Re: [Geeqie-devel] Resource leak in Geeqie 1.2 on debian jessie?

2016-03-08 Thread John Stoffel

Hartmut> You are certainly right that I can (and maybe should) run a
Hartmut> real 64bit kernel. I am already in the process of doing it,
Hartmut> but there are so many programs to install and configure that
Hartmut> it is going to take some time until my 64bit jessie can take
Hartmut> over.

I don't think you need to do anything more than install the 64bit
kernel from the current distro and let apt-get install all the
dependencies for you.  Thne you just reboot and it should (knock on
wood!) work. 

Hartmut> But I had the same problem with 4G of RAM.  So this is not a
Hartmut> "more than 4GBytes of RAM is too much on a 32bit kernel" only
Hartmut> problem.

Sure, but I think the problem is that geeqie (or the libraries it
depends on) are too aggresive with memory usage, which causes problems
when you exhaust all the memory it can access.

When running a 32bit program, even on a 12gb machine, a process can
only access about 3gb worth of data at any one time no matter what as
I recall.  Now you can have multiple processes using lots of memory,
but I think you can only use 3gb (not 4gb, since you need to leave the
other 1gb for system libraries and such) of process memory space. And
thinking on it, it might even be limited to just 2g/2g split.  

Hartmut> I would like to understand why the system becomes unstable.
Hartmut> "Too much RAM" sounds too easy. There must be at least one
Hartmut> bug to find.

I suspect geeqie does have memory problems.  I'll have to try and see
if I can find the time to spin it up in a 32bit VM and see what
happens.  Or even on a rapsberryPi as a test.  

Hartmut> And I want to run all my tests on 64bit jessie as well.  I
Hartmut> have not seen geeqie perform well under that condition yet.

If you can get this same issue to happen under a 64bit OS, kernel and
geeqie version, then that would be interesting for sure.  And
something I could maybe even help out with testing out.

John

--
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
http://makebettercode.com/inteldaal-eval
___
Geeqie-devel mailing list
Geeqie-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geeqie-devel


Re: [Geeqie-devel] Resource leak in Geeqie 1.2 on debian jessie?

2016-03-08 Thread Hartmut Niemann
Hi John!

You are certainly right that I can (and maybe should) run a
real 64bit kernel. I am already in the process of doing it,
but there are so many programs to install and configure
that it is going to take some time until my
64bit jessie can take over.

But I had the same problem with 4G of RAM.
So this is not a "more than 4GBytes of RAM is too much on a 32bit 
kernel" only problem.

I would like to understand why the system becomes unstable.
"Too much RAM" sounds too easy. There must be at least
one bug to find.

And I want to run all my tests on 64bit jessie as well.
I have not seen geeqie perform well under that condition yet.

Hartmut


Am 08.03.2016 um 22:26 schrieb John Stoffel:
> Klaus> Am Mo den  7. Mär 2016 um 23:12 schrieb Hartmut Niemann:
>>> $ cat /proc/version
>>> Linux version 3.16.0-4-686-pae (debian-ker...@lists.debian.org) (gcc version
>>> 4.8.4 (Debian 4.8.4-1) ) #1 SMP Debian 3.16.7-ckt20-1+deb8u3 (2016-01-17)
> I really think this is your problem in alot of ways.  You have an i686
> kernel (32bit) with the PAE extensions installed because you have 12gb
> of RAM in the system.  Why aren't you running a x86_64 (64bit) kernel
> and userland?  You can certainly install the 32bit libraries as well
> if you need them for an application.
>
> But moving to a pure 64bit kernel (can you send the output of cat
> /proc/cpuinfo?)  with a mixed 64bit/32bit userland (if needed) you'll
> get a much bigger address space.
>
> Now I can certainly see how geeqie, or possibly the toolkit and
> libraries, might not handle large amounts of highmem properly on a
> 32bit system using PAE.
>
> In my mind, it's just not worth the hassle.  How hard would it be to
> install a 64bit kernel on this system?
>
> John
>
> --
> Transform Data into Opportunity.
> Accelerate data analysis in your applications with
> Intel Data Analytics Acceleration Library.
> Click to learn more.
> http://makebettercode.com/inteldaal-eval
> ___
> Geeqie-devel mailing list
> Geeqie-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/geeqie-devel


--
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
http://makebettercode.com/inteldaal-eval
___
Geeqie-devel mailing list
Geeqie-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geeqie-devel


Re: [Geeqie-devel] Resource leak in Geeqie 1.2 on debian jessie?

