Re: [PATCH] replace: List replacement along with the object

From: Michael J Gruber <git@drmicha.warpmail.net>
Date: Aug 26, 2011, 07:38 UTC
Message-ID: <4E574D61.8050501@drmicha.warpmail.net>
In-Reply-To: <CAP8UFD2Cr4UQjWa=pRvcqgyX_Ed+qjts=TujWRdyk4dUZsd_7Q@mail.gmail.com>

Christian Couder venit, vidit, dixit 25.08.2011 18:29:

Show 7 quoted lines

> On Thu, Aug 25, 2011 at 4:39 PM, Michael J Gruber
> <git@drmicha.warpmail.net> wrote:
>> The documentation could be misunderstood as if "git replace -l" lists
>> the replacements of the specified objects. Currently, it lists the
>> replaced objects.
> 
> You could just change the documentation to make it more explicit.

Well, sure. I just didn't find the current form that useful.

Show 7 quoted lines

>> Change the output to the form "<object> <replacement>" so that there is
>> an easy way to find the replacement, besides the more difficult to find
>> git show-ref $(git replace -l).
> 
> I shamelessly copied the "-l <pattern>" feature and the documentation
> from "git tag". If you just change the output of "git replace -l" it
> will make the UI inconsistent between both commands.

I don't think many people will expect consistency between branch and tag on the one hand, and replace refs on the other hand. It requires the knowledge that a replacement is basically a lightweight tag stored in a different namespace in refs/, which I would actually consider an implementation detail.

> Maybe you could add a "-L <pattern>" feature to "git replace", "git
> tag" and "git branch" that would output "<ref name> <ref content>"?

I'd use "-v" then if this is about consistency, because that *always* means "verbose", and migrate the misnamed "git tag -v"...

Junio C Hamano venit, vidit, dixit 25.08.2011 21:07:

Show 8 quoted lines

> Michael J Gruber <git@drmicha.warpmail.net> writes:
> 
>> The documentation could be misunderstood as if "git replace -l" lists
>> the replacements of the specified objects. Currently, it lists the
>> replaced objects.
> Seeing that you had to change existing tests, I do not think this is an
> improvement. The existing scripts can read the list of objects and find
> replacement themselves (if they want to find that out, that is), no?

If "replace -l" is considered fair game for scripts then the output should probably not change, though I left the meaning of "$1" for each line of the output as is on purpose.

But, how would scripts find the replacement? rev-parse does not do it, rev-list does not do it, and using show-ref requires the user to know about the actual implementation as refs under refs/replace.

Seems that the doc change is the only option.

Michael