail.com wrote:


[snip]


And FWIW, I can't reproduce this with FreeBSD and the proprietary
nVidia driver.


OK. Does this indicate that it might be affected by different versions of  
some libs, or somesuch?



I would like to help more, if I can, but that stops well short of
looking for a coding error.


I don't think this is a traditional coding error.


OK. I don't have enough of the right experience to know one way or another  
on that.


I made two suggestions that don't seem to have been followed up on:

- Find a set of photos which provoke the problem.
- Try limiting the number of concurrent threads.



On the first front, I do have a set of images that I have been using as my  
test set, and I was able to easily provoke the failure on my system when  
testing versions since the merge of the overview code, but unable to get a  
failure on versions prior to that.
The set consists of 5 images each approx 2 MB. I will see if I can get the  
problem to show itself with reduced size version of those images, which  
might make it easier to make them available.

How to I make them available? I don't have a site where I can put them.

On the threads question, I could test limiting the number of concurrent  
threads, if I knew how to do that. Anyone who can help with that?


Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-02 Thread Gnome Nomad


Terry Duell wrote:

Hullo All,

On Mon, 01 Aug 2011 08:23:23 +1000, Gnome Nomad gnomeno...@gmail.com 
wrote:


[snip]
Whatever version is currently in 64-bit Debian Sid gave me the hang 
yesterday after I (mistakenly) did a distribution-upgrade. The upgrade 
(among other things) got rid of the NV (open source video driver for 
NVidia hardware) and forced me to use the proprietary (binary) NVidia 
video drivers. That changed the GLX library used. The hang happened on 
a uniprocessor machine.


If I understand this correctly, hugin worked OK for you with the NV 
driver but failed when you used the proprietary nvidia driver.


Yes. It works fine after I set it to not open the preview window after 
running the assistant.


My experience with Fedora 15 is that I have had the same FPW failures 
with the proprietary nvidia driver and the with the open source nouveau 
driver.


Could be, I don't know. I've never had to use the proprietary driver 
before. But NV is disappearing from distros, being replaced by Nouveau. 
NV hasn't been updated to support new NVidia cards for some time now.


If Hugin only works properly with NV and has problems with Nouveau, 
Hugin needs to fix them.


--
Gnome Nomad
gnomeno...@gmail.com
wandering the landscape of god
http://www.cafepress.com/otherend/

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-02 Thread Gnome Nomad


Terry Duell wrote:

Hullo Greg,

On Wed, 03 Aug 2011 09:12:42 +1000, Greg 'groggy' Lehey 
groog...@gmail.com wrote:


[snip]


And FWIW, I can't reproduce this with FreeBSD and the proprietary
nVidia driver.


OK. Does this indicate that it might be affected by different versions 
of some libs, or somesuch?


Might be effected by Xorg versions, too?


I would like to help more, if I can, but that stops well short of
looking for a coding error.


I don't think this is a traditional coding error.


OK. I don't have enough of the right experience to know one way or 
another on that.


Unless it's a coding error that has harmful effects only when the app 
calls particular X or video driver APIs? I couldn't duplicate the 
problem using the same version of Linux (Aptosid 2011-02) on my Toshiba 
laptop which uses the Intel915 driver.



I made two suggestions that don't seem to have been followed up on:

- Find a set of photos which provoke the problem.
- Try limiting the number of concurrent threads.


On the first front, I do have a set of images that I have been using as 
my test set, and I was able to easily provoke the failure on my system 
when testing versions since the merge of the overview code, but unable 
to get a failure on versions prior to that.
The set consists of 5 images each approx 2 MB. I will see if I can get 
the problem to show itself with reduced size version of those images, 
which might make it easier to make them available.

How to I make them available? I don't have a site where I can put them.


I suppose you could attach them to the bug report?

--
Gnome Nomad
gnomeno...@gmail.com
wandering the landscape of god
http://www.cafepress.com/otherend/

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-08-01 Thread Frederic Da Vitoria

2011/8/1 Gnome Nomad gnomeno...@gmail.com

 Frederic Da Vitoria wrote:

 2011/7/29 Yuval Levy goo...@levy.ch mailto:goo...@levy.ch


On July 27, 2011 05:24:44 AM Frederic Da Vitoria wrote:
  I still have a mono-processor laptop from 2002 running Windows XP.
  Would it help if I did a few tests on it?

sure!  the more data points we have, the better.


 I found one project which hanged on my Intel Core 2 Duo T6670 (Vista). I
 tested it on my old Intel Pentium 4 M 1.8 GHz (XP) and it hanged in a
 slightly different way (I got a little more of the preview window on the
 Pentium M before being stuck while on the Core 2 I got a completely empty
 window). I don't have any older machine. I can put my hands on 2 years old
 Celeron XP computers, if it's any use.


