Andreas,
Nice job hunting this down. I'd been seeing it as well upstream (http://trac.mapnik.org/ticket/392
) but had not taken the time to debug.
I agree that #2 is critical - that we avoid crashes if pycairo python
module is unavailable at import time even if mapnik was compiled with
pyc
It looks like this change against Mapnik trunk solves the issue:
Index: bindings/python/python_cairo.cpp
===
--- bindings/python/python_cairo.cpp(revision 1277)
+++ bindings/python/python_cairo.cpp(working copy)
@@ -57,6 +57
It looks like this change against Mapnik trunk solves the issue:
Index: bindings/python/python_cairo.cpp
===
--- bindings/python/python_cairo.cpp(revision 1277)
+++ bindings/python/python_cairo.cpp(working copy)
@@ -57,6 +57
Andreas,
Yes, I can see the value behind issuing an error but the user would
also run into an import error when they tried to do anything further
requiring:
>>> import cairo
which would result in an ImportError exception. So I figured Mapnik
just shouldn't crash and pycairo availability
This is due to boost filesystem changing, and can be fixed with this patch:
http://trac.mapnik.org/changeset/2506/branches/0.7.2-dev
Dane
On May 2, 2011, at 5:35 AM, Lucas Nussbaum wrote:
> Source: mapnik
> Version: 0.7.1-3
> Severity: serious
> Tags: wheezy sid
> User: debian...@lists.debian.
On Oct 1, 2010, at 12:52 PM, Peter Schaefer wrote:
> Weitergeleitete Nachricht
>> Von: Peter Schaefer
>> Reply-to: peter.schae...@kettwieseltandem.de
>> An: Julien Cristau
>> Kopie: 598...@bugs.debian.org, da...@debian.org
>> Betreff: Re: Bug#598652: python-mapnik: Misssing sh
6 matches
Mail list logo