2016-03-08 Thread John Stoffel

Klaus> Am Mo den  7. Mär 2016 um 23:12 schrieb Hartmut Niemann:
>> $ cat /proc/version
>> Linux version 3.16.0-4-686-pae (debian-ker...@lists.debian.org) (gcc version
>> 4.8.4 (Debian 4.8.4-1) ) #1 SMP Debian 3.16.7-ckt20-1+deb8u3 (2016-01-17)

I really think this is your problem in alot of ways.  You have an i686
kernel (32bit) with the PAE extensions installed because you have 12gb
of RAM in the system.  Why aren't you running a x86_64 (64bit) kernel
and userland?  You can certainly install the 32bit libraries as well
if you need them for an application.

But moving to a pure 64bit kernel (can you send the output of cat
/proc/cpuinfo?)  with a mixed 64bit/32bit userland (if needed) you'll
get a much bigger address space.

Now I can certainly see how geeqie, or possibly the toolkit and
libraries, might not handle large amounts of highmem properly on a
32bit system using PAE.

In my mind, it's just not worth the hassle.  How hard would it be to
install a 64bit kernel on this system?

John

--
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
http://makebettercode.com/inteldaal-eval
___
Geeqie-devel mailing list
Geeqie-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geeqie-devel


Re: [Geeqie-devel] Resource leak in Geeqie 1.2 on debian jessie?

2016-03-08 Thread Klaus Ethgen
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA512

Hi Folks,

Am Mo den  7. Mär 2016 um 23:12 schrieb Hartmut Niemann:
> $ cat /proc/version
> Linux version 3.16.0-4-686-pae (debian-ker...@lists.debian.org) (gcc version
> 4.8.4 (Debian 4.8.4-1) ) #1 SMP Debian 3.16.7-ckt20-1+deb8u3 (2016-01-17)
> 
> $ geeqie --version
> Geeqie 1.2

That explains somehow.

First I really believe there are some troubles in libgtk and libglib in
stable debian. Then that version is somehow older and there was some
fixes in unstable.

I'll try to get the current 1.2.2 or even 1.2.3 back to stable.

> Unfortunately version 1.2.2 from testing and unstable has some dependencies
> (Upgraded libraries) that I can't fulfill at the moment.

Then, build it from source.

On unstable:
   apt-get source geeqie

Copy that stuff to stable and then on stable as root:
   apt-get build-dep geeqie

And as normal user in the copied directory:
   debuild

After that you should have a nice .deb package that you can install.

> My normal workflow is that I mark some pictures as "1" and then "select all
> "1""
> and "delete".
> I observed that usually the problems start when I deleted a bunch of files.
> I am not sure, but if I select "show only "1"" and then "select all "1"" and
> then "delete",
> the propability of things going bad is even higher.

We might need to dig deeper into that. But first it has to be ruled out
that:
- - No problems with libs are shown (What I suspect as I seen some
  problems like that too in the past)
- - Bug is still in the most current version

> I don't use the gnome trash, by the way.

:-) Good to hear. ;-)

However, geeqie is using libgtk and glib.

> Can I recompile the latest debian package (1.2.2) in my current environment
> or do I need to upgrade libs for that?

Yes. See above. That should work.

Am Di den  8. Mär 2016 um 18:26 schrieb Ian Munsie:
> I hit memory issues as well (all versions I've tried - 1.2 from source,
> Debian's version, even going back to 1.0 still hits it), but it's only
> virtual memory usage - resident memory stays contained. On 32bit geeqie
> will start behaving badly once it hits 4GB virtual memory (which is easy to
> hit scrolling through a directory with a couple of hundred .cr2 images),
> first auto rotation stops working then an image or two later it stops
> loading further images. It does not crash. The 64bit version doesn't
> eliminate the problem, but has enough virtual address space that it is not
> practical to hit.

That's exactly what I seen too in the past. But the problem is gone now
(on debian unstable). So I really have some guess that it is the
libraries.

You can try to run it under valgrind. But expect it to be hell slow.

Regards
   Klaus
- -- 
Klaus Ethgen  http://www.ethgen.ch/
pub  4096R/4E20AF1C 2011-05-16   Klaus Ethgen 
Fingerprint: 85D4 CA42 952C 949B 1753  62B3 79D0 B06F 4E20 AF1C
-BEGIN PGP SIGNATURE-
Version: GnuPG v1
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=fNji
-END PGP SIGNATURE-

--
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
http://makebettercode.com/inteldaal-eval
___
Geeqie-devel mailing list
Geeqie-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geeqie-devel


Re: [Geeqie-devel] Resource leak in Geeqie 1.2 on debian jessie?