 Whatever version is currently in 64-bit Debian Sid gave me the hang
 yesterday after I (mistakenly) did a distribution-upgrade. The upgrade
 (among other things) got rid of the NV (open source video driver for NVidia
 hardware) and forced me to use the proprietary (binary) NVidia video
 drivers. That changed the GLX library used. The hang happened on a
 uniprocessor machine.


Thanks Gnome Nomad, I forgot about the video cards. The Pentium 4 M has a
ATI Mobility Radeon 7500 and the T6670 has a NVidia GeForce 9300M GS

-- 
Frederic Da Vitoria
(davitof)

Membre de l'April - « promouvoir et défendre le logiciel libre » -
http://www.april.org

-- 
You received this message because you are subscribed to the Google Groups 
Hugin and other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-07-31 Thread Frederic Da Vitoria

2011/7/29 Yuval Levy goo...@levy.ch

 On July 27, 2011 05:24:44 AM Frederic Da Vitoria wrote:
  I still have a mono-processor laptop from 2002 running Windows XP.
  Would it help if I did a few tests on it?

 sure!  the more data points we have, the better.


I found one project which hanged on my Intel Core 2 Duo T6670 (Vista). I
tested it on my old Intel Pentium 4 M 1.8 GHz (XP) and it hanged in a
slightly different way (I got a little more of the preview window on the
Pentium M before being stuck while on the Core 2 I got a completely empty
window). I don't have any older machine. I can put my hands on 2 years old
Celeron XP computers, if it's any use.

-- 
Frederic Da Vitoria
(davitof)

Membre de l'April - « promouvoir et défendre le logiciel libre » -
http://www.april.org

-- 
You received this message because you are subscribed to the Google Groups 
Hugin and other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-07-31 Thread Gnome Nomad


Frederic Da Vitoria wrote:

2011/7/29 Yuval Levy goo...@levy.ch mailto:goo...@levy.ch

On July 27, 2011 05:24:44 AM Frederic Da Vitoria wrote:
  I still have a mono-processor laptop from 2002 running Windows XP.
  Would it help if I did a few tests on it?

sure!  the more data points we have, the better.


I found one project which hanged on my Intel Core 2 Duo T6670 (Vista). I 
tested it on my old Intel Pentium 4 M 1.8 GHz (XP) and it hanged in a 
slightly different way (I got a little more of the preview window on the 
Pentium M before being stuck while on the Core 2 I got a completely 
empty window). I don't have any older machine. I can put my hands on 2 
years old Celeron XP computers, if it's any use.


Whatever version is currently in 64-bit Debian Sid gave me the hang 
yesterday after I (mistakenly) did a distribution-upgrade. The upgrade 
(among other things) got rid of the NV (open source video driver for 
NVidia hardware) and forced me to use the proprietary (binary) NVidia 
video drivers. That changed the GLX library used. The hang happened on a 
uniprocessor machine.


--
Gnome Nomad
gnomeno...@gmail.com
wandering the landscape of god
http://www.cafepress.com/otherend/

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-07-29 Thread David Haberthür


On 26.07.2011, at 23:22, Carl von Einem wrote:

 Yuval Levy schrieb am 26.07.11 18:27:
 On July 26, 2011 08:10:28 AM Lukáš Jirkovský wrote:
 I think this is one of the most serious bugs hugin ever had.
 
 Mainly because of the number of users affected by this issue.
 
 Yes, it is growing.  And I see two drivers of growth that would confirm your
 hypothesis that this is a threading issue:
 
 1) as time progresses, people buy newer, better PCs with more threads.  If
 single-thread machines are not affected; dual-thread machines are barely
 affected; and machines with 4 threads and more are significantly affected, it
 seems to indicate that this is a threading issue.
 
 I didn't see this issue so far on both of my Macs: the oldish one is a dual 
 processor PowerPC (2x 2.3 GHz, 6.5 GB RAM), the newer one is a dual 
 hexacore Intel Mac with about 24 or 26 GB of RAM.
 
 I haven't noticed this bug is mentioned by Mac users at all, or did I miss 
 some messages?
 
 Carl

I haven's seen any weird thing happening with the fast preview in Harrys 
(newest) build on my MacBook Pro 2.33 GHz Intel Core 2 Duo with 2 GB of RAM.
Habi

-- 
You received this message because you are subscribed to the Google Groups 
Hugin and other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-07-29 Thread David Haberthür


On 26.07.2011, at 23:51, Yuval Levy wrote:

 On July 26, 2011 05:22:42 PM Carl von Einem wrote:
 I didn't see this issue so far on both of my Macs: the oldish one is a
 dual processor PowerPC (2x 2.3 GHz, 6.5 GB RAM), the newer one is a dual
 hexacore Intel Mac with about 24 or 26 GB of RAM.
 
 thanks for reporting.  what version of OS X are you running?  if you have 
 acces to OS X Lion, we need some data points there.

