Importing Data Set Collection - Segmentation fault
Running graph-tool "2.23dev (commit b626e713, Tue Mar 28 22:29:08 2017 +0100)", compiled using GCC 5.4.0 on Ubuntu 16.04 I get a "Segmentation fault (core dumped)" error when running the following: import graph_tool.all as gt g = gt.collection.data["dolphins"] This issue does not seem to arise for me when running version "2.22 (commit 44bf2b92, Thu Mar 2 23:08:39 2017 +0000)" (i.e. the latest apt-get release). Is this an issue caused in the dev version of graph-tool or is this caused by me compiling something incorrectly? Can anybody reproduce this error? Best, Philipp -- View this message in context: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/... Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com.
On 03.04.2017 19:48, P-M wrote:
Running graph-tool "2.23dev (commit b626e713, Tue Mar 28 22:29:08 2017 +0100)", compiled using GCC 5.4.0 on Ubuntu 16.04 I get a "Segmentation fault (core dumped)" error when running the following:
import graph_tool.all as gt g = gt.collection.data["dolphins"]
This issue does not seem to arise for me when running version "2.22 (commit 44bf2b92, Thu Mar 2 23:08:39 2017 +0000)" (i.e. the latest apt-get release).
Is this an issue caused in the dev version of graph-tool or is this caused by me compiling something incorrectly? Can anybody reproduce this error?
No, I can't reproduce this. The git repository has automatic build tests that would easily catch a bug like this, and I've never seen it happening. I'm skeptical this is an actual bug; it is more likely to be a link-time inconsistency of some sort. Do you have a backtrace of the segfault? Best, Tiago -- Tiago de Paula Peixoto <tiago@skewed.de>
Unfortunately there is no backtrace whatsoever. All I get is Python terminating (if I run it interactively) and the message I posted being displayed. Nothing else appears on my screen. I am able to use it on my other computer using version "2.23dev (commit 316f8a3d, Tue Mar 28 14:07:46 2017 +0100)" so this issue is by no means critical to me if you have other things to get on with. Best, Philipp -- View this message in context: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/... Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com.
On 04.04.2017 14:52, P-M wrote:
Unfortunately there is no backtrace whatsoever. All I get is Python terminating (if I run it interactively) and the message I posted being displayed. Nothing else appears on my screen.
You have to get a backtrace using GDB; it does not happen automatically. You can google "gdb backtrace" to see how it is done.
I am able to use it on my other computer using version "2.23dev (commit 316f8a3d, Tue Mar 28 14:07:46 2017 +0100)" so this issue is by no means critical to me if you have other things to get on with.
I most certainly have other things to do; but it is up to you if you want to provide the backtrace to investigate further. -- Tiago de Paula Peixoto <tiago@skewed.de>
Hi Tiago, The output of gdb is attached ( gdb-python.txt <http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/file/n4027181/gdb-python.txt> ). The issue seems to be: Thread 1 "python" received signal SIGSEGV, Segmentation fault. 0x00007fffe4248960 in boost::iostreams::detail::gzip_header::reset() () from /home/pmj27/anaconda2/lib/libboost_iostreams.so.1.60.0 Is this of any use? Best, Philipp -- View this message in context: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/... Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com.
Sorry, here is the full output: gdb-python.txt <http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/file/n4027182/gdb-python.txt> Philipp -- View this message in context: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/... Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com.
On 04.04.2017 16:12, P-M wrote:
Sorry, here is the full output: gdb-python.txt <http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/file/n4027182/gdb-python.txt>
It seems to be coming from boost-iostreams; not graph-tool. I see you are using boost 1.60. Maybe you should try upgrading to 1.61, 1.62 or 1.63. Best, Tiago -- Tiago de Paula Peixoto <tiago@skewed.de>
It looks like it (though it is funny that the other computer using the same package version doesn't have the issue). It has been a major battle to get all the dependencies to work in Anaconda python so seeing as I have a configuration that currently works I am inclined to stick with the versions I have for now unless other issues also start cropping up. Being able to compile new versions of graph-tool easily is more important to me I think. Thank you for looking into it though! Best, Philipp -- View this message in context: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/... Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com.
participants (2)
-
P-M -
Tiago de Paula Peixoto