RetroSearch Browse

Mon Mar 7 05:32:24 CET 2011 · https://mail.python.org/pipermail/python-dev/2011-March/108870.html

Am 07.03.2011 03:43, schrieb Stephen J. Turnbull:
> "Martin v. Löwis" writes:
>   >  Am 07.03.2011 02:24, schrieb Stephen J. Turnbull:
>   >  >  "Martin v. Löwis" writes:
>   >  >    >   It seems that the dev guide recommends to use the --git option in hg
>   >  >    >   diff. I'm working on the Rietveld integration, and found that this
>   >  >    >   option makes things worse: the regular diff includes the base revision
>   >  >    >   of the patch; hg diff --git doesn't.
>   >  >
>   >  >  Does the regular diff work acceptably for the kinds of changes that
>   >  >  diff --git was designed to be an improvement for?
>   >
>   >  I don't know. What are the kinds of changes that diff --git was designed
>   >  for?
>
> I don't know exactly how much of git diffcore has been implemented in
> hg diff --git.  However, git's diff handles renames and copies
> correctly and pleasantly, including swapping file names (ie, renaming
> a to b and b to a simultaneously), and can change file modes.
>
> That kind of change is rather unpleasant to deal with in a traditional
> diff format.  Eg, renames are represented as deleting all the lines
> from one file and re-adding them as a new file.

Ok, so the next question is what constitutes an acceptable 
representation. I find the original approach to diff completely
acceptable, also considering that people rarely rename files,
and if they do, they typically don't put patches into a bug tracker.

Regards,
Martin

Home - News ( United States | United Kingdom | Italy | Germany ) - Football scores

Showing content from https://mail.python.org/pipermail/python-dev/2011-March/108870.html below:

[Python-Dev] hg diff