From: Karthik Nayak Date: Fri, 18 Sep 2015 15:10:35 GMT Subject: Re: [PATCH v17 14/14] tag.c: implement '--merged' and '--no-merged' options Message-ID: In-Reply-To: <20150918084208.GJ17201@serenity.lan> On Fri, Sep 18, 2015 at 2:12 PM, John Keeping wrote: > On Fri, Sep 18, 2015 at 09:10:08AM +0200, Matthieu Moy wrote: >> Junio C Hamano writes: >> >> > John Keeping writes: >> > >> >>> +--[no-]merged []:: >> >> >> >> We prefer to write --[no-]* as: >> >> >> >> --option:: >> >> --no-option:: >> >> >> >> although this may be the first instance that we see this combination >> >> with an argument. >> >> >> >> I also found the "[]" syntax confusing and had to go and figure >> >> out what PARSE_OPT_LASTARG_DEFAULT does; I wonder if it's worth >> >> appending something like the following to the help for this option: >> >> >> >> The `commit` may be omitted if this is the final argument. >> > >> > "may be omitted" must be followed by a description of what happens >> > when omitted (i.e. "defaults to ..."). >> >> Then: >> >> The `commit` may be omitted and defaults to HEAD if this is the final >> argument. > > I find that slightly confusing, although that might just be me. It's > slightly longer, but I would write: > > The `commit` may be omitted if this is the final argument, in > which case it defaults to `HEAD`. > This seems good to be included. > I also had a look at git-branch(1), which has similar `--merged` and > `--no-merged` options and says: > > Only list branches whose tips are reachable from the specified > commit (HEAD if not specified). Implies `--list`. > > The two options are listed separately in that case. Not sure this is much of a problem with regards to "--[no-]merged" I mean isn't the square brackets self-explanatory? -- Regards, Karthik Nayak