The simplest way I know to debug the current Python script is:

1. Copy the name of the current file (C-x C-j C-0 w)
2. M-x pdb RET yank RET (possibly after deleting previous file name)

This seems like a good place to introduce a "future history" operation.
The attached patch does so, via a new optional argument to
gud-query-cmdline.  The above then simplifies to M-x pdb RET M-n RET.

Existing behavior is unchanged.  I added the same feature to other gdb
commands for which it seemed to make sense (perldb, guiler).  Feedback
or other suggestions would be welcome.