Re: [PATCH 038/194] pack: allow sha1_loose_object_info to handle arbitrary repositories

From: Jonathan Tan <jonathantanmy@google.com>
Date: Feb 7, 2018, 22:33 UTC
Message-ID: <20180207143300.ce1c39ca07f6a0d64fe0e7ca@google.com>
In-Reply-To: <20180205235735.216710-18-sbeller@google.com>

On Mon, 5 Feb 2018 15:54:59 -0800 Stefan Beller <sbeller@google.com> wrote:

Show 7 quoted lines

> From: Jonathan Nieder <jrnieder@gmail.com>
> 
> Signed-off-by: Stefan Beller <sbeller@google.com>
> Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
> ---
>  sha1_file.c | 11 +++++------
>  1 file changed, 5 insertions(+), 6 deletions(-)

Thanks - I can see that this has been a lot of work, and it brings a lot of benefit. Among other things, this will probably allow me to remove the "fetch_if_missing" global variable that we need to set and reset in the partial clone series, replacing it with a setting in either the repo or the object store (and when fetching a missing object, first cloning the repo/store and then setting that setting, so that objects are not recursively fetched).

If we're planning to split the series up for merging in batches (which, as a reviewer, I'm very much in favor of), I think this is a good stopping point for the first batch, so I'll stop my review here for now. After these 38 patches, the net benefit is a better position of the packed object variables (in struct repository) and permitting the reading of loose objects in any repository. (Permitting the reading of any object in any repository, I see, will only come later in patch 74.)