You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, nosetests --pdb-failure or nose2 -D are not optimal for generative tests, as the backtrace goes into the part of nose's code that runs the generator, instead of going into the generator itself. I'd guess it should be possible (if somewhat hackish) to retrieve the traceback of the generator instead (gen.gi_frame) and stitch it into the current traceback, so that the generator's frame becomes accessible in pdb via up.
Sorry I'm quite late on this. You're making a valid point. There are a few issues I'd like to address first though, and some of them might require some siginificant changes in the generative tests plugin (like #80), so I'll keep that in mind when touching it.
Currently,
nosetests --pdb-failure
ornose2 -D
are not optimal for generative tests, as the backtrace goes into the part of nose's code that runs the generator, instead of going into the generator itself. I'd guess it should be possible (if somewhat hackish) to retrieve the traceback of the generator instead (gen.gi_frame
) and stitch it into the current traceback, so that the generator's frame becomes accessible in pdb viaup
.Thoughts?
(crossposted as pytest-dev/pytest#1324)
The text was updated successfully, but these errors were encountered: