[Numpy-discussion] String (and unicode) comparisons and per-thread error handling fixed

[Travis Oliphant]

Tim Hochberg wrote:
> Travis Oliphant wrote:
>
>>
>> String comparisons were added last week.  Today, I added per-thread 
>> error handling to NumPy.  There is 1 more enhancement (scalar math) 
>> prior to 0.9.8 release --- but it will probably take 1-2 weeks.
>
> Oops!  I'm about 2/3 done doing this one too. I think I'll go ahead 
> and finish mine up and see how our approaches stack up performance 
> wise and see if there's any of mine that's useful to roll into yours.
Darn.  I thought I gave you enough time.... :-)

On the other hand,  all I did was change the way the error-mode is being 
looked-up (from the three dictionaries to just one).  It's not much 
different than before except for that.    I didn't do anything about the 
other ideas you spoke of. 

I did add a simple object to reset the error mode when it gets deleted, 
and had to fiddle with the seterr code a little to accept that object so 
that both methods of resetting the error mode work. 

A stack can certainly be built on top of what is now there (I'm thinking 
for numarray compatibility...), but I didn't do that.

Sorry for stepping on your toes.   I'm just anxious...  I'll be gone for 
a couple of days and won't be working on NumPy/SciPy, so feel free to 
adjust.


-Travis