Re: Enumerated string comparisons

["Johnson, Andrew N." <anj@aps.anl.gov>]

It would be easy to use case insensitive matching instead of the strcmp() that we use at present, but I think we should check the string against all possibilities using a case sensitive comparison before using the more fuzzy matching, and doing that would make the code a little longer. The underline/space thing would make the comparison a bit more fuzzy than I'm really comfortable with myself. I will see what I can do to improve the error message, but IIRC the message is a bit removed from where the error is detected.

Thanks for all your comments,

- Andrew

-- 
Sent from my iPad

On Oct 9, 2013, at 4:09 PM, "Eric Norum" <eric@norum.ca> wrote:

> On Oct 9, 2013, at 3:58 PM, Benjamin Franksen <benjamin.franksen@helmholtz-berlin.de> wrote:
> 
>> 
>> 
>> Therefore I see a case for some very limited sort of dwimming, namely:
>> 
>> (1) Case-insensitive matching.
>> 
>> (2) Identifying underline '_' and space ' '.
> I'm more in favour of completely strict matching as is now the case (but with really descriptive diagnostic messages on failure).   My opinion is that 'dwimming' results in making things less understandable because it gets too hard to remember the rules and thus databases and such get harder to comprehend.
> 
> -- 
> Eric Norum
> eric@norum.ca
> 
> 
> 
> 
>