unknown terminal “cygwin”

[pblelloch]

I’m using Spyder (2.2) from an Anaconda distribution on Windows 7, and every time I print an object I get the following warning (usually twice):

 

unknown terminal “cygwin”

 

This doesn’t occur on my laptop, but that doesn’t have cygwin installed, while my desktop does.  It’s only a minor annoyance, but is there anyway to disable that message?

 

-Paul Blelloch

[Jed Ludlow]

Hi, Paul,

I suspect your issue is more related to Anaconda itself rather than Spyder. You might try asking it on their forum here:

https://groups.google.com/a/continuum.io/forum/#!forum/anaconda

Hope that helps,

Jed 

[pblelloch]

It took me a while to check this out, but the version of Spyder that comes with WinPython acts in exactly the same way as the Anaconda distribution, so it does appear to be Spyder and not Anaconda.  Whenever I print an object it gives me the "Unknown terminal "Cygwin"" error.  This only occurs in the default interpreter and not in an iPython interpreter.

 

 

In addition both numpy 1.7.1 and scipy 0.12.0 linalg modules cause Python to close when doing calculations on matrices larger than about 200x200.  That problem appears to be related to AVX calls in MKL, but I can't work it out further than that.  It occurs on all workstations in our company, but I've tried asking on the scipy-users list and haven't gotten any response, which probably means that it's limited to my particular class of desktop (though I can't believe that it's that rare).  These errors occur whether I download the numpy and scipy through Anaconda, WinPython or directly from Christoph Gohlke's site (I think that they all come from there anyway). 

 

The first issue is an annoyance, but the second is somewhat of a showstopper :-(.

 

 

-Paul

[pblelloch]

This is a minor update.  While the problems with Spyder persist, the more serious problem that I was having with numpy and scipy crashing was found and fixed by Christoph Gohlke, so if you update to his latest builds of 64-bit MKL numpy and scipy those issues go away.

On Friday, May 10, 2013 12:51:22 PM UTC-7, pblelloch wrote:

[Carlos Córdoba]

Hi Paul,

I'm afraid I don't have time to debug this, but a quick goggling seems to indicate that you could solve the problem by changing you TERM env variable, as it's mentioned for example here:

http://cygwin.com/ml/cygwin/2000-06/msg00412.html

HTH,
Carlos

PS. By the way, what was causing your problem with numpy and scipy?

El 08/06/13 22:49, pblelloch escribió:

--
You received this message because you are subscribed to the Google Groups "spyder" group.
To unsubscribe from this group and stop receiving emails from it, send an email to spyderlib+...@googlegroups.com.
To post to this group, send email to spyd...@googlegroups.com.
Visit this group at http://groups.google.com/group/spyderlib?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Paul Blelloch]

Christoph Gohlke surmised that the problem with numpy and scipy was related to <http://software.intel.com/en-us/articles/svd-multithreading-bug-in-mkl>.I did in fact have 6 cores on the machine that was giving me trouble.  He first rebuilt numpy with update 4 of Composer XE and MKL 11.0.  That fixed numpy, but scipy.linalg routines were still crashing.  He then rebuilt scipy the same way and that fixed that too. 

I'll try the fix that you suggest at work tomorrow.

THANKS

--
You received this message because you are subscribed to a topic in the Google Groups "spyder" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/spyderlib/rE4ygiXMEoU/unsubscribe?hl=en.
To unsubscribe from this group and all its topics, send an email to spyderlib+...@googlegroups.com.