The question was not aimed at me, but I'm gonna chime in anyway: I'm running 
10.6.8 and everything is fine. I plan to update to Lion in a week or so (when 
the University provides the site license image). I'll try to chime in with bug 
814301 regarding control point detectors as soon as possible

Habi

-- 
You received this message because you are subscribed to the Google Groups 
Hugin and other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-07-28 Thread Yuval Levy

On July 27, 2011 09:24:25 PM Greg 'groggy' Lehey wrote:
 Which version should I be trying?  hugin-2011.2_beta?

Sure.  I think any version after 2010.4.0 is good for trying.  Terry has been 
able to reproduce the bug as far back as revision 4808 (which is the last time 
when major features were added to the fast preview).

If the culprit is the ImageCache, it could go a few weeks back, and if it is 
CPfind, a few weeks later.

Yuv


signature.asc
Description: This is a digitally signed message part.

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-07-28 Thread Yuval Levy

On July 27, 2011 05:24:44 AM Frederic Da Vitoria wrote:
 I still have a mono-processor laptop from 2002 running Windows XP.
 Would it help if I did a few tests on it?

sure!  the more data points we have, the better.
Yuv


signature.asc
Description: This is a digitally signed message part.

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-07-27 Thread Rogier Wolff

On Tue, Jul 26, 2011 at 12:27:19PM -0400, Yuval Levy wrote:

 2) as Hugin has evolved, threads have become more widely used and
 conflicts more likely.  Revision 4524 (2010-11-04 18:31:15) is as
 far back as I went in my investigations so far looking for the use
 of threads.  This is when the ImageCache was transferred to a
 separate thread, making the application more responsive.  In it
 alone not a problem, but maybe the introduction of later threaded
 features introduced the conflicts that are becoming apparent now.

There have been longstanding bugs that I suspect come from
ImageCache. Crashes, image corruption etc. etc. These are very likely
to change character when you move the ImageCache part to a separate
thread. They might suddenly happen all the time when before they
happened only sometimes. Or the other way around. Or a crash might
turn in a corruption, etc etc. 

Roger. 

-- 
** r.e.wo...@bitwizard.nl ** http://www.BitWizard.nl/ ** +31-15-2600998 **
**Delftechpark 26 2628 XH  Delft, The Netherlands. KVK: 27239233**
*-- BitWizard writes Linux device drivers for any device you may have! --*
The plan was simple, like my brother-in-law Phil. But unlike
Phil, this plan just might work.

-- 
You received this message because you are subscribed to the Google Groups 
Hugin and other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-07-27 Thread Frederic Da Vitoria

2011/7/26, Yuval Levy goo...@levy.ch:
  On July 26, 2011 08:10:28 AM Lukáš Jirkovský wrote:
  I think this is one of the most serious bugs hugin ever had.
 
  What makes you reach this judgment?

 Mainly because of the number of users affected by this issue.

 Yes, it is growing.  And I see two drivers of growth that would confirm your
 hypothesis that this is a threading issue:

 1) as time progresses, people buy newer, better PCs with more threads.  If
 single-thread machines are not affected; dual-thread machines are barely
 affected; and machines with 4 threads and more are significantly affected,
 it
 seems to indicate that this is a threading issue.

 2) as Hugin has evolved, threads have become more widely used and conflicts
 more likely.  Revision 4524 (2010-11-04 18:31:15) is as far back as I went
 in
 my investigations so far looking for the use of threads.  This is when the
 ImageCache was transferred to a separate thread, making the application more
 responsive.  In it alone not a problem, but maybe the introduction of later
 threaded features introduced the conflicts that are becoming apparent now.


 Also the
 fact that it renders the fast preview, which is IMHO becoming the
 center of the workflow, unusable makes it more important.

 Yes, the FPW is becoming the center of the workflow IMHO too, and making it
 usable is very important.

 I suspect that getting the threading right will be a big and intensive job.
 The more I look at recent reports in the bug tracker, the more dots there
 are
 to connect and they all seem to lead to the same pesky thing.

 Could the conflict be related to the use of different threading libraries?

 we have boost threads since revision 4524 [0] for the images cache and
 cpfind
 uses zthread?

I still have a mono-processor laptop from 2002 running Windows XP.
Would it help if I did a few tests on it?

-- 
Frederic Da Vitoria
(davitof)

Membre de l'April - « promouvoir et défendre le logiciel libre » -
http://www.april.org

-- 
You received this message because you are subscribed to the Google Groups 
Hugin and other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-07-27 Thread Greg 'groggy' Lehey

On Tuesday, 26 July 2011 at  9:40:50 -0400, Yuval Levy wrote:
 On July 24, 2011 11:52:53 PM Greg 'groggy' Lehey wrote:

 I wasn't able to provoke this bug.

