Ubuntu 13.10 cannot configure properly

graph-tool
1

Hi,everyone

I have some problems with installing graph-tools.
After I execure ./configure graph-tools, it stops with a message saying
libstdc++ config.status: executing libtool commands
configure: Modifying libtool to correctly link libstdc++ before anything
else...I am not fimiliar with Linux...so I don't know how I can I link my
libstd to libtool.

I install boost-1.55.0 ipython-dev, python with spyder, scipy, numpy,
expat, cgal and sparsehash. My OS is Ubuntu 13.10 64bit with gnome 3.10.

The Internet access is not good. So I cannot use the sources that you
provide.

thx for help,
Liu

2

Hi,everyone

I have some problems with installing graph-tools.
After I execure ./configure graph-tools, it stops with a message saying
libstdc++ config.status: executing libtool commands
configure: Modifying libtool to correctly link libstdc++ before anything
else...I am not fimiliar with Linux...so I don't know how I can I link my
libstd to libtool.

This is not an error message. After the configure script finishes, you
have to compile and install the package with "make" and "make install".

I install boost-1.55.0 ipython-dev, python with spyder, scipy, numpy,
expat, cgal and sparsehash. My OS is Ubuntu 13.10 64bit with gnome 3.10.

The Internet access is not good. So I cannot use the sources that you
provide.

Are you referring to the pre-compiled packages for ubuntu? Why can you
download the source but not the binary packages? It is much easier than
compiling it by hand.

Best,
Tiago

3

Hello, Mr.Tiago de Paula Peixoto,

I have installed python-graph-tool with apt-get command. But another
problem occurs. When I execute the example, Building a Price Network and
then draw the network with graph_draw() function, my python2.7.5 crashs
down. It reports that python has a segmentation fault. I have the core
dumped. I have installed all requirements to use graph_tool. Is there any
wrong setting I made? Or it is a bug?

best regards, Liu
在 2014年1月3日星期五UTC+8上午2时20分58秒，Tiago Peixoto写道：

attachment.html (2.05 KB)

4

I cannot reproduce this with Ubuntu 13.10. Could you provide a minimal
but complete example of when you observer the crash?

A new version of graph-tool (2.2.28) was released today. Do you observe
the crash with this version as well?

Best,
Tiago

5

Seems like I have a similiar Problem, using ubuntu 13.10 on a 64-Bit
machine.
The boost library + python seems to have a Problem, Everything works fine
until the first function uses the boos library... my guess.
Example:

# -*- coding: utf-8 -*-

import numpy as np
import graph_tool.all as gt

import graph_tool.topology as gt_topo

G = gt.Graph()

G_sub = gt.Graph()

#Decleration of needed properties
v_prop_entity = G.new_vertex_property("string")
v_prop_globalName = G.new_vertex_property('string')
e_prop_type = G.new_edge_property('string')

#gt_topo.subgraph_isomorphism()

v1 = G.add_vertex()
v2 = G.add_vertex()

v3 = G_sub.add_vertex()

e1 = G.add_edge(v1, v2)

#v_prop_entity[v1] = 'car'

#make internal property maps
G.vertex_properties['type'] = v_prop_entity
G.vertex_properties['globalName'] = v_prop_globalName
G.edge_properties['type'] = e_prop_type

G_sub.vertex_properties['type'] = v_prop_entity
G_sub.vertex_properties['globalName'] = v_prop_globalName
G_sub.edge_properties['type'] = e_prop_type

#write/change internal property maps
G.vertex_properties['type'][v1] = 'Car'
G.vertex_properties['type'][v2] = 'T-Crossing'

G_sub.vertex_properties['type'][v1] = 'Car'

#gt_topo.subgraph_isomorphism(G_sub, G, (G_sub.vertex_properties['type'], \
# G.vertex_properties['type']) )
print gt_topo.subgraph_isomorphism(G_sub, G)

following error:
xxx:~/Code$ python graph-tool_test2.py
Speicherzugriffsfehler (Speicherabzug geschrieben)

which means: memory <http://www.dict.cc/englisch-deutsch/memory.html&gt; access<http://www.dict.cc/englisch-deutsch/access.html&gt;
error:

The problem occurs, when using the subgraph_isomorphism function...

attachment.html (5.38 KB)

6

Hello, Mr.Tiago de Paula Peixoto

I am sorry for the late reply. I have not checked Google Forums for a
while. It seems that there is something wrong with pre-compiled packages
for Ubuntu 13.10 64Bit version. If I complie the source code and install
graph-tool, graph_draw() works well under graph-tool 2.27 and graph-tool
2.28 version. But after I install python-graph-tool with apt-get command,
graph_draw() fails drawing any graph which has more than 1000 vertexes.

Here is the test python code I used:

attachment.html (2.88 KB)

price.xml.gz (164 KB)

7

Hello, Mr.Klingelschmitt

I have just tested your example on my machine(Ubuntu 13.10 64Bit Gnome
3.10). I compiled the source code of graph-tool v.2.28 and then intalled
it. The same problem occurs when using subgraph_isomorphism function.
Python console reports "UMD has deleted: gi.repository.Gtk,
gi.repository.Gdk, gi.repository.GdkPixbuf, gi.repository.GLib,
gi.repository.GObject, gi.repository.cairo, gi.repository.Atk".

在 2014年2月14日星期五UTC+8上午3时13分21秒，Stefan Klingelschmitt写道：

attachment.html (6.9 KB)

8

Hello, Mr. Mr.Tiago de Paula Peixoto

I tested the example Mr.Klingelschmitt's example. And the same problem
occurs with Ubuntu internal error which says python2.7 crashed with SIGSEGV
in void get_mapping::operator() and python concole also reports "Python
console reports "UMD has deleted: gi.repository.Gtk, gi.repository.Gdk,
gi.repository.GdkPixbuf, gi.repository.GLib, gi.repository.GObject,
gi.repository.cairo, gi.repository.Atk".

I hope these may help.

Best,
Liu

在 2014年1月5日星期日UTC+8上午12时14分19秒，Tiago Peixoto写道：

attachment.html (2.04 KB)

9

What do you mean by failing to draw the graph?

I tried the script you sent in Ubuntu 13.10 with graph-tool 2.2.29.1 and
I found no problems. See output attached.

Best,
Tiago

price.jpg

10

Seems like I have a similiar Problem, using ubuntu 13.10 on a 64-Bit
machine. The boost library + python seems to have a Problem,
Everything works fine until the first function uses the boos
library... my guess.

<...>

The problem occurs, when using the subgraph_isomorphism function...

This seems indeed to be a bug in the subgraph_isomorphism function, but
it only occurs if the subgraph contains zero edges. I'll provide a fix
soon.

(Note that finding a subgraph with zero edges is a silly task, since the
result will be trivial.)

Best,
Tiago

11

>(Note that finding a subgraph with zero edges is a silly task, since the
>result will be trivial.)

This is true, I just started a very basic example to get used to the
interface.
Indeed, adding an edge solves the problem.

Thanks!

attachment.html (428 Bytes)