Re: [RFC PATCH] checkout: Force matching mtime between files

From: Duy Nguyen <pclouds@gmail.com>
Date: Apr 28, 2018, 06:16 UTC
Message-ID: <CACsJy8D9=f8KSEne3RKka_QkS9-c05bZ3D3vRcJ3Wq9NXaBkhQ@mail.gmail.com>
In-Reply-To: <38e961ca-e354-6390-982d-4d0c9c9e55da@xiplink.com>

On Fri, Apr 27, 2018 at 11:08 PM, Marc Branchaud <marcnarc@xiplink.com> wrote:

Show 25 quoted lines

>> On Wed, Apr 25, 2018 at 5:18 PM, Marc Branchaud <marcnarc@xiplink.com>
>> This is a limitation of the current post-checkout hook. $3==0 from the
>> hook lets us know this is not a branch switch, but it does not really
>> tell you the affected paths. If it somehow passes all the given
>> pathspec to you, then you should be able to do "git ls-files --
>> $pathspec" which gives you the exact same set of paths that
>> git-checkout updates. We could do this by setting $4 to "--" and put
>> all the pathspecs in $5+ [1] e.g. "HEAD@{1} HEAD 0 -- path/to/file" in
>> the above example.
>>
>> There is  third case here, if you do "git checkout <tree-ish> --
>> path/to/file" then it cannot be covered by the current design. I guess
>> we could set $3 to '2' (retrieve from a tree) to indicate this in
>> addition to 0 (from index) and 1 (from switching branch) and then $1
>> could be the tree in question (pathspecs are passed the same way
>> above)
>>
>> [1] I wonder if we could have a more generic approach to pass
>> pathspecs via environment, which could work for more than just this
>> one hook. Not sure if it's a good idea though.
>
>
> I think there needs to be something other than listing all the paths in the
> command is viable, because it's too easy to hit some command-line-length
> limit.

I send pathspecs, not paths. If you type "git checkout -- foo/" then I send exactly "foo/" not every paths in it. You can always figure that out with git-ls-files.

Sure this can still hit command length limit when you do "git checkout -- foo/*" and have lots of files in foo just one more param from hitting the limit, then the hook may hit the limit because we need more command line arguments. But this is the corner case I don't think we should really need to care about.

> It would also be good if hook authors didn't have to re-invent the
> wheel of determining the changed paths for every corner-case.

Flexibility vs convenience I guess. A sample hook as template should help the reinvention.

Show 6 quoted lines

> My first instinct is to write them one-per-line on the hook's stdin. That's
> probably not generic enough for most hooks, but it seems like a good
> approach for this proposal.
>
> Throwing them into a temporary file with a known name is also good ---
> better, I think, than stuffing them into an environment variable.

This goes back to my post-checkout-modified proposal. If you're writing to file, might as well reuse the index format. Then you can read it with ls-files (which lets you decide path separator or even quoting, I'm not sure) and it also provides some more info like file hashes, access time...

-- 
Duy