 Anything else I should try?

 I am no expert.  You probably know better than me, and indeed you
 seem to be on the right track with the threading thing.  What would
 you suggest we try next?

Difficult to say.  Everybody has his own procedures, and mine would
have been with gdb if I had been able to reproduce the bug.  What I'm
going to do now is to try it out with Apple and Microsoft and see if I
can reproduce it on those platforms with the same data.  But if I do,
I don't know what I can do about it, at least on Microsoft, where I
don't have any development tools.

Which version should I be trying?  hugin-2011.2_beta?

 I just tried to apply a strategy that I know from solving finance
 problems to solve this code problem, helping those who can provoke
 the bug zero-in on the changeset that introduced it; and trying to
 find correlations between their specific situations and the buggy
 behavior while not being able to reproduce it myself.

Certainly getting a collection of data on where it happens and where
it doesn't might be useful.  It seems we need at least:

  OS, including specific release
  Way the executable was built (libraries, etc)
  Number of processors

It would also be good to find a standard set of images with which the
bug shows itself.

Another thing that might be of interest would be to set the
preferences to use only one CPU and see if the problem still occurs.
It might also be nice to add an option to limit the number of threads,
but I suspect that's non-trivial.

Greg
--
Sent from my real computer
Finger g...@freebsd.org for PGP public key.
See complete headers for address and phone numbers.
This message is digitally signed.  If your Microsoft MUA reports
problems, please read http://tinyurl.com/broken-mua


pgp1DnA2JiyNQ.pgp
Description: PGP signature

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-07-26 Thread Terry Duell


Hullo Greg,

On Mon, 25 Jul 2011 13:52:53 +1000, Greg 'groggy' Lehey  
groog...@gmail.com wrote:


[snip]


This was with my own images, a collection of 25.  I tried playing
around with the buttons in the fast preview window and kept an eye on
the process memory image with ps, but there's no evidence of any
memory leak.

Anything else I should try?  Until I can reproduce it with tip, there
doesn't seem to be much point to try a binary search.


I have found slightly different behaviour with the latter versions, where  
they are often OK if the FPW opens with the overview mode hidden. If Hugin  
has been quit with the FPW showing the overview mode, when hugin is next  
started, and when the FPW is opened it will be showing the overview mode,  
and THAT is when it gives trouble.
My testing of much earlier versions (eg rev 4814) on my system, has shown  
that even if the FPW first opens with the overview mode hidden, the FPW is  
completely unresponsive and either hugin crashes or the system asks if I  
want to quit the FPW, which takes hugin down.
My testing today indicates (if I haven't made any blunders) that the  
problem was introduced with the overview mode. Versions later than the  
overview mode clearly show the problem, and I can't get versions prior to  
overview mode to do anything other expected behaviour.
If you haven't already done so, try your version after quitting with  
overview mode shown in FPW. It will be interesting to see if you can find  
a fault.
I wonder if the varying behaviour of different systems is being influenced  
by the varying versions of dependent libraries that different systems have  
installed, each reacting a bit differently to some bit of code in the  
overview mode. Just my 2 bobs worth.


Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-07-26 Thread Greg 'groggy' Lehey

On Tuesday, 26 July 2011 at 16:24:32 +1000, Terry Duell wrote:
 Hullo Greg,

 On Mon, 25 Jul 2011 13:52:53 +1000, Greg 'groggy' Lehey
 groog...@gmail.com wrote:

 [snip]

 Anything else I should try?  Until I can reproduce it with tip, there
 doesn't seem to be much point to try a binary search.

 I have found slightly different behaviour with the latter versions, where
 they are often OK if the FPW opens with the overview mode hidden.

Ah!  I don't like the overview mode (maybe because I don't understand
its purpose), but I didn't know I could turn it off.  Instead I
detached it and pushed it off the edge of the screen.  But now I

 If Hugin has been quit with the FPW showing the overview mode, when
 hugin is next started, and when the FPW is opened it will be showing
 the overview mode, and THAT is when it gives trouble.

Well, all that applies, but it doesn't give me any trouble.  I thought
that maybe that's because I detached the overview window, but I've
tried again with the window attached, and it still works fine.

 If you haven't already done so, try your version after quitting with
 overview mode shown in FPW. It will be interesting to see if you can
 find a fault.

That's what I've been doing all the time.  No problems.

 I wonder if the varying behaviour of different systems is being
 influenced by the varying versions of dependent libraries that
 different systems have installed, each reacting a bit differently to
 some bit of code in the overview mode. Just my 2 bobs worth.

Quite possibly.  Did you see the bug report I put in today?  It seems
that some component crashed on exiting because of a consistency check.
Since it was exiting anyway, it seems to be harmless, but it's clearly
related to specific threading libraries.