2016-03-08 Thread Ian Munsie
I hit memory issues as well (all versions I've tried - 1.2 from source,
Debian's version, even going back to 1.0 still hits it), but it's only
virtual memory usage - resident memory stays contained. On 32bit geeqie
will start behaving badly once it hits 4GB virtual memory (which is easy to
hit scrolling through a directory with a couple of hundred .cr2 images),
first auto rotation stops working then an image or two later it stops
loading further images. It does not crash. The 64bit version doesn't
eliminate the problem, but has enough virtual address space that it is not
practical to hit.

-Ian

-- 
http://darkstarsword.net
http://darkstarshout.blogspot.com
http://github.com/DarkStarSword
--
Please avoid sending me Word or PowerPoint attachments.
See http://www.gnu.org/philosophy/no-word-attachments.html
--
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
http://makebettercode.com/inteldaal-eval___
Geeqie-devel mailing list
Geeqie-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geeqie-devel


Re: [Geeqie-devel] Resource leak in Geeqie 1.2 on debian jessie?

2016-03-08 Thread Hartmut Niemann

Hi!

John, you asked about versions:

$ cat /proc/version
Linux version 3.16.0-4-686-pae (debian-ker...@lists.debian.org) (gcc 
version 4.8.4 (Debian 4.8.4-1) ) #1 SMP Debian 3.16.7-ckt20-1+deb8u3 
(2016-01-17)


$ geeqie --version
Geeqie 1.2

$ ldd `which geeqie`
linux-gate.so.1 (0xb77d5000)
libgtk-x11-2.0.so.0 => /usr/lib/i386-linux-gnu/libgtk-x11-2.0.so.0 
(0xb72b8000)
libgdk-x11-2.0.so.0 => /usr/lib/i386-linux-gnu/libgdk-x11-2.0.so.0 
(0xb71f7000)
libpangocairo-1.0.so.0 => 
/usr/lib/i386-linux-gnu/libpangocairo-1.0.so.0 (0xb71e8000)

libatk-1.0.so.0 => /usr/lib/i386-linux-gnu/libatk-1.0.so.0 (0xb71c)
libcairo.so.2 => /usr/lib/i386-linux-gnu/libcairo.so.2 (0xb7077000)
libgdk_pixbuf-2.0.so.0 => 
/usr/lib/i386-linux-gnu/libgdk_pixbuf-2.0.so.0 (0xb704f000)

libgio-2.0.so.0 => /usr/lib/i386-linux-gnu/libgio-2.0.so.0 (0xb6e99000)
libpangoft2-1.0.so.0 => 
/usr/lib/i386-linux-gnu/libpangoft2-1.0.so.0 (0xb6e8)
libpango-1.0.so.0 => /usr/lib/i386-linux-gnu/libpango-1.0.so.0 
(0xb6e2e000)
libgobject-2.0.so.0 => /usr/lib/i386-linux-gnu/libgobject-2.0.so.0 
(0xb6dd)

libglib-2.0.so.0 => /lib/i386-linux-gnu/libglib-2.0.so.0 (0xb6ca8000)
libfontconfig.so.1 => /usr/lib/i386-linux-gnu/libfontconfig.so.1 
(0xb6c66000)
libfreetype.so.6 => /usr/lib/i386-linux-gnu/libfreetype.so.6 
(0xb6bb3000)
libgthread-2.0.so.0 => /usr/lib/i386-linux-gnu/libgthread-2.0.so.0 
(0xb6bb)

libjpeg.so.62 => /usr/lib/i386-linux-gnu/libjpeg.so.62 (0xb6b51000)
libtiff.so.5 => /usr/lib/i386-linux-gnu/libtiff.so.5 (0xb6ad4000)
liblcms2.so.2 => /usr/lib/i386-linux-gnu/liblcms2.so.2 (0xb6a72000)
libexiv2.so.13 => /usr/lib/i386-linux-gnu/libexiv2.so.13 (0xb67e6000)
liblua5.1.so.0 => /usr/lib/i386-linux-gnu/liblua5.1.so.0 (0xb67b4000)
liblirc_client.so.0 => /usr/lib/liblirc_client.so.0 (0xb67ad000)
libstdc++.so.6 => /usr/lib/i386-linux-gnu/libstdc++.so.6 (0xb66bb000)
libm.so.6 => /lib/i386-linux-gnu/i686/cmov/libm.so.6 (0xb6675000)
libgcc_s.so.1 => /lib/i386-linux-gnu/libgcc_s.so.1 (0xb6657000)
libpthread.so.0 => /lib/i386-linux-gnu/i686/cmov/libpthread.so.0 
(0xb663b000)

