git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Junio C Hamano <gitster@pobox.com>
To: "Matěj Cepl" <mcepl@cepl.eu>
Cc: git@vger.kernel.org
Subject: Re: [PATCH] RFC: git-request-get maintainer receiving from git-request-pull
Date: Tue, 07 Jan 2025 11:37:24 -0800	[thread overview]
Message-ID: <xmqqcygys9e3.fsf@gitster.g> (raw)
In-Reply-To: <20250107185716.1562311-1-mcepl@cepl.eu> ("Matěj Cepl"'s message of "Tue, 7 Jan 2025 19:56:35 +0100")

Matěj Cepl <mcepl@cepl.eu> writes:

> This should be fulfilment of my comment on the GitLab bug [1] from 2016-05-01:
>
>> Well, the minimal solution would be a parser for the
>> git-request-pull(1), which would check every comment and if recognized
>> the comment as the pull request, it would add button for the owner of
>> the repo (e.g., [Create Pull Request]). After pressing that (so it is
>> constantly under the control of the owner repo), gitlab would setup
>> new remote (if doesn't exist still), fetch it, and create a merge
>> request.
>
> In the end, the best UI I could come up with is to pull all
> commits to the new review branch and send all of them via
> git-send-email(1) to the review email list.
>
> The script can either receive git-request-pull(1) output on
> stdin, or it can be driven by the command line arguments with URL
> and (optionally) branch names for all those “I have some commits
> on SOMEWHERE, but I don’t want to fiddle with git-send-email(1)
> just because of that.”
>
> [1] https://gitlab.com/gitlab-org/gitlab/-/issues/14116
> ---
>
> I would like to get these types of comments (before I finalize
> a proper commit for git):
>
> * whether the idea of the script makes sense at all (or
>   suggestions on the UI improvement),

You'd have much better chance if you made the patch standalone,
instead of discouraging those who might be interested by forcing
them to first visit an external website for 8-year old "bug" that
may or may not be interesting to them.  From the above, I am not
sure if people can guess what problem is being solved, and without
knowing that, they cannot judge if script makes sense in the first
place.

> * a general code review

There are a handful of scripts written in Perl still in our
codebase; study the way how their sources are written (e.g., notice
that they all end with ".perl") and are made into executables by
Makefile rules, and mimick them.

> * advice how to proceed with testing before rewriting it into the
>   proper git test harness: is there only way to add those ugly
>   `if test -z "$DEBUG"` constructs everywhere, or there is some
>   better way?

Also send the e-mailed patch to yourself, and then run "git am" to
make sure it applies cleanly.  It seems the patch is creating a new
Makefile, but our project already has one, so I do not see how this
patch would apply.

Thanks for your interest in our project.

