Re: [PATCH 02/15] ref-filter: implement %(if:equals=<string>) and %(if:notequals=<string>)

From: Junio C Hamano <gitster@pobox.com>
Date: Mar 8, 2016, 06:04 UTC
Message-ID: <CAPc5daVvD53KcX3RQkLiUtc5fe+hksYtQw7dK5+C=1eQdTT24w@mail.gmail.com>
In-Reply-To: <CAOLa=ZSFD39HXSw8jo1=ehiKpeqjLUSfc0NkRJD0BSbqAvDwuQ@mail.gmail.com>

On Mon, Mar 7, 2016 at 9:41 PM, Karthik Nayak <karthik.188@gmail.com> wrote:

Show 5 quoted lines

>
> You're correct, the "handler" functions run once for each "ref". But WRT
> to the %(if)...%(then)...%(else)...%(end) atoms, it needs to be. Because
> each outcome of these atoms depend on the current outcome of the fields
> used between then WRT to the current ref.

Yeah, there is no argument against that part at all. My comment was purely about preparsing the if_equals field. All other fields are preparsed in *parse() helper functions and it looked strange that only these two are scanned in the handle() function every time it is evaluated for each ref, only to yield the same result.