From: Mark C. Chu-Carroll Date: Wed, 28 Oct 2026 14:11:36 GMT Subject: Re: [PATCH 02/12] rerere: convert flags into an enum Message-ID: <541b09ff60a5acd851aff471cf6fba75@fastmail.com> In-Reply-To: On Wed, 28 Oct 2026 09:33:25 +0100, Patrick Steinhardt wrote: > On Tue, Oct 27, 2026 at 12:40:45PM -0700, Junio C Hamano wrote: > > "Mark C. Chu-Carroll" writes: > > > > > On Tue, 27 Oct 2026 11:46:45 +0100, Patrick Steinhardt wrote: > > >> Our modern coding guidelines say that flags should typically be passed > > >> around as enums so that it becomes trivial to see what kind of flags are > > >> expected. Adapt the rerere flags accordingly. > > >> > > >> > > >> +enum rerere_flags { > > >> + RERERE_AUTOUPDATE = (1 << 0), > > >> + RERERE_NOAUTOUPDATE = (1 << 1), > > >> + RERERE_READONLY = (1 << 2), > > >> + /* If MERGE_RR.lock is taken, return -1 as if rerere were disabled */ > > >> + RERERE_NOWAIT = (1 << 3), > > >> +}; > > > > > > If we're moving to using enum flags, then wouldn't it be better to > > > explicitly add a "RERE_MERGE_LOCK_TOKEN" to the enum for that case, > > > rather than have this value only documented by a comment? > > > > Perhaps you misunderstand what the comment says. It is an > > explanation for the next line, i.e., it tells the API users "If you > > want the 'immediate failure when locked' behaviour, pass the _NOWAIT > > bit". > > Yeah, this. I guess we can help that though by documenting all of the > flags while at it. Ah, yeah. I misunderstood that. Obvious in retrospect, but I was puzzling over that for a while. -Mark