Greg
--
Finger g...@freebsd.org for PGP public key.
See complete headers for address and phone numbers.
This message is digitally signed.  See
http://www.lemis.com/grog/email/signed-mail.php for more details.
If your Microsoft MUA reports problems, please read
http://tinyurl.com/broken-mua


pgpoaUZGTfACt.pgp
Description: PGP signature

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-07-26 Thread Lukáš Jirkovský

Hello,

On 24 July 2011 15:47, Yuval Levy goo...@levy.ch wrote:
 Hi all,

 We know that the current Hugin is plagued with a pesky spurious bug that is
 difficult to reproduce.  It is causing grievance to a lot of users as seen on
 the ML and on the main tracker ticket where the bug is discussed  [0].

I think this is one of the most serious bugs hugin ever had. I'm
affected by this bug too. I spent quite some trying to debug it, but
it never showed up with debug build for some reason.

 This is most likely caused by a memory leak.  The most difficult part of
 fixing a memory leak is to identify where it happens.  Would you like to help
 nuke this bug?

In my opinion it's some threading issue, because even though it
happens very often, it doesn't happen every time.

 I herewith call on everybody interested to help hunt down and fix that pesky
 bug.  Here is a strategy and a plan.  I cross-post this information to the
 hugin-ptx mailing list, but please continue discussion and contribution on the
 tracker ticket dedicated to this bug [0], where it is more easily structured
 and does not get lost in time.


 1) Strategy

 In a nutshell:  identify the range of revisions that are likely to contain the
 error and narrow it down with a binary tree search.  Split the revision range
 in two, build and test the version in the middle.  If the error is present, we
 know it is in the earlier revisions range.  Else in the later.  Take the
 identified narrower revision range and repeat the process.  There is a high
 likelyhood that after about eight repetitions the source of the error is
 identified.

 In detail:

 Today, we are at revision 5435.  It is most likely that the bug was introduced
 after 2010.4 was released.  Revision 4780 is 2010.4 final.

 So the error is most likely between 4780 and 5435.  That would be 656
 different versions of Hugin, but it is much less than that: many of these
 changesets are not on the default branch.


 2) The plan

 List the relevant changesets on the default branch:

 hg log -b default -r4780:tip --template 'r: {rev}\n'  revlist.txt

 That's 448 changesets to examine.

 Start from the middle, i.e. the 224th changeset in that list.  build that
 version and try to trigger the error.  If the bug is there, it is likely to be
 in the newer half of the list.  If not, in the older half.  Rebuild the list,
 now 224 entries long, and cut it in half again.  Repeat.

 With such a binary strategy, we should zero in on the bug in a maximum of
 eight iterations.

 224, 112, 56, 28, 14, 7, 4, 2, 1

Good idea! I didn't thought about bisecting hugin before. My guess is
that it's somehow related to the panorama overview. BTW there's bisect
extension [1] which does this automatically.

[1] http://mercurial.selenic.com/wiki/BisectExtension

Lukas

-- 
You received this message because you are subscribed to the Google Groups 
Hugin and other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

Hi Lukáš,

On July 26, 2011 08:10:28 AM Lukáš Jirkovský wrote:
 I think this is one of the most serious bugs hugin ever had.

What makes you reach this judgment?


 In my opinion it's some threading issue, because even though it
 happens very often, it doesn't happen every time.

You know more about this than me.  I've added the question to the bug report, 
if we get enough observations to correlate the number of threads in the system 
with the observed behavior we'll verify your hypothesis.


  With such a binary strategy, we should zero in on the bug in a maximum of
  eight iterations.
  
  224, 112, 56, 28, 14, 7, 4, 2, 1
 
 Good idea! I didn't thought about bisecting hugin before. My guess is
 that it's somehow related to the panorama overview. BTW there's bisect
 extension [1] which does this automatically.
 
 [1] http://mercurial.selenic.com/wiki/BisectExtension

Oh, nice to see that somebody implemented the binary strategy in Mercurial 
(copying it from Git).

OT:  The binary strategy is applicable to many practical problems.  I first 
encountered it when researching financial valuation models [1].

[0] https://bugs.launchpad.net/hugin/+bug/792896
[1] http://en.wikipedia.org/wiki/Binomial_options_pricing_model  


signature.asc
Description: This is a digitally signed message part.

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

On July 24, 2011 11:52:53 PM Greg 'groggy' Lehey wrote:
 I've just cloned the hg repository and built the latest version
 (5345).

Thank you for chiming in, your help is appreciated.


 I had some difficulties building, which I'll mention in a
 separate message

Please do. [0] can use an update, and the port as well.


 I wasn't able to provoke this bug.

Some of us are luckier than others, but we seem to be zeroing-in on it 
together.


 Anything else I should try? 

