From edder at tkwsping.nl  Sun Feb  4 10:18:10 2007
From: edder at tkwsping.nl (Edwin)
Date: Sun, 04 Feb 2007 15:18:10 -0000
Subject: [Tioga-users] tioga 1.4
Message-ID: <op.tm8csk1xngac0g@localhost>

Hi,

I'm curious to know if anyone else has problems with tioga-1.4?

The gem seems to install the libs wrongly. There is no directory Tioga  
under the libs, so when tioga.rb requires Tioga/FigureMaker it can't find  
it. The easiest way around it was to make the directory and copy  
everything in the lib dir to it.

Next I get an segmentation fault when trying to plot a plot:
edwin at vimes:~/documents/rhul/popdyn/article$ tioga popdyn_plots.rb -13
loading /home/edwin/documents/rhul/popdyn/article/popdyn_plots.rb
/usr/lib/ruby/site_ruby/1.8/Tioga/FigMkr.rb:897: [BUG] Segmentation fault
ruby 1.8.5 (2006-12-04) [i686-linux]

Aborted

The corresponding line in FigMkr.rb:
         show_axis(self.xaxis_loc) if self.xaxis_visible

(This happened with both the gem version and the other version).
I'm using gentoo linux, gcc 4.1.1.

On a more positive note, the new commandline tools seem to be very nice  
and thanks for providing a gem version, which makes installing and  
updating a lot easier. I'm also curious about the new arrow fixes, because  
I did have some problems with plotting vectorfields (when the arrowheads  
were scaled down they wouldn't attach to the tail anymore)

Edwin

From edder at tkwsping.nl  Sun Feb  4 13:54:35 2007
From: edder at tkwsping.nl (Edwin)
Date: Sun, 04 Feb 2007 18:54:35 -0000
Subject: [Tioga-users] tioga 1.4
In-Reply-To: <FE20FECC-986D-442F-ADB7-A504AC742FD4@kitp.ucsb.edu>
References: <op.tm8csk1xngac0g@localhost>
	<FE20FECC-986D-442F-ADB7-A504AC742FD4@kitp.ucsb.edu>
Message-ID: <op.tm8ms9yhngac0g@localhost>

On Sun, 04 Feb 2007 16:50:28 -0000, Bill Paxton <paxton at kitp.ucsb.edu>  
wrote:

> The segmentation fault when making a plot is something I'd like to look  
> into.  Does it happen with all plots?  Or only with your  
> "popdyn_plots"?  Please send me more info so I can try to recreate the  
> problem.

It happens with every plot I tried (for example sine.rb and plots.rb which  
are both part of the tioga-sample package)

I'm afraid I don't really know how I would go about debugging this, but I  
attached a strace log

Edwin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: tioga_trace
Type: application/octet-stream
Size: 124640 bytes
Desc: not available
Url : http://rubyforge.org/pipermail/tioga-users/attachments/20070204/e0f4dddd/attachment-0001.obj 

From vincent.fourmond at 9online.fr  Sun Feb  4 15:14:39 2007
From: vincent.fourmond at 9online.fr (Vincent Fourmond)
Date: Sun, 04 Feb 2007 21:14:39 +0100
Subject: [Tioga-users] [Fwd: Re:  tioga 1.4]
Message-ID: <45C63EAF.9000404@9online.fr>

[sorry, always forget to reply to the list]
Edwin wrote:
> Hi,
> 
> I'm curious to know if anyone else has problems with tioga-1.4?
> 
> The gem seems to install the libs wrongly. There is no directory Tioga  
> under the libs, so when tioga.rb requires Tioga/FigureMaker it can't find  
> it. The easiest way around it was to make the directory and copy  
> everything in the lib dir to it.

  I'm curious about this part. Could you output the command tree whithin
your gem directory ?

BTW, you need to have
require 'rubygems'
before
require 'Tioga/FigureMaker'
in the script, else the rubygem lookup mechanism simply doesn't work and
require will fail. [I and Bill and JJ got caught by this as well, so
don't feel bad if that is the case ;-)...]

