On Feb 10, 2008, at 9:13 PM, William Stein wrote:

> Any line with "random" anywhere in it is replaced by
>
>    sage: _ = [the original line]
> <newline>
>    [original output]      <--- gets ignored because of the newline
>
> This is so doctests with random output can still be run using exactly
> the same doctesting framework, but the output is not tested.
>
> This is a hack, clearly.  It would be better to run the random input
> line *exactly* as is, but just somehow ignore the output.  When I  
> wrote

[...]

okay I see.

A much better solution would be for sage to have a sane pseudo-random  
number framework, so we could be guaranteed sane behaviour for  
doctests like this.

I realise how difficult this is, given how many systems there are in  
sage with different random number generators etc.

Actually this burns me quite often for various reasons. I frequently  
have some code that I'm playing with, and I'm debugging deep into  
some complicated piece of code, and I have a piece of paper with  
various numbers written on it keeping track of what's supposed to be  
happening.... and then I run it again and I get totally different  
numbers and I have to work it all out again. It's a real pain, and  
quite unworkable.

BTW this is a hack for two reasons. The "lesser hack" is that it  
rewrites code that is supposed to have random output. I suppose I can  
live with this. But the "greater hack", which is much more confusing  
to me, is that it gets applied for any line containing the string  
"random"!! That's totally unexpected and weird. i.e. I could live with:

sage: some_func(5)      # random
234736

where it's clearer that there is some decoration on the line.

What about

sage: rand_stuff()

or

sage: rnd_stuff()

?

david


--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---

Reply via email to