Junio C Hamano wrote: > Stephan Feder <sf@xxxxxxxx> writes: > >> I have to send patches of binary data to a customer but the builtin diff >> was no help in this case. > > Given the previous patch, and also your point #2 below, I would > have expected you to introduce an option to force files to be > treated as binary even when they are otherwise misidentified as > text, but this patch is going the other way. > > Interesting. Not really. I was surprised that the GNU diff option --text is unsupported in the builtin diff. > >> 1. The shorthand -a for --text is not implemented. Is there a conflicting >> shorthand? > > I do not think of one offhand, but it's the responsibility for > the party to propose such an enhancement to do the study ;-) Of course. I did not find any conflict but as the builtin diff and its options are used by quite a lot of git commands I wanted to make sure. If no objections arise I am going to add the shorthand. >> 2. For diffstat --text is ignored. It seems pointless because binary >> patch data is not for human consumption anyway. > >> 3. No documentation yet. If the patch is accepted I will add a short >> description. To Documentation/diff-options.txt? > > Most likely that would be the place. Good. - : send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html