diff mbox series

git-send-email: die if sendmail.* config is set

Message ID 20200718132311.27248-1-sir@cmpwn.com (mailing list archive)
State New, archived
Headers show
Series git-send-email: die if sendmail.* config is set | expand

Commit Message

Drew DeVault July 18, 2020, 1:23 p.m. UTC
I've seen several people mis-configure git send-email on their first
attempt because they set the sendmail.* config options - not
sendemail.*. This patch detects this mistake and bails out with a
friendly warning.

Signed-off-by: Drew DeVault <sir@cmpwn.com>
---
 git-send-email.perl |  6 ++++++
 perl/Git.pm         | 26 ++++++++++++++++++++++++++
 2 files changed, 32 insertions(+)

Comments

Junio C Hamano July 18, 2020, 6:02 p.m. UTC | #1
Drew DeVault <sir@cmpwn.com> writes:

> I've seen several people mis-configure git send-email on their first
> attempt because they set the sendmail.* config options - not
> sendemail.*. This patch detects this mistake and bails out with a
> friendly warning.

It is not very friendly thing to do, though.

It just closes the door for anybody to add something that works
independent from "git send-email", to which "sendmail.*" variables
may be appropriate knob to use.

Demoting the "die" to "warn" is OK, and limiting the check to
variables that actually has corresponding and likely-misspelt
"sendemail.*" counterparts would be even better, but "you are not
allowed to have any 'sendmail.*' variables, ever" is way too much,
I am afraid.

> Signed-off-by: Drew DeVault <sir@cmpwn.com>
> ---
>  git-send-email.perl |  6 ++++++
>  perl/Git.pm         | 26 ++++++++++++++++++++++++++
>  2 files changed, 32 insertions(+)
>
> diff --git a/git-send-email.perl b/git-send-email.perl
> index 36c47bae1d..8e42ba00c1 100755
> --- a/git-send-email.perl
> +++ b/git-send-email.perl
> @@ -478,6 +478,12 @@ sub read_config {
>      usage();
>  }
>  
> +if ((scalar Git::config_regexp("sendmail.*")) != 0) {
> +	die __("fatal: found configuration options for 'sendmail'\n" .
> +		"git-send-email is configured with the sendemail.* options - note the 'e'.\n" .
> +		"Assuming this is a mistake and bailing out.\n");
> +}
Drew DeVault July 18, 2020, 6:25 p.m. UTC | #2
On Sat Jul 18, 2020 at 7:02 AM EDT, Junio C Hamano wrote:
> It is not very friendly thing to do, though.
>
> It just closes the door for anybody to add something that works
> independent from "git send-email", to which "sendmail.*" variables
> may be appropriate knob to use.
>
> Demoting the "die" to "warn" is OK, and limiting the check to
> variables that actually has corresponding and likely-misspelt
> "sendemail.*" counterparts would be even better, but "you are not
> allowed to have any 'sendmail.*' variables, ever" is way too much,
> I am afraid.

I originally had this just a warning, but it can be difficult to see.
There's a lot of text printed from git send-email, a lot of it looks
like diagnostics, and it can be easy to lose a message in there. And if
you pass --annotate, which I encourage people to do, your editor
immediately covers up the warning.

I wonder if a happy medium would be adding a config option which
squelches the message? If sendemail.squelchConfigError is set, for
example. If not, I'll experiment with a noisy warning.
Junio C Hamano July 18, 2020, 8:10 p.m. UTC | #3
"Drew DeVault" <sir@cmpwn.com> writes:

> I originally had this just a warning, but it can be difficult to see.
> There's a lot of text printed from git send-email, a lot of it looks
> like diagnostics, and it can be easy to lose a message in there. And if
> you pass --annotate, which I encourage people to do, your editor
> immediately covers up the warning.

OK.

> I wonder if a happy medium would be adding a config option which
> squelches the message? If sendemail.squelchConfigError is set, for
> example.

I think with "squelches the message" replaced with "bypasses the
check", it would probably be a good idea.  It may be that a
reasonable behaviour would be to check for any "sendemail.*" and
die, like your version does, by default, an optional opt-out
configuration variable (sendemail.forbidSendmailVariables) to
disable the check entirely, and then add a hint that mentions the
variable to the "die" message in your patch.
diff mbox series

Patch

diff --git a/git-send-email.perl b/git-send-email.perl
index 36c47bae1d..8e42ba00c1 100755
--- a/git-send-email.perl
+++ b/git-send-email.perl
@@ -478,6 +478,12 @@  sub read_config {
     usage();
 }
 
+if ((scalar Git::config_regexp("sendmail.*")) != 0) {
+	die __("fatal: found configuration options for 'sendmail'\n" .
+		"git-send-email is configured with the sendemail.* options - note the 'e'.\n" .
+		"Assuming this is a mistake and bailing out.\n");
+}
+
 die __("Cannot run git format-patch from outside a repository\n")
 	if $format_patch and not $repo;
 
diff --git a/perl/Git.pm b/perl/Git.pm
index 54c9ed0dde..10df990959 100644
--- a/perl/Git.pm
+++ b/perl/Git.pm
@@ -723,6 +723,32 @@  sub config_int {
 	return scalar _config_common({'kind' => '--int'}, @_);
 }
 
+=item config_regexp ( RE )
+
+Retrieve the list of configuration key names matching the regular
+expression C<RE>. The return value is a list of strings matching
+this regex.
+
+=cut
+
+sub config_regexp {
+	my ($self, $regex) = _maybe_self(@_);
+	try {
+		my @cmd = ('config', '--name-only', '--get-regexp', $regex);
+		unshift @cmd, $self if $self;
+		my @matches = command(@cmd);
+		return @matches;
+	} catch Git::Error::Command with {
+		my $E = shift;
+		if ($E->value() == 1) {
+			my @matches = ();
+			return @matches;
+		} else {
+			throw $E;
+		}
+	};
+}
+
 # Common subroutine to implement bulk of what the config* family of methods
 # do. This currently wraps command('config') so it is not so fast.
 sub _config_common {