Hi Erick,

This error may indicate some problem unrelated to libgd. Could you do me a favor and insert the following diagnostic message into line 206 of Karyotype.pm (easiest to do it directly in /usr/local/lib/perl5/site_perl...), and then send me the error log output?

 warn "n1=$n1, n2=$n2";

Thanks,

Lincoln

On Tue, May 26, 2009 at 10:47 AM, Erick Antezana <erick.antezana@...> wrote:

Lincoln,

I have just installed libgd 2.0.36RC1 and the problem is still there (I checked that I do not need to recompile GD; thanks for the ldd tip!).

here is an excerpt of my apache error log file:

 [error]  which: no inkscape in (/tools/bioinfo/bioim_bin:/usr/java/default/bin:/usr/local/ant/bin:/usr/bin:/bin:/usr/lib/oracle/11.1.0.1/client64/bin:/tools/bioinfo/bin:/tools/bioinfo/app/blast2/bin:/tools/bioinfo/app/Emboss/bin:/tools/bioinfo/app/psort/:/tools/bioinfo/app/tssp:/tools/bioinfo/app/t_coffee/plugins/linux),
 [error]  Argument "" isn't numeric in numeric comparison (<=>) at,
 [error]  \t/usr/local/lib/perl5/site_perl/5.10.0/x86_64-linux/Bio/Graphics/Karyotype.pm line 210 (#1),
 [error]      (W numeric) The indicated string was fed as an argument to an operator,
 [error]      that expected a numeric value instead.  If you're fortunate the message,
 [error]      will identify which operator was so unfortunate.,

thanks again,

Erick

2009/5/26 Lincoln Stein <lincoln.stein@...>

Hi Erick,

You said:

the problem is still there although now the apache error log is more informative

What do you see in the error log?

Perhaps this isn't the libgd bug at all but something different. That would be a relief to me.

inkscape doesn't use libgd, so you don't have to worry about that. You only need to recompile GD if it is linked to libgd dynamically. The way to do this is to find GD.so, and then:

% ldd /usr/lib/perl5/auto/GD/GD.so
    linux-gate.so.1 =>  (0xb809b000)
    libjpeg.so.62 => /usr/lib/libjpeg.so.62 (0xb802e000)
    libfreetype.so.6 => /usr/lib/libfreetype.so.6 (0xb7fb7000)
    libpng12.so.0 => /usr/lib/libpng12.so.0 (0xb7f90000)
    libz.so.1 => /lib/libz.so.1 (0xb7f7a000)
    libm.so.6 => /lib/tls/i686/cmov/libm.so.6 (0xb7f54000)
    libgd.so.2 => /usr/lib/libgd.so.2 (0xb7f13000)
    libc.so.6 => /lib/tls/i686/cmov/libc.so.6 (0xb7db0000)
    /lib/ld-linux.so.2 (0xb809c000)

This shows that my version of GD is dynamically linked and doesn't need to be recompiled.

Lincoln

On Tue, May 26, 2009 at 10:14 AM, Erick Antezana <erick.antezana@...> wrote:

Lincoln,

 I forgot to mention that I am not getting the "child process terminated" in my error log...

 on the other hand, do I need to recompile any other package apart from GD? perhaps inkscape?

thanks
Erick

2009/5/26 Lincoln Stein <lincoln.stein@...>

Hi Erick,

Are you getting the "child process terminated" error in your apache log?

I am using GD version 2.39 without getting crashes, but the GD version is not related to the crash, as it occurs deep in libgd. However, after installing the new libgd, you may need to recompile GD.

Also, try adjusting the appearance of the karyotype:

 [builtin:karyotype]
 chrom_width  =  20
 chrom_height = 150

and variants of the above.

I had thought this problem was fixed, but apparently I need to do some more work.

Lincoln

On Tue, May 26, 2009 at 9:25 AM, Erick Antezana <erick.antezana@...> wrote:

Hi Lincoln,

 Our system had 2.0.33, we have just installed the lastest release (2.0.35) but the problem is still there although now the apache error log is more informative.... I have checked on the libgd site (http://www.libgd.org/Downloads) that the latest release is 2.0.35 but there is also a candidate release (2.0.36RC1) as well as a develompent one (2.1.0.-dev).  which version of GD would you recommend to use? (I have Bio::Graphics version 1.94)

thanks,
Erick

2009/5/25 Lincoln Stein <lincoln.stein@...>

Hi Erick,

There is a persistent bug in libgd that the karyotype display (used to show multiple hits; see enclosed) somehow triggers. Usually you will see a "child process unexpectedly terminated" in one of your apache logs -- if you are using virtual hosts, then it will be in the main error log, and not in the virutal host one.

Here are steps to try to fix the problem:

1. Install the latest Bio::Graphics from CPAN. It contains code that works around the problem.
2. Recompile libgd from the latest source.
3. Complain to me if you try (1) and (2) and the problem persists.

Lincoln

On Mon, May 25, 2009 at 11:21 AM, Erick Antezana <erick.antezana@...> wrote:

Hi,

  Our local installation of GBrowse 2.0 was successfully installed some days ago and we were testing the "Search" capabilities on specific regions and landmarks. If we look for a particular region, GBrowse is able to find it and show it. However, we got some problems to find a landmark (free text): if we use for instance the sample data source "Yeast chromosomes 1+2 (basic)" and we look for "membrane trafficking", the server shows on the top:

Yeast chromosomes 1+2 (basic): The following 66 regions match your request.

However, no region is shown below. I have checked on the Apache error log file, and there is no message. We have also tried the "Search" functionality out with our internal data and the behaviour is identical. Any suggestions?

cheers,
Erick

------------------------------------------------------------------------------
Register Now for Creativity and Technology (CaT), June 3rd, NYC. CaT
is a gathering of tech-side developers & brand creativity professionals. Meet
the minds behind Google Creative Lab, Visual Complexity, Processing, &
iPhoneDevCamp asthey present alongside digital heavyweights like Barbarian
Group, R/GA, & Big Spaceship. http://www.creativitycat.com
_______________________________________________
Gmod-gbrowse mailing list
Gmod-gbrowse@...
https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse

--
Lincoln D. Stein
Director, Informatics and Biocomputing Platform
Ontario Institute for Cancer Research
101 College St., Suite 800
Toronto, ON, Canada M5G0A3
416 673-8514
Assistant: Renata Musa <Renata.Musa@...>

--
Lincoln D. Stein
Director, Informatics and Biocomputing Platform
Ontario Institute for Cancer Research
101 College St., Suite 800
Toronto, ON, Canada M5G0A3
416 673-8514
Assistant: Renata Musa <Renata.Musa@...>

--
Lincoln D. Stein
Director, Informatics and Biocomputing Platform
Ontario Institute for Cancer Research
101 College St., Suite 800
Toronto, ON, Canada M5G0A3
416 673-8514
Assistant: Renata Musa <Renata.Musa@...>

--
Lincoln D. Stein
Director, Informatics and Biocomputing Platform
Ontario Institute for Cancer Research
101 College St., Suite 800
Toronto, ON, Canada M5G0A3
416 673-8514
Assistant: Renata Musa <Renata.Musa@...>