> Next I get an segmentation fault when trying to plot a plot:
> edwin at vimes:~/documents/rhul/popdyn/article$ tioga popdyn_plots.rb -13
> loading /home/edwin/documents/rhul/popdyn/article/popdyn_plots.rb
> /usr/lib/ruby/site_ruby/1.8/Tioga/FigMkr.rb:897: [BUG] Segmentation fault
> ruby 1.8.5 (2006-12-04) [i686-linux]
> 
> Aborted
> 
> The corresponding line in FigMkr.rb:
>          show_axis(self.xaxis_loc) if self.xaxis_visible
> 
> (This happened with both the gem version and the other version).
> I'm using gentoo linux, gcc 4.1.1.

  If this one is a reproducible bug, here is how to get useful input for
 me to debug: run

gdb ruby
[then, whithin gdb]
run your_script.rb and its command line arguments
[then, when it fails with a segfault,]
bt

  This should allow us to track the problem. My guess is it might be
packaging-related (ie, my fault).

  Cheers, and thanks for the input !

	Vincent

-- 
Vincent Fourmond, PhD student (not for long anymore)
http://vincent.fourmond.neuf.fr/


From edder at tkwsping.nl  Mon Feb  5 06:21:01 2007
From: edder at tkwsping.nl (edder)
Date: Mon, 5 Feb 2007 12:21:01 +0100 (CET)
Subject: [Tioga-users] [Fwd: Re:  tioga 1.4]
Message-ID: <51255.134.219.41.106.1170674461.squirrel@wm.tkwsping.nl>

> I'm curious about this part. Could you output the command tree whithin
> your gem directory ?

edwin at zaphod:~/tmp/samples$ /usr/bin/ls
/usr/lib/ruby/gems/1.8/gems/tioga-1.4/
Tioga_README  lgpl.txt  lib  split  tests

edwin at zaphod:~/tmp/samples$ /usr/bin/ls
/usr/lib/ruby/gems/1.8/gems/tioga-1.4/lib
Arcs_and_Circles.rb        FigureMaker.so        Strokes.rb
ColorConstants.rb          Figures_and_Plots.rb  TeX_Text.rb
Colorbars.rb               Flate.so              TexPreamble.rb
Colormaps.rb               Function.so           Titles_and_Labels.rb
Coordinate_Conversions.rb  Function_extras.rb    Transparency.rb
Creating_Paths.rb          Images.rb             Using_Paths.rb
Doc.rb                     Legends.rb            Utils.rb
Dtable.so                  MarkerConstants.rb    X_and_Y_Axes.rb
Dtable_extras.rb           Markers.rb            irb_tioga.rb
Dvector.so                 Numeric_extras.rb     tioga.rb
Dvector_extras.rb          Page_Frame_Bounds.rb  tioga_ui.rb
Executive.rb               Rectangles.rb         tioga_ui_cmds.rb
FigMkr.rb                  Shading.rb
FigureConstants.rb         Special_Paths.rb



> BTW, you need to have
> require 'rubygems'
> before
> require 'Tioga/FigureMaker'
> in the script, else the rubygem lookup mechanism simply doesn't work and
> require will fail. [I and Bill and JJ got caught by this as well, so
> don't feel bad if that is the case ;-)...]

I'm pretty sure that's all in order, especially since it works as soon as
I move everything to Tioga:
edwin at zaphod:~$ cat /usr/bin/tioga
#!/usr/bin/env ruby
#
# This file was generated by RubyGems.
#
# The application 'tioga' is installed as part of a gem, and
# this file is here to facilitate running it.
#

require 'rubygems'
version = "> 0"
if ARGV.size > 0 && ARGV[0][0]==95 && ARGV[0][-1]==95
    if Gem::Version.correct?(ARGV[0][1..-2])
      version = ARGV[0][1..-2]
      ARGV.shift
    end
end
require_gem 'tioga', version
load 'tioga'

edwin at zaphod:~/tmp/samples$ tioga sample.rb -l
/usr/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:21:in `require__':
no such file to load -- Tioga/tioga_ui.rb (LoadError)
          from /usr/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:21:in
`require'
          from /usr/lib/ruby/gems/1.8/gems/tioga-1.4/split/scripts/tioga:2
          from /usr/bin/tioga:18:in `load'
          from /usr/bin/tioga:18

