On 01/03, Duy Nguyen wrote: > On Wed, Dec 14, 2016 at 6:14 AM, Brandon Williams <bmwill@xxxxxxxxxx> wrote: > > @@ -340,8 +336,9 @@ static void NORETURN unsupported_magic(const char *pattern, > > continue; > > if (sb.len) > > strbuf_addch(&sb, ' '); > > - if (short_magic & m->bit) > > - strbuf_addf(&sb, "'%c'", m->mnemonic); > > + > > + if (m->mnemonic) > > + strbuf_addf(&sb, "'(%c)%s'", m->mnemonic, m->name); > > else > > strbuf_addf(&sb, "'%s'", m->name); > > } > > The die() call is out of diff context, but it'll print > > pathspec magic not supported by this command: (!)top > > which looks too much like :(<name>)<mnemonic> pathspec syntax too me > and threw me off a bit. And it's a bit cryptic, isn't it? Since this > is meant for human, maybe we can just write > > pathspec magic not supported by this command: top (mnemonic: '!') I was trying to keep it short and sweet, turns out that ends up being more difficult to understand. I like your suggestion, it definitely makes things much clearer. -- Brandon Williams