I should not that all single-case letters used to be defined, and I
for one was often bitten by this in strange ways.
On Fri, May 11, 2012 at 11:13 AM, Nils Bruin <nbr...@sfu.ca> wrote:
> On May 11, 12:37 am, Julien Puydt <julien.pu...@laposte.net> wrote:
>
>> NameError: name 'y' is not defined -- perhaps you need to type var('y')
>> first ?
>
> A nice idea, but I don't think NameError is sufficiently aware of the
> context it is raised in to make reliable suggestions:
>
> sage: sinus(pi/3)
> NameError: name 'sinus' is not defined -- perhaps you need to type
> var('sinus') first?
> sage: var('sinus')
> sage: sinus(pi/3)
> [Deprecation warning]
> 1/3*pi
> sage: sinus(pi/3)
> 1/3*pi
>
> Following the advice indeed makes the error go away, but likely
> doesn't produce the result the user intended.
When the declaration warning becomes an error, it will be a bit
better. Certainly an argument for not making all undefined symbols
into symbolic variables. One can write SR("a*x+b") and it defines the
variables for you (though not globally IIRC).
> Also, I'm afraid NameError gets raised somewhere in the innards of
> Python, so to implement this idea we probably would need to patch
> Python.
I bet we could catch this and re-throw a modified version; we have
hooks into the read-eval-print loop to do preparsing, for example.
- Robert
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