I also tried the following, but got the same error message:
require 'rubygems'
version = "> 0"
if ARGV.size > 0 && ARGV[0][0]==95 && ARGV[0][-1]==95
    if Gem::Version.correct?(ARGV[0][1..-2])
      version = ARGV[0][1..-2]
      ARGV.shift
    end
end
#require_gem 'tioga', version
#load 'tioga'
require 'Tioga/FigureMaker'


>> Next I get an segmentation fault when trying to plot a plot:
>> edwin at vimes:~/documents/rhul/popdyn/article$ tioga popdyn_plots.rb -13
>> loading /home/edwin/documents/rhul/popdyn/article/popdyn_plots.rb
>> /usr/lib/ruby/site_ruby/1.8/Tioga/FigMkr.rb:897: [BUG] Segmentation
>> fault
>> ruby 1.8.5 (2006-12-04) [i686-linux]
>>
>> Aborted
>>
>> The corresponding line in FigMkr.rb:
>>          show_axis(self.xaxis_loc) if self.xaxis_visible
>>
>> (This happened with both the gem version and the other version).
>> I'm using gentoo linux, gcc 4.1.1.
>
>   If this one is a reproducible bug, here is how to get useful input for
>  me to debug: run
>
> gdb ruby
> [then, whithin gdb]
> run your_script.rb and its command line arguments
> [then, when it fails with a segfault,]
> bt
>
>   This should allow us to track the problem. My guess is it might be
> packaging-related (ie, my fault).

(gdb) run /usr/bin/tioga sine.rb -0
Starting program: /usr/bin/ruby /usr/bin/tioga sine.rb -0
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
loading /home/edwin/tmp/samples/sine.rb

Program received signal SIGSEGV, Segmentation fault.
0xb79ef9e8 in End_Axis_Standard_State at plt ()
     from /usr/lib/ruby/gems/1.8/gems/tioga-1.4/lib/Tioga/FigureMaker.so
(gdb)

BTW I now tried it on another machine (although still gentoo and gcc
4.1.1) and still the same problem

The end :)

Edwin




From vincent.fourmond at 9online.fr  Mon Feb  5 16:56:20 2007
From: vincent.fourmond at 9online.fr (Vincent Fourmond)
Date: Mon, 05 Feb 2007 22:56:20 +0100
Subject: [Tioga-users] [Fwd: Re:  [Fwd: Re:  tioga 1.4]]
In-Reply-To: <39437.134.219.41.106.1170684393.squirrel@wm.tkwsping.nl>
References: <39437.134.219.41.106.1170684393.squirrel@wm.tkwsping.nl>
Message-ID: <45C7A804.8060500@9online.fr>

edder wrote:
> I send this to the list, but it doesn't seem to appear on there, so here
> is a copy of it.
> 
> ---------------------------- Original Message ----------------------------
> Subject: Re: [Tioga-users] [Fwd: Re:  tioga 1.4]
> From:    "edder" <edder at tkwsping.nl>
> Date:    Mon, February 5, 2007 12:21 pm
> To:      tioga-users at rubyforge.org
> --------------------------------------------------------------------------
> 
>> I'm curious about this part. Could you output the command tree whithin
>> your gem directory ?
> 
> edwin at zaphod:~/tmp/samples$ /usr/bin/ls
> /usr/lib/ruby/gems/1.8/gems/tioga-1.4/
> Tioga_README  lgpl.txt  lib  split  tests
> 
> edwin at zaphod:~/tmp/samples$ /usr/bin/ls
> /usr/lib/ruby/gems/1.8/gems/tioga-1.4/lib
> Arcs_and_Circles.rb        FigureMaker.so        Strokes.rb
> ColorConstants.rb          Figures_and_Plots.rb  TeX_Text.rb
> Colorbars.rb               Flate.so              TexPreamble.rb
> Colormaps.rb               Function.so           Titles_and_Labels.rb
> Coordinate_Conversions.rb  Function_extras.rb    Transparency.rb
> Creating_Paths.rb          Images.rb             Using_Paths.rb
> Doc.rb                     Legends.rb            Utils.rb
> Dtable.so                  MarkerConstants.rb    X_and_Y_Axes.rb
> Dtable_extras.rb           Markers.rb            irb_tioga.rb
> Dvector.so                 Numeric_extras.rb     tioga.rb
> Dvector_extras.rb          Page_Frame_Bounds.rb  tioga_ui.rb
> Executive.rb               Rectangles.rb         tioga_ui_cmds.rb
> FigMkr.rb                  Shading.rb
> FigureConstants.rb         Special_Paths.rb

