[TIP] while on the subject of unittest...

Fernando Perez fperez.net at gmail.com
Sun Sep 27 12:23:00 PDT 2009

Hi Michael,

Thanks!  For some odd reason, this message only showed up on my inbox
today, though it appears as from '7 days ago'.  Weird...

On Mon, Sep 21, 2009 at 3:55 AM, Michael Foord
<fuzzyman at voidspace.org.uk> wrote:
> That sounds like a reasonable request - although I'm marginally
> unsympathetic to people who are subclassing TestProgram in the firstplace as
> it is an abomination that should never have existed. :-)
> At some point I'd like to deprecate TestProgram altogether and replace it
> with something less horrible.
> Anyway: http://bugs.python.org/issue6956

Thanks a lot!  I tried to file a bug report but the mail forwarding
through sourceforge was so slow that I only got the password reset
many days later, and I'd gotten busy with other things.  Much

I should say that I have no special attachment to TestProgram at all,
and in fact all I'd like to have is, *in the standard library*
something to be able to do what I can get with

- nosetests  foo
- nosetests --with-doctest foo
- nosetests --pdb/--pdb-failures foo

where foo has reasonably flexible semantics to work at the
package/module/class/function/single-method levels.

I use nose for that reason, because it lets me work quickly, though
nose is probably too much magic for the stdlib.  But anything you guys
are working on that can make this type of (IMHO critically necessary)
workflow possible within the stdlib, will only get loud cheers from me

Best regards,


More information about the testing-in-python mailing list