SZEDER Gábor <szeder....@gmail.com> writes: >> > inifinite recursion)? Or are they "correct but not immediately >> > necessary" (e.g. because calling read_cache() does not hurt until >> > that function gets removed, so rewriting the callers to call >> > read_index() with &the_index may be correct but not immediately >> > necessary)? >> >> the latter. I assume correctness (of the semantic patch) to be a given, > > I'm afraid we can't assume that. As far as correctness is concerned, > I think semantic patches are not different from any other code we > write, i.e. they are potentially buggy. Perhaps even more so than the > "usual" Git code, because we have long experience writing C and shell > code (with all their quirks and error-proneness), but none of us is > really an expert in writing semantic patches.
All correct. And applying semantic patches generated from buggy sources can produce buggy code, just like merging a buggy topic branch does. These days, I ran coccicheck at the tip of 'pu' (even though this cost me quite a lot a few times every day) and feed its findings back to authors of topics that introduce new ones, so that their topics next time do not need the fix-up at the tip of 'pu' in the next integration cycle. That way, the changes mechanically suggested by coccicheck can still be reviewed in small bite-sized chunks and hopefully possible problems caused by buggy sources can either be found in the review process, or discovered at the tip of 'pu'.