  That is really weird, see the structure created on my machine:

lib
|-- Dobjects
|   |-- Dtable.so
|   |-- Dtable_extras.rb
|   |-- Dvector.so
|   |-- Dvector_extras.rb
|   |-- Function.so
|   |-- Function_extras.rb
|   `-- Numeric_extras.rb
|-- Flate.so
`-- Tioga
    |-- Arcs_and_Circles.rb
    |-- ColorConstants.rb
    |-- Colorbars.rb
    |-- Colormaps.rb
    |-- Coordinate_Conversions.rb
    |-- Creating_Paths.rb
    |-- Doc.rb
    |-- Executive.rb
    |-- FigMkr.rb
    |-- FigureConstants.rb
    |-- FigureMaker.so
    |-- Figures_and_Plots.rb
    |-- Images.rb
    |-- Legends.rb
    |-- MarkerConstants.rb
    |-- Markers.rb
    |-- Page_Frame_Bounds.rb
    |-- Rectangles.rb
    |-- Shading.rb
    |-- Special_Paths.rb
    |-- Strokes.rb
    |-- TeX_Text.rb
    |-- TexPreamble.rb
    |-- Titles_and_Labels.rb
    |-- Transparency.rb
    |-- Using_Paths.rb
    |-- Utils.rb
    |-- X_and_Y_Axes.rb
    |-- irb_tioga.rb
    |-- tioga.rb
    |-- tioga_ui.rb
    `-- tioga_ui_cmds.rb

  Could you post the complete log of the installation ?
gem install tioga > log.txt 2>&1

  You might need to uninstall tioga first.

> I'm pretty sure that's all in order, especially since it works as soon as
> I move everything to Tioga:

  Even this is weird: it really shouldn't work this way, as Tioga
expects some libraries in Dobject/, which are not there. I'll be able to
deal more accurately with it as soon as I get the installation log.

  Cheers,

	Vincent

-- 
Vincent Fourmond, PhD student (not for long anymore)
http://vincent.fourmond.neuf.fr/

From andreas_s at hotmail.com  Wed Feb 14 20:35:56 2007
From: andreas_s at hotmail.com (Andreas S)
Date: Thu, 15 Feb 2007 01:35:56 +0000
Subject: [Tioga-users] tioga 1.4 segmentation fault
Message-ID: <BAY120-F239625DDBC3585872E61508D960@phx.gbl>

Following the tioga 1.4 thread on the similar issue, I thought another data 
point might be helpful since the last mail in the thread doesn't indicate 
any insight on the source of the problem

I build Tioga from source file and I've got this segmentation fault while 
running tc_Dtable.rb test

Loaded suite tc_Dtable
Started
....tc_Dtable.rb:167: [BUG] Segmentation fault
ruby 1.8.5 (2006-12-25) [i686-linux]

Abort (core dumped)

Red Hat Linux 3.2.3-54
I tried gcc 3.2.3 and gcc 3.3.2
ruby 1.8.5 (2006-12-25 patchlevel 12) [i686-linux]


(gdb) run tc_Dtable.rb
Starting program: /login/andresi/bin/ruby tc_Dtable.rb
Loaded suite tc_Dtable
Started
....
Program received signal SIGSEGV, Segmentation fault.
0x001419a0 in Dvector_Create ()
   from /login/andresi/lib/ruby/Dobjects/Dtable.so


I saw no error nor warning from make. This could be totally unrelated to 
Edwin's case.