I am no expert.  You probably know better than me, and indeed you seem to be 
on the right track with the threading thing.  What would you suggest we try 
next?

I just tried to apply a strategy that I know from solving finance problems to 
solve this code problem, helping those who can provoke the bug zero-in on the 
changeset that introduced it; and trying to find correlations between their 
specific situations and the buggy behavior while not being able to reproduce 
it myself.  In the meantime I know that I can reproduce it on my netbook.

I only do the release management.  I occasionally hack in the codebase, but my 
serious coding times (mostly assembler) are more than twenty years back.  My 
formal training is in economics and business and in September I go to law 
school.  Logic, analytic, and strategic thinking are transferable skills, but 
I am definitely limited by my lack of formal C++ training and limited trial 
and error experience.

Plus, there is this limitation I keep bumping into, that the day only has 24 
hours.

So, the build just finished.  Out for a test ride.

Yuv


[0] http://wiki.panotools.org/Hugin_Compiling_FreeBSD


signature.asc
Description: This is a digitally signed message part.

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-07-26 Thread Lukáš Jirkovský

On 26 July 2011 15:26, Yuval Levy goo...@levy.ch wrote:
 Hi Lukáš,

 On July 26, 2011 08:10:28 AM Lukáš Jirkovský wrote:
 I think this is one of the most serious bugs hugin ever had.

 What makes you reach this judgment?


Mainly because of the number of users affected by this issue. Also the
fact that it renders the fast preview, which is IMHO becoming the
center of the workflow, unusable makes it more important.

Lukas

-- 
You received this message because you are subscribed to the Google Groups 
Hugin and other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

  On July 26, 2011 08:10:28 AM Lukáš Jirkovský wrote:
  I think this is one of the most serious bugs hugin ever had.
  
  What makes you reach this judgment?
 
 Mainly because of the number of users affected by this issue.

Yes, it is growing.  And I see two drivers of growth that would confirm your 
hypothesis that this is a threading issue:

1) as time progresses, people buy newer, better PCs with more threads.  If 
single-thread machines are not affected; dual-thread machines are barely 
affected; and machines with 4 threads and more are significantly affected, it 
seems to indicate that this is a threading issue.

2) as Hugin has evolved, threads have become more widely used and conflicts 
more likely.  Revision 4524 (2010-11-04 18:31:15) is as far back as I went in 
my investigations so far looking for the use of threads.  This is when the 
ImageCache was transferred to a separate thread, making the application more 
responsive.  In it alone not a problem, but maybe the introduction of later 
threaded features introduced the conflicts that are becoming apparent now.


 Also the
 fact that it renders the fast preview, which is IMHO becoming the
 center of the workflow, unusable makes it more important.

Yes, the FPW is becoming the center of the workflow IMHO too, and making it 
usable is very important.

I suspect that getting the threading right will be a big and intensive job.  
The more I look at recent reports in the bug tracker, the more dots there are 
to connect and they all seem to lead to the same pesky thing.

Could the conflict be related to the use of different threading libraries?

we have boost threads since revision 4524 [0] for the images cache and cpfind 
uses zthread?

Yuv

[0] http://hugin.hg.sourceforge.net/hgweb/hugin/hugin/rev/4524


signature.asc
Description: This is a digitally signed message part.

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

On July 26, 2011 03:16:43 AM Greg 'groggy' Lehey wrote:
 I don't like the overview mode (maybe because I don't understand
 its purpose)

the relationship between the preview and the overview is like that between a 
map of the earth and a globe.  The overview shows the globe itself, and you 
can turn it around in 3D, while the preview is a flat projection.

The overview is helpful when making full spherical panoramas.  You can look at 
the poles (zenith and nadir) without distortion, while they are too distorted 
in the preview.

  I wonder if the varying behaviour of different systems is being
  influenced by the varying versions of dependent libraries that
  different systems have installed, each reacting a bit differently to
  some bit of code in the overview mode. Just my 2 bobs worth.
 
 Quite possibly.  Did you see the bug report I put in today?  It seems
 that some component crashed on exiting because of a consistency check.
 Since it was exiting anyway, it seems to be harmless, but it's clearly
 related to specific threading libraries.

To make things even more complicated, we are using boost threads and zthreads, 
potentially at the same time.

https://bugs.launchpad.net/hugin/+bug/792896/comments/40

Yuv


signature.asc
Description: This is a digitally signed message part.

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

On July 25, 2011 10:14:08 PM Terry Duell wrote:
 Time to let the computer cool down a bit while I take a break.

thanks for going so quickly through the dissecting process and helping build 
understanding for this pesky bug!

Yuv


signature.asc
Description: This is a digitally signed message part.

