Re: [igraph] How to build development version?

[Jeff Alstott]

Well, I hacked cairocffi into the 0.7 branch of python-graph. But now it turns out I would really rather use the development version, due to this (now fixed) problem plotting in Ipython Notebooks on Python 3:

http://stackoverflow.com/questions/26887513/igraph-plot-function-does-not-show-images-in-an-ipython-notebook

But I can't actually make the development igraph still. I built flex from source, then:

./configure --prefix=${HOME}

make

make check

This is what I get at the end of make check:

make[2]: Leaving directory `/home/jeffrey_alstott/igraph/src'

make[1]: Leaving directory `/home/jeffrey_alstott/igraph/src'

Making check in tests

make[1]: Entering directory `/home/jeffrey_alstott/igraph/tests'

make  check-local

make[2]: Entering directory `/home/jeffrey_alstott/igraph/tests'

{                                       \

          echo '# Signature of the current package.'; \

          echo 'm4_define([AT_PACKAGE_NAME],      [igraph])'; \

          echo 'm4_define([AT_PACKAGE_TARNAME],   [igraph])'; \

          echo 'm4_define([AT_PACKAGE_VERSION],   [0.8.0-pre+594.701edd8])'; \

          echo 'm4_define([AT_PACKAGE_STRING],    [igraph 0.8.0-pre+594.701edd8])'; \

          echo 'm4_define([AT_PACKAGE_BUGREPORT], [address@hidden])'; \

        } >./package.m4

/bin/sh /home/jeffrey_alstott/igraph/missing --run autom4te --language=autotest -I ../tests ../tests/testsuite.at -o testsuite.tmp

/usr/bin/m4:community.at:58: ERROR: end of file in argument list

autom4te: /usr/bin/m4 failed with exit status: 1

make[2]: *** [testsuite] Error 1

make[2]: Leaving directory `/home/jeffrey_alstott/igraph/tests'

make[1]: *** [check-am] Error 2

make[1]: Leaving directory `/home/jeffrey_alstott/igraph/tests'

make: *** [check-recursive] Error 1

On Wed, Jan 28, 2015 at 10:29 AM, Jeff Alstott <address@hidden> wrote:

Thanks for the info to you both!

I now need to decide if I should set up the development branch on my machine, or hack my python-igraph to allow for cairocffi. Which is likely a better move? Possible considerations:

- How stable is the development branch right now?

- Is there any estimate of when the next version will be released? (with cairocffi presumably baked in)

- Is either move likely to cause me any other problems down the line?

On Tue, Jan 27, 2015 at 3:35 PM, Jeff Alstott <address@hidden> wrote:

I am trying to build and install the development version of igraph and python-igraph, because I want to use cairocffi instead of pycairo (which has been implemented in the development, but not version 0.7). However, I'm having trouble configuring igraph:

$ls

AUTHORS       configure.ac     debian    igraph_Info.plist.in  INSTALL.WINDOWS  msvc   optional  tests

bootstrap.sh  CONTRIBUTING.md  doc       igraph.pc.in          interfaces       NEWS   README    TODO

ChangeLog     COPYING          examples  include               Makefile.am      nexus  src       tools

$./configure

bash: ./configure: No such file or directory

$chmod +x configure.ac

$./configure.ac

./configure.ac: line 1: syntax error near unexpected token `igraph,'

./configure.ac: line 1: `AC_INIT(igraph, esyscmd([tr -d '\n' < VERSION]), address@hidden)'

What am I missing?