-andre

_________________________________________________________________
Don?t miss your chance to WIN 10 hours of private jet travel from Microsoft 
Office Live http://clk.atdmt.com/MRT/go/mcrssaub0540002499mrt/direct/01/


From edder at tkwsping.nl  Thu Feb 15 11:33:38 2007
From: edder at tkwsping.nl (Edwin)
Date: Thu, 15 Feb 2007 16:33:38 -0000
Subject: [Tioga-users] tioga 1.4 segmentation fault
In-Reply-To: <BAY120-F239625DDBC3585872E61508D960@phx.gbl>
References: <BAY120-F239625DDBC3585872E61508D960@phx.gbl>
Message-ID: <op.tnstmckengac0g@localhost>

On Thu, 15 Feb 2007 01:35:56 -0000, Andreas S <andreas_s at hotmail.com>  
wrote:

> Following the tioga 1.4 thread on the similar issue, I thought another  
> data
> point might be helpful since the last mail in the thread doesn't indicate
> any insight on the source of the problem

Another data point for me, I got it working on another machine (a 64bit  
machine) without any problems, but my other two (32bit) are still not  
working. All three are using gentoo and I also tried it with another  
version of gcc (3.4.?).

Edwin

From vincent.fourmond at 9online.fr  Thu Feb 15 14:51:31 2007
From: vincent.fourmond at 9online.fr (Vincent Fourmond)
Date: Thu, 15 Feb 2007 20:51:31 +0100
Subject: [Tioga-users] tioga 1.4 segmentation fault
In-Reply-To: <BAY120-F239625DDBC3585872E61508D960@phx.gbl>
References: <BAY120-F239625DDBC3585872E61508D960@phx.gbl>
Message-ID: <45D4B9C3.5060202@9online.fr>

Andreas S wrote:
> Following the tioga 1.4 thread on the similar issue, I thought another
> data point might be helpful since the last mail in the thread doesn't
> indicate any insight on the source of the problem
> 
> I build Tioga from source file and I've got this segmentation fault
> while running tc_Dtable.rb test
> 
> Loaded suite tc_Dtable
> Started
> ....tc_Dtable.rb:167: [BUG] Segmentation fault
> ruby 1.8.5 (2006-12-25) [i686-linux]
> 
> Abort (core dumped)
> 
> Red Hat Linux 3.2.3-54
> I tried gcc 3.2.3 and gcc 3.3.2
> ruby 1.8.5 (2006-12-25 patchlevel 12) [i686-linux]
> 
> 
> (gdb) run tc_Dtable.rb
> Starting program: /login/andresi/bin/ruby tc_Dtable.rb
> Loaded suite tc_Dtable
> Started
> ....
> Program received signal SIGSEGV, Segmentation fault.
> 0x001419a0 in Dvector_Create ()
>   from /login/andresi/lib/ruby/Dobjects/Dtable.so

  Yes, I think I start to see the problem. This all has to do with
shared symbols... I thought I had designed a clever way to do
everything. Well not quite ;-) !

  I'll work on it and make a pre-release as soon as I can. It is
something I should have worked on a long time ago...

  Thanks for reporting,

	Vincent
-- 
Vincent Fourmond, PhD student (not for long anymore)
http://vincent.fourmond.neuf.fr/

From vincent.fourmond at 9online.fr  Thu Feb 15 15:08:29 2007
From: vincent.fourmond at 9online.fr (Vincent Fourmond)
Date: Thu, 15 Feb 2007 21:08:29 +0100
Subject: [Tioga-users] tioga 1.4 segmentation fault
In-Reply-To: <op.tnstmckengac0g@localhost>
References: <BAY120-F239625DDBC3585872E61508D960@phx.gbl>
	<op.tnstmckengac0g@localhost>
Message-ID: <45D4BDBD.2050803@9online.fr>

Edwin wrote:
> On Thu, 15 Feb 2007 01:35:56 -0000, Andreas S <andreas_s at hotmail.com>  
> wrote:
> 
>> Following the tioga 1.4 thread on the similar issue, I thought another  
>> data
>> point might be helpful since the last mail in the thread doesn't indicate
>> any insight on the source of the problem
> 
> Another data point for me, I got it working on another machine (a 64bit  
> machine) without any problems, but my other two (32bit) are still not  
> working. All three are using gentoo and I also tried it with another  
> version of gcc (3.4.?).