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

On July 26, 2011 05:22:42 PM Carl von Einem wrote:
 I didn't see this issue so far on both of my Macs: the oldish one is a
 dual processor PowerPC (2x 2.3 GHz, 6.5 GB RAM), the newer one is a dual
 hexacore Intel Mac with about 24 or 26 GB of RAM.

thanks for reporting.  what version of OS X are you running?  if you have 
acces to OS X Lion, we need some data points there.

 
 I haven't noticed this bug is mentioned by Mac users at all, or did I
 miss some messages?

No Mac users in this bug, but https://bugs.launchpad.net/hugin/+bug/814301 (OS 
X Lion) may be related.

The current hypothesis is that this is a threading issue, and a different 
threading library means a potentially different behavior / effect.  So we have 
to watch for different CPUs (i.e. number of cores / threads) different O/S 
Versions (i.e. how those threads are managed) and different threading 
libraries used (i.e. those present in the system which can be very different 
across same systems; and/or those present in the SDK used to build the 
binaries).

Yuv



signature.asc
Description: This is a digitally signed message part.

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview


Hullo All,

On Mon, 25 Jul 2011 15:05:53 +1000, Terry Duell tdu...@iinet.net.au  
wrote:



Just built and ran 4921 on Fedora 15 x86_64 with the same project as  
previously.

[snip]

I think we can say this one has the bug.


In an attempt to save a bit of time I have been going through my previous  
builds to see if I built a fedora rpm and src.rpm for a rev somewhere  
around 4847...my guess at the next one to try.

I found 4855:8042c8b1dd05 (2010.5.0) which I had build with Fedora 14.
I have attempted to rebuild from the src.rpm on Fedora 15, but run into  
many errors in the vigra code.
I think the vigra bundled with hugin is 1.5-ish, whereas my Fedora 15 has  
vigra 1.7.1.
Is this likely to be the cause of the errors I am seeing when rebuilding  
the 4855:8042c8b1dd05 src.rpm?
For some reason (dense I guess :-)) I am not seeing a way around this at  
the moment.

Any suggestions?

Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview


Hullo All,

On Mon, 25 Jul 2011 16:52:44 +1000, Terry Duell tdu...@iinet.net.au  
wrote:



Hullo All,

On Mon, 25 Jul 2011 15:05:53 +1000, Terry Duell tdu...@iinet.net.au  
wrote:



Just built and ran 4921 on Fedora 15 x86_64 with the same project as  
previously.

[snip]

I think we can say this one has the bug.


In an attempt to save a bit of time I have been going through my  
previous builds to see if I built a fedora rpm and src.rpm for a rev  
somewhere around 4847...my guess at the next one to try.

I found 4855:8042c8b1dd05 (2010.5.0) which I had build with Fedora 14.
I have attempted to rebuild from the src.rpm on Fedora 15, but run into  
many errors in the vigra code.


I have the solution to that issue...I am now running gcc-4.6.0 and hence  
need the patch that Bruno added as changeset 4893 (added stddef.h to  
src/foreign/vigra/vigra/multi_iterator.hxx).


I have now built 4847 and it exhibits the same behaviour as 4921, hence I  
will now try 4814.


Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview


Hullo All,

On Tue, 26 Jul 2011 10:45:16 +1000, Terry Duell tdu...@iinet.net.au  
wrote:



I have now built 4847 and it exhibits the same behaviour as 4921, hence  
I will now try 4814.


Well, I think we are getting close.
I have just built and tested 4814 and it appears to be OK.
In all my previous builds in this exercise, the FPW has been broken at the  
outset...either it opens blank and unresponsive then system asks if I want  
to force a quit, or it opens with pano displayed and unresponsive and then  
crashes.

With this build the FPW does everything I have asked it do.
The problem may still be there but now harder to invoke, but I think the  
best approach is for me to move forward now to (say) 4805 or thereabouts  
(I will look at the change comments and select on the basis of what has  
been changed) and see how it behaves. If it immediately shows the problem  
then I think we can be sure that 4814 really is OK.


Cheers,


--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview


Hullo All,

On Tue, 26 Jul 2011 10:45:16 +1000, Terry Duell tdu...@iinet.net.au  
wrote:



Hullo All,
I have now built 4847 and it exhibits the same behaviour as 4921, hence  
I will now try 4814.



Well, I think we are getting close.
I have just built and tested 4798 and it appears to be OK.
In all my previous builds in this exercise, the FPW has been broken at the  
outset...either it opens blank and unresponsive then system asks if I want  
to force a quit, or it opens with pano displayed and unresponsive and then  
crashes.

With this build the FPW does everything I have asked it do.
The problem may still be there but now harder to invoke, but I think the  
best approach is for me to move forward now to (say) 4805 or thereabouts  
(I will look at the change comments and select on the basis of what has  
been changed) and see how it behaves. If it immediately shows the problem  
then I think we can be sure that 4798 really is OK.


Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview


Hullo Yuv,

On Sun, 24 Jul 2011 23:47:14 +1000, Yuval Levy goo...@levy.ch wrote:


Hi all,

We know that the current Hugin is plagued with a pesky spurious bug that  
is
difficult to reproduce.  It is causing grievance to a lot of users as  
seen on

the ML and on the main tracker ticket where the bug is discussed  [0].

This is most likely caused by a memory leak.  The most difficult part of
fixing a memory leak is to identify where it happens.  Would you like to  
help

nuke this bug?


I will do what I can to help.
If my experiences are anything to go by, it may prove difficult to be sure  
that you have a version with the bug. I had a build a week or more ago  
where the fpw would fail every time hugin was first started after a boot,  
but have not been able to get my latest build (2011.3.0.478a71d71f90) to  
show any problems at all.


[snip]

These are the steps I did, in case somebody wants to step in my shoes.   
This

would be helpful for the continuation of the process after September 5 if
necessary.

cd hugin.default
hg pull  hg up
hg up -r 5057
hg identify -b -n -i

cd ..
mkdir hugin-tarball
cd hugin-tarball
cmake ../hugin.default -DENABLE_LAPACK=YES -DCPACK_BINARY_DEB:BOOL=OFF -
DCPACK_BINARY_NSIS:BOOL=OFF \
  -DCPACK_BINARY_RPM:BOOL=OFF -DCPACK_BINARY_STGZ:BOOL=OFF -
DCPACK_BINARY_TBZ2:BOOL=OFF \
  -DCPACK_BINARY_TGZ:BOOL=ON -DCPACK_BINARY_TZ:BOOL=OFF
make package_source

mv hugin-2011.1.0.tar.bz2 hugin-2011.1.0.5057.tar.bz2



-
just to clarify, what is this next section (to the -) for? Don't  
you already have the tarball you want to upload?



tar xvfj hugin-2011.1.0.5057.tar.bz2
mkdir hugin-tarball-build
cd hugin-tarball-build
cmake ../hugin-2011.1.0 -DENABLE_LAPACK=YES -DCPACK_BINARY_DEB:BOOL=ON -
DCPACK_BINARY_NSIS:BOOL=OFF \
  -DCPACK_BINARY_RPM:BOOL=OFF -DCPACK_BINARY_STGZ:BOOL=OFF -
DCPACK_BINARY_TBZ2:BOOL=OFF \
  -DCPACK_BINARY_TGZ:BOOL=OFF -DCPACK_BINARY_TZ:BOOL=OFF \
  -DBUILD_HSI:BOOL=ON -DSWIG_EXECUTABLE=/usr/bin/swig2.0
make package
sudo dpkg -i hugin-2011.1.0-Linux.deb


--

cd ..
rsync --partial --progress -e ssh hugin-2011.1.0.5057.tar.bz2
yuv,hu...@frs.sourceforge.net:/home/frs/project/h/hu/hugin/hugin/pesky-bug/

The tarball is available for download at [3] now.


Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview


Hullo Yuv,

On Sun, 24 Jul 2011 23:47:14 +1000, Yuval Levy goo...@levy.ch wrote:



So the error is most likely between 4780 and 5435.  That would be 656
different versions of Hugin, but it is much less than that: many of these
changesets are not on the default branch.


I just built and tested 5057 on Fedora 15 x86_64 and it has the bug.
I am running an nvidia gt9600 card and the nvidia prop. driver.

I will add a comment to the tracker.

Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview

2011-07-24 Thread Yuval Levy

Hi Terry,

On July 24, 2011 07:21:16 PM Terry Duell wrote:
 just to clarify, what is this next section (to the -) for? Don't
 you already have the tarball you want to upload?

never upload a tarball without having tested at least that it builds on your 
machine.

the next tarball is up: 

https://sourceforge.net/projects/hugin/files/hugin/pesky-
bug/hugin-2010.5.0.4921.tar.bz2/download

Yuv


signature.asc
Description: This is a digitally signed message part.

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview


Hullo Yuv,

On Mon, 25 Jul 2011 14:05:31 +1000, Yuval Levy goo...@levy.ch wrote:


Hi Terry,

On July 24, 2011 07:21:16 PM Terry Duell wrote:

just to clarify, what is this next section (to the -) for? Don't
you already have the tarball you want to upload?


never upload a tarball without having tested at least that it builds on  
your machine.


Yes, thank you.



the next tarball is up:

https://sourceforge.net/projects/hugin/files/hugin/pesky-
bug/hugin-2010.5.0.4921.tar.bz2/download


Working on this one now.

Cheers,
--
Regards,
Terry Duell

--
You received this message because you are subscribed to the Google Groups Hugin and 
other free panoramic software group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx

Re: [hugin-ptx] CRITICAL: join us to nuke the pesky bug that is plaguing the fast preview