Re: [QUESTION] how to quickly find an oid in the git repository

From: ZheNing Hu <adlternative@gmail.com>
Date: May 31, 2021, 11:30 UTC
Message-ID: <CAOLTT8Rsy-XRys6fv3EbUJSUbHFCG4=w9uR7Bz8N8u8PWHgPhw@mail.gmail.com>
In-Reply-To: <YLP7kuFNfVZK+xCK@coredump.intra.peff.net>

Jeff King <peff@peff.net> 于2021年5月31日周一 上午4:54写道：

Show 12 quoted lines

> > On the other hand, oid_object_info_extended() will parse an object
> > in depth, using it will seriously affect performance. So I want to know
> > if there is a function that can quickly find the oid? I searched in
> > `object-name.c`, `object-file.c`, and there seems to be no clue...
>
> You should be able to use oid_object_info_extended() here. It will try
> to do as little work as possible to fulfill the items requested in the
> object_info struct. So a blank one (or even passing NULL) will return an
> error if the object doesn't exist, but not otherwise.
>
> There's also has_object_file(), which is essentially a wrapper for this.
>

This is exactly what I want. It seems that the real performance problem is not because of the use of oid_object_info_extended(), but the possible copy of the object data in it.

> -Peff

Thanks! -- ZheNing Hu