  Edwin, Andrea,

  Could you both try this:

  edit the file split/Dtable/dtable_intern.h to add the following lines

#undef PRIVATE
#define PRIVATE static

just before

#include "dvector.h"

and the sames lines just before

#include "dvector_intern.h"

in split/Dvector/dvector.c

  It might not stop segfaults, but they really should appear on another
test (not Dtable anymore).

  Cheers,

	Vincent

-- 
Vincent Fourmond, PhD student (not for long anymore)
http://vincent.fourmond.neuf.fr/

From edder at tkwsping.nl  Thu Feb 15 15:57:32 2007
From: edder at tkwsping.nl (Edwin)
Date: Thu, 15 Feb 2007 20:57:32 -0000
Subject: [Tioga-users] tioga 1.4 segmentation fault
In-Reply-To: <45D4BDBD.2050803@9online.fr>
References: <BAY120-F239625DDBC3585872E61508D960@phx.gbl>
	<op.tnstmckengac0g@localhost> <45D4BDBD.2050803@9online.fr>
Message-ID: <op.tns5t62bngac0g@localhost>

On Thu, 15 Feb 2007 20:08:29 -0000, Vincent Fourmond  
<vincent.fourmond at 9online.fr> wrote:

> Edwin wrote:
>> On Thu, 15 Feb 2007 01:35:56 -0000, Andreas S <andreas_s at hotmail.com>
>> wrote:
>>
>>> Following the tioga 1.4 thread on the similar issue, I thought another
>>> data
>>> point might be helpful since the last mail in the thread doesn't  
>>> indicate
>>> any insight on the source of the problem
>>
>> Another data point for me, I got it working on another machine (a 64bit
>> machine) without any problems, but my other two (32bit) are still not
>> working. All three are using gentoo and I also tried it with another
>> version of gcc (3.4.?).
>
>   Edwin, Andrea,
>
>   Could you both try this:
>
>   edit the file split/Dtable/dtable_intern.h to add the following lines
>
> #undef PRIVATE
> #define PRIVATE static
>
> just before
>
> #include "dvector.h"

#include <ruby.h>
#undef PRIVATE
#define PRIVATE static
#include "dvector.h"

>
> and the sames lines just before
>
> #include "dvector_intern.h"
>
> in split/Dvector/dvector.c

#include "intern.h"
#undef PRIVATE
#define PRIVATE static
#include "dvector_intern.h"

This causes the following warning at compilation:
In file included from Dvector/dvector_intern.h:25,
                  from Dvector/dvector.c:33:
./namespace.h:40:1: warning: "PRIVATE" redefined


>
>   It might not stop segfaults, but they really should appear on another
> test (not Dtable anymore).

The tests seem to run fine (not sure if they always did):
edwin at vimes:~/tmp/tioga-1.4/tests$ ruby ts_Tioga.rb
This is the test suite for Tioga 1.4
Loaded suite ts_Tioga
Started
...................................................................
compressed from 0 to 8
compressed from 13 to 21
compressed from 70 to 69
compressed from 3631 to 1642
.


>>> NOTE: please look at tests/Icon_Test.pdf and compare it to  
>>> samples/Icon.pdf
..........
Finished in 0.530279 seconds.

78 tests, 724 assertions, 0 failures, 0 errors


But still a segmentation fault while running tioga
(gdb) run /usr/bin/tioga sine.rb -0
Starting program: /usr/bin/ruby /usr/bin/tioga sine.rb -0
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
loading /home/edwin/tmp/samples/sine.rb

Program received signal SIGSEGV, Segmentation fault.
0x007d6630 in ?? ()

Edwin

From vincent.fourmond at 9online.fr  Thu Feb 15 18:13:50 2007
From: vincent.fourmond at 9online.fr (Vincent Fourmond)
Date: Fri, 16 Feb 2007 00:13:50 +0100
Subject: [Tioga-users] tioga 1.4 segmentation fault
In-Reply-To: <op.tns5t62bngac0g@localhost>
References: <BAY120-F239625DDBC3585872E61508D960@phx.gbl>
	<op.tnstmckengac0g@localhost> <45D4BDBD.2050803@9online.fr>
	<op.tns5t62bngac0g@localhost>
Message-ID: <45D4E92E.7060905@9online.fr>

Edwin wrote:
> #include "intern.h"
> #undef PRIVATE
> #define PRIVATE static
> #include "dvector_intern.h"
> 
> This causes the following warning at compilation:
> In file included from Dvector/dvector_intern.h:25,
>                   from Dvector/dvector.c:33:
> ./namespace.h:40:1: warning: "PRIVATE" redefined

  Well, I knew it. I'll make it a little cleaner later on, if that
happens to work.

