I like this idea. And this kind of "breaking" wouldn't be that bad: instead of their files being changed by the script, users will get a diff on their screen and should figure out easily what's wrong.
And this mode will be consistent with clang-format's command line interface, which is really useful. On Fri, Oct 4, 2013 at 10:28 PM, Alp Toker <[email protected]> wrote: > Thinking outside the box, it almost makes sense to have Alexander's mode > as the default for this tool because it already works in terms of > patches, then have -i to do an inplace update if requested. > > I like the semantics of diff-in-diff-out and it avoids the question of > what to call the new flag, but it would be a breaking change. Of course > since there are no tests, it wouldn't technically be breaking anything ;-) > > Alp. > > On 04/10/2013 21:02, Daniel Jasper wrote: > > > > ================ > > Comment at: tools/clang-format/clang-format-diff.py:40 > > @@ -36,1 +39,3 @@ > > 'Reformat changed lines in diff.') > > + parser.add_argument('-n', action='store_true', default=False, > > + help='don\'t change files, instead display what > diffs' > > ---------------- > > Why "-n" is there precedence in other tools? Otherwise, how about a > named parameter? Maybe --dry_run? > > > > Or maybe two parameters "--dry_run" (do not apply changes) and "-v" > (display diffs)? > > > > > > http://llvm-reviews.chandlerc.com/D1840 > > _______________________________________________ > > cfe-commits mailing list > > [email protected] > > http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits >
_______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
