From: Felipe Contreras Date: Sun, 08 Apr 2012 14:58:08 GMT Subject: Re: [PATCH 02/12] completion: simplify __gitcomp Message-ID: In-Reply-To: <20120408143621.GA2631@burratino> On Sun, Apr 8, 2012 at 5:36 PM, Jonathan Nieder wrote: > (dropping Shawn from cc since I'm not under the impression he works >  on the completion script these days) > Felipe Contreras wrote: > >> Whoa! Breaking their tab completion? Where? Can you show me some evidence? > > If you weren't listening before, I'm not sure what I can add now[*]. This is what you said; "in exchange for us breaking their tab completion". There's a big difference between "breaking", and "*potentially* breaking". I have never seen any evidence of tab completion actually being broken. > Luckily, I already said what I needed to say.  Yes, cleanups can be > good when they don't break things, and no, cleanups that break things > are not good.  Sometimes it is not obvious which category each case > falls into.  As you well know, "tests pass" is not enough (e.g., > sometimes there are no tests!).  If you want someone to argue with, > you can find someone else. Fortunately there's no evidence of anything being broken; all the issues mentioned are *theoretical*, and there's not even an example of a custom completion that would be broken by this. -- Felipe Contreras