>>   It might not stop segfaults, but they really should appear on another
>> test (not Dtable anymore).
> 
> The tests seem to run fine (not sure if they always did):
> edwin at vimes:~/tmp/tioga-1.4/tests$ ruby ts_Tioga.rb
> [...]
> 78 tests, 724 assertions, 0 failures, 0 errors

  OK. That should be good news, shouldn't it ?

> But still a segmentation fault while running tioga
> (gdb) run /usr/bin/tioga sine.rb -0
> Starting program: /usr/bin/ruby /usr/bin/tioga sine.rb -0
> loading /home/edwin/tmp/samples/sine.rb

  Hum. Did you make sure you have no files from an old tioga
installation somewhere ? (I got caught by those many times...).

  Did you try running other files in samples ? Do they all show the same
behavior ? What about ctioga (if you're using it) ?

> Program received signal SIGSEGV, Segmentation fault.
> 0x007d6630 in ?? ()

  What does the bt command of gdb says ?

  Thanks for the informations. I'm off to bed, so don't expect anything
before a (decently long) while. Cheers !

	Vincent

-- 
Vincent Fourmond, PhD student (not for long anymore)
http://vincent.fourmond.neuf.fr/

From vincent.fourmond at 9online.fr  Thu Feb 22 17:04:05 2007
From: vincent.fourmond at 9online.fr (Vincent Fourmond)
Date: Thu, 22 Feb 2007 23:04:05 +0100
Subject: [Tioga-users] tioga 1.4 segmentation fault
In-Reply-To: <BAY120-F17EBFA705438F308F8DD478D8F0@phx.gbl>
References: <BAY120-F17EBFA705438F308F8DD478D8F0@phx.gbl>
Message-ID: <45DE1355.4050703@9online.fr>

Andreas S wrote:
> I apologize for such slow response. I've tried adding
> #undef PRIVATE
> #define PRIVATE static
> into dvector.c and dtable_intern.h as you suggested. That actually fixed
> the segmentation fault in tc_Dtable.rb test.

  Great !

  OK... That is good news (for you) and bad news (for me). Basically,
the scheme which I had designed for sharing C symbols across ruby
extensions has some flaws in it. I'll have to redesign it somehow... But
I really can't promise it soon, as it will require more skill than I
have accumulated so far ;-)... Sorry. But be sure, I'm already thinking
on it !

  Cheers,

	Vincent

-- 
Vincent Fourmond, PhD student (not for long anymore)
http://vincent.fourmond.neuf.fr/

From edder at tkwsping.nl  Sat Feb 24 10:25:44 2007
From: edder at tkwsping.nl (Edwin)
Date: Sat, 24 Feb 2007 15:25:44 -0000
Subject: [Tioga-users] tioga 1.4 segmentation fault
In-Reply-To: <45D4E92E.7060905@9online.fr>
References: <BAY120-F239625DDBC3585872E61508D960@phx.gbl>
	<op.tnstmckengac0g@localhost> <45D4BDBD.2050803@9online.fr>
	<op.tns5t62bngac0g@localhost> <45D4E92E.7060905@9online.fr>
Message-ID: <op.tn9eg6bgngac0g@localhost>

Sorry for reacting so slow

>>>   It might not stop segfaults, but they really should appear on another
>>> test (not Dtable anymore).
>>
>> The tests seem to run fine (not sure if they always did):
>> edwin at vimes:~/tmp/tioga-1.4/tests$ ruby ts_Tioga.rb
>> [...]
>> 78 tests, 724 assertions, 0 failures, 0 errors
>
>   OK. That should be good news, shouldn't it ?

Ofcourse, but I didn't test this before, so not a hundred percent sure  
that this was a problem before.