libc.so.6 => /lib/i386-linux-gnu/i686/cmov/libc.so.6 (0xb648e000)
libgmodule-2.0.so.0 => /usr/lib/i386-linux-gnu/libgmodule-2.0.so.0 
(0xb6489000)

libX11.so.6 => /usr/lib/i386-linux-gnu/libX11.so.6 (0xb6337000)
libXcomposite.so.1 => /usr/lib/i386-linux-gnu/libXcomposite.so.1 
(0xb6332000)

libXdamage.so.1 => /usr/lib/i386-linux-gnu/libXdamage.so.1 (0xb632e000)
libXfixes.so.3 => /usr/lib/i386-linux-gnu/libXfixes.so.3 (0xb6327000)
libXrender.so.1 => /usr/lib/i386-linux-gnu/libXrender.so.1 (0xb631b000)
libXinerama.so.1 => /usr/lib/i386-linux-gnu/libXinerama.so.1 
(0xb6317000)

libXi.so.6 => /usr/lib/i386-linux-gnu/libXi.so.6 (0xb6303000)
libXrandr.so.2 => /usr/lib/i386-linux-gnu/libXrandr.so.2 (0xb62f7000)
libXcursor.so.1 => /usr/lib/i386-linux-gnu/libXcursor.so.1 (0xb62eb000)
libXext.so.6 => /usr/lib/i386-linux-gnu/libXext.so.6 (0xb62d6000)
libpixman-1.so.0 => /usr/lib/i386-linux-gnu/libpixman-1.so.0 
(0xb621d000)

libpng12.so.0 => /lib/i386-linux-gnu/libpng12.so.0 (0xb61ef000)
libxcb-shm.so.0 => /usr/lib/i386-linux-gnu/libxcb-shm.so.0 (0xb61eb000)
libxcb-render.so.0 => /usr/lib/i386-linux-gnu/libxcb-render.so.0 
(0xb61e)

libxcb.so.1 => /usr/lib/i386-linux-gnu/libxcb.so.1 (0xb61ba000)
libz.so.1 => /lib/i386-linux-gnu/libz.so.1 (0xb619d000)
librt.so.1 => /lib/i386-linux-gnu/i686/cmov/librt.so.1 (0xb6193000)
libselinux.so.1 => /lib/i386-linux-gnu/libselinux.so.1 (0xb616b000)
libresolv.so.2 => /lib/i386-linux-gnu/i686/cmov/libresolv.so.2 
(0xb6154000)
libharfbuzz.so.0 => /usr/lib/i386-linux-gnu/libharfbuzz.so.0 
(0xb60f7000)

libthai.so.0 => /usr/lib/i386-linux-gnu/libthai.so.0 (0xb60ed000)
libffi.so.6 => /usr/lib/i386-linux-gnu/libffi.so.6 (0xb60e4000)
libpcre.so.3 => /lib/i386-linux-gnu/libpcre.so.3 (0xb6072000)
libexpat.so.1 => /lib/i386-linux-gnu/libexpat.so.1 (0xb6049000)
liblzma.so.5 => /lib/i386-linux-gnu/liblzma.so.5 (0xb602)
libjbig.so.0 => /usr/lib/i386-linux-gnu/libjbig.so.0 (0xb6011000)
libdl.so.2 => /lib/i386-linux-gnu/i686/cmov/libdl.so.2 (0xb600b000)
/lib/ld-linux.so.2 (0xb77d8000)
libXau.so.6 => /usr/lib/i386-linux-gnu/libXau.so.6 (0xb6007000)
libXdmcp.so.6 => /usr/lib/i386-linux-gnu/libXdmcp.so.6 (0xb6001000)
libgraphite2.so.3 => /usr/lib/i386-linux-gnu/libgraphite2.so.3 
(0xb5fda000)

libdatrie.so.1 => /usr/lib/i386-linux-gnu/libdatrie.so.1 (0xb5fd)

Unfortunately version 1.2.2 from testing and unstable has some dependencies
(Upgraded libraries) that I can't fulfill at the moment.


My normal workflow is that I mark some pictures as "1" and then "select 
all "1""

and "delete".
I observed that usually the problems start when I deleted a bunch of files.
I am not sure, but if I select "show only "1"" and then "