Web lists-archives.com

Re: [PATCH 4/4] get_oid(): when an object was not found, try harder




"Johannes Schindelin via GitGitGadget" <gitgitgadget@xxxxxxxxx>
writes:

> @@ -442,6 +442,18 @@ static enum get_oid_result get_short_oid(const char *name, int len,
>  	find_short_packed_object(&ds);
>  	status = finish_object_disambiguation(&ds, oid);
>  
> +	/*
> +	 * If we didn't find it, do the usual reprepare() slow-path,
> +	 * since the object may have recently been added to the repository
> +	 * or migrated from loose to packed.
> +	 */
> +	if (status == MISSING_OBJECT) {
> +		reprepare_packed_git(the_repository);
> +		find_short_object_filename(&ds);
> +		find_short_packed_object(&ds);
> +		status = finish_object_disambiguation(&ds, oid);
> +	}
> +

This looks obviously correct, but two things that made me wonder
briefly were:

 1. is reprepare_packed_git() a bit too heavy-weight, if the only
    thing we are addressing is the loose-object cache going stale?

 2. is there a way to cleanly avoid the three-line duplicate?

My tentative answers are (1) even if it is, but get_short_oid() is
already heavy-weight enough; it won't be worth restructuring the
code to make it possible to clear only the loose-object cache, and
(2) a loop that runs twice when the first result is MISSING_OBJECT
and otherwise leaves after once would need an extra variable, its
iniialization, check and increment, which is more than what we might
save with such a restructuring, so it won't be worth pursuing.

But others may have better ideas, as always ;-)