>> But still a segmentation fault while running tioga
>> (gdb) run /usr/bin/tioga sine.rb -0
>> Starting program: /usr/bin/ruby /usr/bin/tioga sine.rb -0
>> loading /home/edwin/tmp/samples/sine.rb
>
>   Hum. Did you make sure you have no files from an old tioga
> installation somewhere ? (I got caught by those many times...).

I uninstalled the gem and then checked by hadn if it got everything. I am  
pretty confident I got everything (e.g. using locate ioga).

>   Did you try running other files in samples ? Do they all show the same
> behavior ? What about ctioga (if you're using it) ?

I'm not using ctioga, I tried to install it by hand, but that also lead to  
problems (see bottom of the email), then I installed the gem without the  
tioga dependency (because I needed it to use the tioga version installed  
by hand), but then ctioga can't find tioga while running it.

Other files (such as samples/plots.rb) cause the same segmentation fault

>> Program received signal SIGSEGV, Segmentation fault.
>> 0x007d6630 in ?? ()
>
>   What does the bt command of gdb says ?

(gdb) run /usr/bin/tioga plots.rb -1
Starting program: /usr/bin/ruby /usr/bin/tioga plots.rb -1
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
loading /home/edwin/tmp/tioga-1.4/samples/plots/plots.rb

Program received signal SIGSEGV, Segmentation fault.
0x007d6630 in ?? ()
(gdb) bt
#0  0x007d6630 in ?? ()
#1  0xbffb2c34 in ?? ()
#2  0x00000007 in ?? ()
#3  0x00000000 in ?? ()


Problems installing Ctioga:

  make install
/usr/bin/install -c -m 644 ./lib/movingarrays.rb  
/usr/lib/ruby/site_ruby/1.8/CTioga/movingarrays.rb
/usr/bin/install -c -m 644 ./lib/backends.rb  
/usr/lib/ruby/site_ruby/1.8/CTioga/backends.rb
/usr/bin/install -c -m 644 ./lib/debug.rb  
/usr/lib/ruby/site_ruby/1.8/CTioga/debug.rb
/usr/bin/install -c -m 644 ./lib/styles.rb  
/usr/lib/ruby/site_ruby/1.8/CTioga/styles.rb
/usr/bin/install -c -m 644 ./lib/curves.rb  
/usr/lib/ruby/site_ruby/1.8/CTioga/curves.rb
/usr/bin/install -c -m 644 ./lib/elements.rb  
/usr/lib/ruby/site_ruby/1.8/CTioga/elements.rb
/usr/bin/install -c -m 644 ./lib/tioga_primitives.rb  
/usr/lib/ruby/site_ruby/1.8/CTioga/tioga_primitives.rb
/usr/bin/install -c -m 644 ./lib/plotmaker.rb  
/usr/lib/ruby/site_ruby/1.8/CTioga/plotmaker.rb
ruby18 -r fileutils -e 'FileUtils.install(ARGV[0],ARGV[1],:mode => 0755)'  
bin/ctioga /usr/bin/ctioga
ruby18 -r fileutils -e 'FileUtils.install(ARGV[0],ARGV[1],:mode => 0755)'  
bin/ctable /usr/bin/ctable
make: *** No rule to make target `backends.rb', needed by  
`install_lib_sciyag'.  Stop.

ctioga
/usr/lib/ruby/site_ruby/1.8/CTioga/tioga_primitives.rb:17:in `require': no  
such file to load -- MetaBuilder/parameters (LoadError)
         from /usr/lib/ruby/site_ruby/1.8/CTioga/tioga_primitives.rb:17
         from /usr/lib/ruby/site_ruby/1.8/CTioga/plotmaker.rb:28:in  
`require'
         from /usr/lib/ruby/site_ruby/1.8/CTioga/plotmaker.rb:28
         from /usr/bin/ctioga:22:in `require'
         from /usr/bin/ctioga:22

I then tried to install MetaBuilder, but there was no mkmf.rb file =>
ruby extconf.rb
extconf.rb:1:in `require': no such file to load -- mkmf2 (LoadError)
         from extconf.rb:1
make install
make: *** No rule to make target `install'.  Stop


Thanks for all the help.

Edwin