Re: [igraph] R igraph 0.6 crashing on assortativity.nominal

[Gábor Csárdi]

Marcel,

'types' should have positive integer values. The manual was wrong, not
from zero, but from one.

I have corrected the manual and added a check for this.

Thanks for the report, best,
Gabor

On Mon, Mar 28, 2011 at 4:55 PM, Marcel Salathé
<address@hidden> wrote:
> Hi
>
> I run into problems with assortativity calculations. When using 
> assortativity.nominal on a directed graph, it crashes (segfault).
>
> Here's an example code that reproduces the behavior on my system (R version 
> 2.12.2 (2011-02-25), Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit), 
> igraph 0.6 downloaded today):
>
> library(igraph)
> toy = graph.formula(1-+2, 1-+3, 2-+1, 2-+3, 2-+4, 3-+2, 4-+5, 5-+2)
> V(toy)$prop = c(-1,-1,0,1,1)
> assortativity.nominal(toy,types=V(toy)$prop)
>
> It seems to calculate and return a value but then crashes immediately. 
> Because of the segfault, I'm not sure I trust the result though....
>
> Here's the output:
>
>> assortativity.nominal(toy,types=V(toy)$prop)
> [1] 0.06666667
>
>  *** caught segfault ***
> address 0x0, cause 'unknown'
>
> Possible actions:
> 1: abort (with core dump, if enabled)
> 2: normal R exit
> 3: exit R without saving workspace
> 4: exit R saving workspace
> [1] 0.06666667
>
>  *** caught segfault ***
> address 0x0, cause 'unknown'
>
> Possible actions:
> 1: abort (with core dump, if enabled)
> 2: normal R exit
> 3: exit R without saving workspace
> 4: exit R saving workspace
>
>  *** caught segfault ***
> address 0x0, cause 'unknown'
>
> Possible actions:
> 1: abort (with core dump, if enabled)
> 2: normal R exit
> 3: exit R without saving workspace
> 4: exit R saving workspace
> Selection:
> Selection:
>
>
>
> Thanks
> Marcel
>
>
>
> On Mar 28, 2011, at 3:01 PM, Gábor Csárdi wrote:
>
>> This is not very surprising, if you compile it for x86_64, then it
>> works only on x86_64. Compile it for i386 as well, and place it in
>> another R package directory. Another option is to create a fat binary,
>> that contains both the x86_64 and the i386 versions. I don't know how
>> to do that, but most probably it is covered in the R OS X guide.
>>
>> Best,
>> Gabor
>>
>> On Mon, Mar 28, 2011 at 1:13 PM, Marcel Salathé
>> <address@hidden> wrote:
>>> Compiling works now. However, library can only be loaded in 64-bit version 
>>> of R (Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)).
>>>
>>> On regular 32 bit version (Platform: i386-apple-darwin9.8.0/i386 (32-bit)), 
>>> I get
>>>
>>> Error: package 'igraph' is not installed for 'arch=i386'
>>>
>>> Thanks,
>>> Marcel
>>>
>>> PS: I'm using R 2.12.2
>>>
>>>
>>>
>>>
>>> On Mar 28, 2011, at 12:02 PM, Gábor Csárdi wrote:
>>>
>>>> Oh, this is the R package, sorry, I just noticed the 'R' in the
>>>> subject. This is a bug in the nightly bug service, the package
>>>> contains some extra files that should not be there. I have uploaded a
>>>> new nightly build that should fix this problem. Let me know if it does
>>>> not.
>>>>
>>>> Best,
>>>> Gabor
>>>>
>>>> On Mon, Mar 28, 2011 at 10:22 AM, Marcel Salathé
>>>> <address@hidden> wrote:
>>>>> Hi Gabor
>>>>>
>>>>> I've just ran into the exact same problem as Przemyslaw.
>>>>>
>>>>> I'm on Mac OS X 10.6.6, and I've been trying to install 
>>>>> igraph_nightly_0.6-2410-20110324.tar.gz which I downloaded an hour ago 
>>>>> from http://code.google.com/p/igraph/downloads/list
>>>>>
>>>>> The error I get is
>>>>> dnaupd.c:450: error: conflicting types for ‘igraphdnaupd_’
>>>>> igraph_arpack_internal.h:147: error: previous declaration of 
>>>>> ‘igraphdnaupd_’ was here
>>>>> make: *** [dnaupd.o] Error 1
>>>>> ERROR: compilation failed for package ‘igraph’
>>>>>
>>>>>
>>>>> Best,
>>>>> Marcel
>>>>>
>>>>>
>>>>> On Mar 25, 2011, at 7:21 PM, Gábor Csárdi wrote:
>>>>>
>>>>>> Przemyslaw,
>>>>>>
>>>>>> then please tell us about your platform and configuration because it
>>>>>> works fine for me.
>>>>>>
>>>>>> Best,
>>>>>> Gabor
>>>>>>
>>>>>> On Fri, Mar 25, 2011 at 7:02 PM, Przemek Grabowicz
>>>>>> <address@hidden> wrote:
>>>>>>>  Hi Gabor,
>>>>>>>
>>>>>>> In the newest development version from code.google.com this problem is 
>>>>>>> still
>>>>>>> present.
>>>>>>>
>>>>>>>
>>>>>>> Przemyslaw Grabowicz.
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> On 03/25/2011 01:45 PM, Gábor Csárdi wrote:
>>>>>>>>
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> I can't recall the exact reason, but this has been fixed.
>>>>>>>>
>>>>>>>> Best,
>>>>>>>> Gabor
>>>>>>>>
>>>>>>>> On Fri, Mar 25, 2011 at 11:23 AM, Przemek Grabowicz
>>>>>>>> <address@hidden>  wrote:
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> igraph-help mailing list
>>>>>>> address@hidden
>>>>>>> http://lists.nongnu.org/mailman/listinfo/igraph-help
>>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>> Gabor Csardi <address@hidden>     MTA KFKI RMKI
>>>>>>
>>>>>> _______________________________________________
>>>>>> igraph-help mailing list
>>>>>> address@hidden
>>>>>> http://lists.nongnu.org/mailman/listinfo/igraph-help
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> igraph-help mailing list
>>>>> address@hidden
>>>>> http://lists.nongnu.org/mailman/listinfo/igraph-help
>>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> Gabor Csardi <address@hidden>     MTA KFKI RMKI
>>>
>>>
>>> _______________________________________________
>>> igraph-help mailing list
>>> address@hidden
>>> http://lists.nongnu.org/mailman/listinfo/igraph-help
>>>
>>
>>
>>
>> --
>> Gabor Csardi <address@hidden>     MTA KFKI RMKI
>
>
> _______________________________________________
> igraph-help mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/igraph-help
>



-- 
Gabor Csardi <address@hidden>     MTA KFKI RMKI