> Thank you for any reply,
>
> Matěj
>
>  Makefile        |   8 +++
>  git-request-get | 186 ++++++++++++++++++++++++++++++++++++++++++++++++
>  2 files changed, 194 insertions(+)
>  create mode 100644 Makefile
>  create mode 100755 git-request-get
>
> diff --git a/Makefile b/Makefile
> new file mode 100644
> index 0000000..4d686f5
> --- /dev/null
> +++ b/Makefile
> @@ -0,0 +1,8 @@
> +export DEBUG := true
> +export WANT_REMOTE := 1
> +
> +test:
> +	./git-request-get < gitk-email.mbx 2>&1 | \
> +	grep -e "ensure_remote: getting REMOTE = j6t and URL = https://github.com/j6t/gitk.git" \
> +		-e "make_branch: getting BRANCH_NAME = mcepl-07012025 and HEAD = master" \
> +		-e "main: DEST_ADDR= ~mcepl/git-test@lists.sr.ht, AUTHOR=Matěj Cepl <mcepl@cepl.eu>, BASE=c18400c6bb04f4e8875c1930db72ddd9b94649ca, HEAD=master"
> diff --git a/git-request-get b/git-request-get
> new file mode 100755
> index 0000000..1f6edfa
> --- /dev/null
> +++ b/git-request-get
> @@ -0,0 +1,186 @@
> +#!/bin/sh
> +# Requires: perl, sed, awk
> +set -eux
> +
> +# Initialize internal variables
> +DEBUG="${DEBUG:-}"
> +WANT_REMOTE="${WANT_REMOTE:-"$(git config get --type int --default=0 git-request-get.wantRemote)"}"
> +BASE=""
> +HEAD=""
> +DEST_ADDR=""
> +
> +usage() {
> +    echo "usage: $(basename "$0") [-t ADDRTO] [-s] [-h] URL [HEAD] [BASE]" >&2
> +    echo "Generate git-send-email messages from remote URL or from the email message"
> +    echo "formatted by git-request-pull"
> +    echo
> +    echo "-t ADDRTO send emails to this email address"
> +    exit 1
> +}
> +
> +die() {
> +    echo "$1" >&2
> +    if test $# -gt 1
> +    then
> +        exit "$2"
> +    else
> +        exit 1
> +    fi
> +}
> +
> +log() {
> +    echo "$1" >&2
> +}
> +
> +parse_message() {
> +    STR="$(perl -pe 'use MIME::QuotedPrint; $_=MIME::QuotedPrint::decode($_);')"
> +    # Collect URL of the incoming repo
> +    USTR="$(echo "$STR" | sed -n -e '/^are available in the Git repository at:/{n;n;p}')"
> +    U="$(echo "$USTR" | awk '{print $1}')"
> +
> +    # Find the tip of the offered branch
> +    # for you to fetch changes up to 102cd4d5c1818ecb5926dd538962f8b42c43b022:
> +    B="$(echo "$STR" | awk '/^The following changes since commit / { print $NF }' | sed -e 's/[=:]*$//')"
> +
> +    E="$(echo "$USTR" | awk '{print $2}')"
> +    if test -z "$E"
> +    then
> +        E="$(echo "$STR" | awk '/^for you to fetch changes up to / { print $NF }' | sed -e 's/[=:]*$//')"
> +    fi
> +    printf "%s^%s^%s" "$U" "$B" "$E"
> +}
> +
> +ensure_remote () {
> +    URL="$1"
> +    REMSTR="$(git remote -v | awk -v url="$1" '(NR%2 == 1) && index($0, url) {print $1, $2}')"
> +
> +    if test -n "$REMSTR"
> +    then
> +        log "REMSTR found: ${REMSTR}\n"
> +        REMOTE="$(echo "$REMSTR"|cut -d ' ' -f 1)"
> +    else
> +        # try to generate some good name for the remote
> +        case "$URL" in
> +        *github.com*|*gitlab.com*)
> +            REMOTE="$(echo "$URL"|cut -d/ -f 4)"
> +            ;;
> +        *sr.ht*)
> +            REMOTE="$(echo "$URL"|awk -F'[~/]' '{print $5}')"
> +            ;;
> +        *)
> +            REMOTE="_4review"
> +            ;;
> +        esac
> +        if test -z "$DEBUG"
> +        then
> +            git remote add "$REMOTE" "$URL"
> +            git remote update "$REMOTE"
> +        else
> +            printf "ensure_remote: getting REMOTE = %s and URL = %s\n" "$REMOTE" "$URL" >&2
