> Sure, but my reasonable assumption was that when you insert an exception
> message in a doctest you reread it another one time and fix typos if any.
Relying on ideal behavior is rarely a "reasonable assumption" ;-)
What actually happens is that people write new doctests, run them, and in
the places an expected exception occurs they just copy+paste the traceback
from their terminal screen into the doctest. The only brain cells that
engage there are "expected an exception, got an exception -- great, let's
paste it in and move on". You're in "finish the doctest" mode then, not
"review messages for spelling and grammar" mode.
> I'll be more careful in the future about such assumptions.
There's a very simple rule, avoiding all need for philosophical rumination,
let alone tedious self-improvement ;-):
Run the tests before doing a checkin.
I was sorely tempted to skip that for a trivial checkin yesterday, but
forced myself, and was amazed to see dozens of tests fail. I apparently let
a finger slip right after opening the file, and managed to delete the first
quote character in the module's triple-quoted docstring. It certainly cost
me more time to run the tests, but that saved countless others the time to
deal with the consequences of my typo. "Don't think, just do it." <0.5
For more information about ZODB, see the ZODB Wiki:
ZODB-Dev mailing list - ZODB-Dev@zope.org