[PATCH v2] pull: warn if the user didn't say whether to rebase or to merge

[PATCH v2] pull: warn if the user didn't say whether to rebase or to merge

[Alex Henrie]

Often novice Git users forget to say "pull --rebase" and end up with an
unnecessary merge from upstream. What they usually want is either "pull
--rebase" in the simpler cases, or "pull --ff-only" to update the copy
of main integration branches, and rebase their work separately. The
pull.rebase configuration variable exists to help them in the simpler
cases, but there is no mechanism to make these users aware of it.

Issue a warning message when no --[no-]rebase option from the command
line and no pull.rebase configuration variable is given. This will
inconvenience those who never want to "pull --rebase", who haven't had
to do anything special, but the cost of the inconvenience is paid only
once per user, which should be a reasonable cost to help a number of new
users.

Signed-off-by: Alex Henrie <alexhenrie24@gmail.com>
---
v2:
- Use the commit message that Junio suggested
- Drop "!= 0" from the if condition
- Mention in the warning that --ff-only will also suppress the warning
---
 builtin/pull.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/builtin/pull.c b/builtin/pull.c
index 3e624d1e00..96512a4b27 100644
--- a/builtin/pull.c
+++ b/builtin/pull.c
@@ -327,6 +327,13 @@ static enum rebase_type config_get_rebase(void)
 	if (!git_config_get_value("pull.rebase", &value))
 		return parse_config_rebase("pull.rebase", value, 1);
 
+	if (strcmp(opt_ff, "--ff-only")) {
+		warning(_("Pulling without specifying whether to rebase or to merge is discouraged\n"
+			"and will be disallowed in a future Git release.\n"
+			"Next time, run `git config pull.rebase (true|false)` first\n"
+			"or pass --rebase, --no-rebase, or --ff-only on the command line.\n"));
+	}
+
 	return REBASE_FALSE;
 }
 
-- 
2.25.1

Re: [PATCH v2] pull: warn if the user didn't say whether to rebase or to merge

[Junio C Hamano]

Alex Henrie <alexhenrie24@gmail.com> writes:

> +		warning(_("Pulling without specifying whether to rebase or to merge is discouraged\n"
> +			"and will be disallowed in a future Git release.\n"

Sorry for not catching this in the earlier round, but I do not think
anybody argued, let alone the community came to a concensus on
doing, such a strong move.  Am I mistaken?

I certainly did not intend to, at least when I commented on the
earlier round and proposed an updated log message, I wasn't even
aware of the possibility that we may want to turn this into die()
after a transition period.

Not that I'd object strongly to the idea, but it was somewhat
unexpected.

If we are proposing to make it a long-term plan, that should
certainly be written down in the proposed log message.

> +			"Next time, run `git config pull.rebase (true|false)` first\n"
> +			"or pass --rebase, --no-rebase, or --ff-only on the command line.\n

I am somewhat puzzled by "first, or".  You certainly mean the config
to be "set and forget", and you do not want to say "before you pull,
do this first, always", but somehow the latter is the impression I
got.

But it does not sound to me like "Next time, and only next time, do
this configuration.  You can countermand the choice you make from
the command line later if needed", which I think is what you meant
to convey to your readers.
 
    You can squelch this message by `pull.rebase` configuration
    variable to show your preference.  By passing --[no-]rebase
    or --ff-only from the command line, you can countermand the
    choice per invocation.

is what I came up with, but I am not quite happy with it.  It is
overly long to start with X-<.

"));
> +	}
> +
>  	return REBASE_FALSE;
>  }

Re: [PATCH v2] pull: warn if the user didn't say whether to rebase or to merge

[Alex Henrie]

On Sat, Feb 29, 2020 at 9:51 AM Junio C Hamano <gitster@pobox.com> wrote:
>
> Alex Henrie <alexhenrie24@gmail.com> writes:
>
> > +             warning(_("Pulling without specifying whether to rebase or to merge is discouraged\n"
> > +                     "and will be disallowed in a future Git release.\n"
>
> Sorry for not catching this in the earlier round, but I do not think
> anybody argued, let alone the community came to a concensus on
> doing, such a strong move.  Am I mistaken?
>
> I certainly did not intend to, at least when I commented on the
> earlier round and proposed an updated log message, I wasn't even
> aware of the possibility that we may want to turn this into die()
> after a transition period.
>
> Not that I'd object strongly to the idea, but it was somewhat
> unexpected.
>
> If we are proposing to make it a long-term plan, that should
> certainly be written down in the proposed log message.

I really don't think it's a good idea to start making a merge commit
without confirming that that's what the user wants. Elijah and Robert
seemed to agree. However, we can start by just putting in a friendly
educational message and then continue the discussion about whether or
not to make a future version of Git abort upon printing the warning.

> > +                     "Next time, run `git config pull.rebase (true|false)` first\n"
> > +                     "or pass --rebase, --no-rebase, or --ff-only on the command line.\n
>
> I am somewhat puzzled by "first, or".  You certainly mean the config
> to be "set and forget", and you do not want to say "before you pull,
> do this first, always", but somehow the latter is the impression I
> got.
>
> But it does not sound to me like "Next time, and only next time, do
> this configuration.  You can countermand the choice you make from
> the command line later if needed", which I think is what you meant
> to convey to your readers.

I'll send a new version that has a longer and more clear warning message.

-Alex