> +        fi
> +    fi
> +
> +    printf "%s" "$REMOTE"
> +}
> +
> +#### main
> +while getopts ":dht:" arg ; do
> +    case $arg in
> +        h)
> +            usage
> +            ;;
> +        t)
> +            DEST_ADDR="$OPTARG"
> +            ;;
> +        ?)
> +            die "Invalid option: -${OPTARG}." 5
> +            ;;
> +    esac
> +done
> +
> +shift $((OPTIND - 1))
> +
> +MASTER="master"
> +if git rev-parse -q --verify main >/dev/null 2>&1
> +then
> +    MASTER="main"
> +fi
> +
> +if test $# -lt 1
> +then
> +    PARSED_STR="$(parse_message)"
> +    test -n "$PARSED_STR" || die "Not enough arguments and no message in stdin!" 2
> +    URL="$(echo "$PARSED_STR"|cut -d '^' -f 1)"
> +    BASE="$(echo "$PARSED_STR"|cut -d '^' -f 2)"
> +    HEAD="$(echo "$PARSED_STR"|cut -d '^' -f 3)"
> +# Ignore all command line arguments when stdin parsed
> +else
> +    if test $# -ge 1
> +    then
> +        URL="$1"
> +    fi
> +
> +    if test -z "$URL"
> +    then
> +        die "Insufficient input: no parseable input or command line arguments." 3
> +    fi
> +
> +    if test $# -ge 2
> +    then
> +        HEAD="$2"
> +    else
> +        HEAD="FETCH_HEAD"
> +    fi
> +
> +    # find the base commit where PR roots in the default branch
> +    if test $# -ge 3
> +    then
> +        BASE="$3"
> +    fi
> +fi
> +
> +if test "$WANT_REMOTE" -eq 1
> +then
> +    REMOTE="$(ensure_remote "$URL")"
> +    if test -n "$HEAD" && ! git rev-parse -q --verify "$HEAD"
> +    then
> +        git rev-parse -q --verify "$REMOTE/$HEAD" || die "Cannot find $HEAD branch on the remote $REMOTE!" 4
> +        HEAD="${REMOTE}/${HEAD}"
> +    fi
> +
> +    BRANCH_NAME="$(git log --pretty="%aL" -1)-$(date '+%d%m%Y')"
> +
> +    if test -z "$DEBUG"
> +    then
> +        git branch "$BRANCH_NAME" "$HEAD"
> +    else
> +        printf "make_branch: getting BRANCH_NAME = %s and HEAD = %s\n" "$BRANCH_NAME" "$HEAD" >&2
> +    fi
> +else
> +    git fetch "$URL"
> +fi
> +
> +if test -z "$DEST_ADDR"
> +then
> +    DEST_ADDR="$(git config --default='' sendemail.to)"
> +    test -z "$DEST_ADDR" && die "Unknown destination for the emails" 4
> +fi
> +
> +test "$HEAD" = "FETCH_HEAD" && HEAD="$(git rev-parse -q --verify FETCH_HEAD)"
> +
> +ROOT="$(git merge-base "$MASTER" "$HEAD")"
> +
> +# commit is not part of our tree
> +test "$ROOT" = "$HEAD" && die "Final commit cannot be included in our tree" 4
> +
> +test -z "$BASE" && BASE="$ROOT"
> +
> +test "$BASE" = "$HEAD" && die "No change observed" 1
> +
> +AUTHOR="$(git log --pretty="%an <%ae>" -1)"
> +if test -z "$DEBUG"
> +then
> +    git send-email --to="$DEST_ADDR" --from="$AUTHOR" --no-annotate --no-cc --no-bcc --no-identity --quiet "${BASE}..${HEAD}"
> +else
> +    printf "main: DEST_ADDR= %s, AUTHOR=%s, BASE=%s, HEAD=%s\n" "$DEST_ADDR" "$AUTHOR" "$BASE" "$HEAD" 
> +fi

  reply	other threads:[~2025-01-07 19:37 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-01-07 18:56 [PATCH] RFC: git-request-get maintainer receiving from git-request-pull Matěj Cepl
2025-01-07 19:37 ` Junio C Hamano [this message]
2025-01-07 20:20   ` Matěj Cepl
2025-01-07 21:05     ` Junio C Hamano

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=xmqqcygys9e3.fsf@gitster.g \
    --to=gitster@pobox.com \
    --cc=git@vger.kernel.org \
    --cc=mcepl@cepl.eu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).