Re: [PATCH v4] builtin/remote.c: teach `-v` to list filters for promisor remotes

From: Abhradeep Chakraborty <chakrabortyabhradeep79@gmail.com>
Date: May 9, 2022, 17:21 UTC
Message-ID: <20220509172157.28593-1-chakrabortyabhradeep79@gmail.com>
In-Reply-To: <Ynk0mADTSJU/xVUd@nand.local>

Taylor Blau <me@ttaylorr.com> wrote:

Show 6 quoted lines

> But there was a good question raised by Phillip in
>
>     https://lore.kernel.org/git/ab047b4b-6037-af78-1af6-ad35ac6d7c90@iee.email/
>
> that I didn't see addressed in your response, which was "why not put
> this behind a new `--show-partial-filter` option"?

Actually, I addressed it[1] -

Show 18 quoted lines

> ... Another point is that
> it's important for an user to know which one is a promisor remote and what
> filter type they use. If we go with the current implementation the output
> would be let's say - 
> origin <remote-url> (fetch)
> origin <remote-url> (push)
> upstream <remote-url> (fetch)
> upstream <remote-url> (push)
>
> By seeing the above output anyone may assume that all the remotes are
> normal remotes. If the user now try to run `git pull origin` and suddenly
> he/she discover that some blobs are not downloaded. He/she run the above
> mentioned (1) command and find that this is a promisor remote!
>
> Here `remote -v` didn't warn the user about the origin remote being an
> promisor remote. Instead it makes him/her assume that all are normal
> remotes. Providing only these three info (i.e. <remote-name>, <remote-url>
> and <direction>) is not sufficient - it only shows the half of the picture.

If we use a new `--show-partial-clone` flag, users can get to know about promisor remotes only if he/she use this flag. As I said in the refered comment, it may happen that the user unfortunately use the flag AFTER the accident - to know about if that was the promisor remote!

See this also[2] -

> ... If
> we can specify `(fetch)` in the output then why not the filter of that
> `fetch` on which the behaviour of `fetch` functionality highly depends?

Taylor Blau <me@ttaylorr.com> wrote:

> But I can see where it _would_ be useful. So it would be nice to be able
> to turn the extra output on in those cases, but _only_ those cases, and
> a flag would be a nice way to go about doing that.

Adding the extra flag is not a good approach to me due to the above reason. But at the end of the day, all of you have a lots of experience in this field than me. You all could better tell which one is better approach.

[1] https://lore.kernel.org/git/20220501193807.94369-1-chakrabortyabhradeep79@gmail.com/ [2] https://lore.kernel.org/git/20220502145624.12702-1-chakrabortyabhradeep79@gmail.com/

Thanks :)