get_maintainer.pl sends bogus addresses to git send-email

For me sending patches via git send-email fails because email address conversion is failing. Something appends a ')' to x86/lkml@kernel.org. I suspect the double '))' in MAINTAINERS is confusing the command. I tried to send the trivial patch from v5.0 and v5.6 tag.

Is this a failure in ./scripts/get_maintainer.pl, or is this something git does internally? I'm sure others use such command on a daily basis, so likely something on my end became broken at some point in the past.

Olaf

linux.git $ git send-email --reroll-count 1 --confirm=always --annotate --to-cmd ./scripts/get_maintainer.pl  HEAD^

(mbox) Adding cc: Olaf Hering <olaf@aepfle.de> from line 'From: Olaf Hering <olaf@aepfle.de>' (body) Adding cc: Olaf Hering <olaf@aepfle.de> from line 'Signed-off-by: Olaf Hering <olaf@aepfle.de>' (to-cmd) Adding to: "K. Y. Srinivasan" <kys@microsoft.com> from: './scripts/get_maintainer.pl' (to-cmd) Adding to: Haiyang Zhang <haiyangz@microsoft.com> from: './scripts/get_maintainer.pl' (to-cmd) Adding to: Stephen Hemminger <sthemmin@microsoft.com> from: './scripts/get_maintainer.pl' (to-cmd) Adding to: Sasha Levin <sashal@kernel.org> from: './scripts/get_maintainer.pl' (to-cmd) Adding to: Thomas Gleixner <tglx@linutronix.de> from: './scripts/get_maintainer.pl' (to-cmd) Adding to: Ingo Molnar <mingo@redhat.com> from: './scripts/get_maintainer.pl' (to-cmd) Adding to: Borislav Petkov <bp@alien8.de> from: './scripts/get_maintainer.pl' (to-cmd) Adding to: "H. Peter Anvin" <hpa@zytor.com> from: './scripts/get_maintainer.pl' (to-cmd) Adding to: x86@kernel.org (maintainer:X86 ARCHITECTURE (32-BIT AND 64-BIT)) from: './scripts/get_maintainer.pl' (to-cmd) Adding to: devel@linuxdriverproject.org (open list:Hyper-V CORE AND DRIVERS) from: './scripts/get_maintainer.pl' (to-cmd) Adding to: linux-kernel@vger.kernel.org (open list:X86 ARCHITECTURE (32-BIT AND 64-BIT)) from: './scripts/get_maintainer.pl'

From: Olaf Hering <olaf@aepfle.de>
To: "K. Y. Srinivasan" <kys@microsoft.com>,
        Haiyang Zhang <haiyangz@microsoft.com>,
        Stephen Hemminger <sthemmin@microsoft.com>,
        Sasha Levin <sashal@kernel.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>,
        Borislav Petkov <bp@alien8.de>,
        "H. Peter Anvin" <hpa@zytor.com>,
        x86@kernel.org) (maintainer:X86 ARCHITECTURE (32-BIT AND 64-BIT),
        devel@linuxdriverproject.org (open list:Hyper-V CORE AND DRIVERS),
        linux-kernel@vger.kernel.org) (open list:X86 ARCHITECTURE (32-BIT AND 64-BIT)
Cc: Olaf Hering <olaf@aepfle.de>
X-Mailer: git-send-email 2.16.4
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit

Send this email? ([y]es|[n]o|[q]uit|[a]ll): y 5.1.3 Invalid character in domain: <x86@kernel.org)>

On Tue, Apr 07, 2020 at 05:40:46PM +0200, Olaf Hering wrote:

It's a bug in send-email's hand-rolled address parser, which was fixed in bd869f67b9 (send-email: add and use a local copy of Mail::Address, 2018-01-05). Upgrade to Git v2.17.0 or newer.

-Peff

On Tue, 2020-04-07 at 13:02 -0400, Jeff King wrote:

Not really. You need to add --norolestats on the get_maintainer command line

git send-email expects bare email addresses, not ones annotated with additional content.

For instance:

$ ./scripts/get_maintainer.pl -f lib/vsprintf.c Petr Mladek <pmladek@suse.com> (maintainer:VSPRINTF) Steven Rostedt <rostedt@goodmis.org> (maintainer:VSPRINTF) Sergey Senozhatsky <sergey.senozhatsky@gmail.com> (maintainer:VSPRINTF) Andy Shevchenko <andriy.shevchenko@linux.intel.com> (reviewer:VSPRINTF) Rasmus Villemoes <linux@rasmusvillemoes.dk> (reviewer:VSPRINTF) linux-kernel@vger.kernel.org (open list)

vs:

$ ./scripts/get_maintainer.pl -f --norolestats lib/vsprintf.c Petr Mladek <pmladek@suse.com> Steven Rostedt <rostedt@goodmis.org> Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Andy Shevchenko <andriy.shevchenko@linux.intel.com> Rasmus Villemoes <linux@rasmusvillemoes.dk> linux-kernel@vger.kernel.org

Am Tue, 07 Apr 2020 10:18:41 -0700 schrieb Joe Perches <joe@perches.com>:

Thanks, this can be used as a workaround for the time being. Not sure why anyone would actually care about such details in default mode...

Olaf

On Tue, 2020-04-07 at 19:29 +0200, Olaf Hering wrote:

Because the default is "tell me more about the maintainers of a particular file". which can include not just the default maintainers of a particular file within a subsystem, but the also info about the people that actually apply patches to files.

For instance, a patch made to a file often has a nominal maintainer that doesn't actually apply the patches but that maintainer may review or approve but not actually be the upstream path for acceptance of the patch.

On Tue, Apr 07, 2020 at 10:18:41AM -0700, Joe Perches wrote:

I agree that dropping them from the output is even better, if you'd never want them to be sent.

Syntactically they are rfc822 comments, and send-email _should_ be able to handle them (and does in recent versions).

In all of these cases send-email will drop the bit in parentheses.

In this one, I think that the comment will be used as the name field, since there isn't one.

-Peff

On Tue, 2020-04-07 at 13:44 -0400, Jeff King wrote:

I'm not certain that comments are allowed _after_ a tld in an email address. In any case, I guess it's a good thing I used parentheses for the get_maintainer rolestats block.

I think that slightly unexpected as the name field is not required.

cheers, Joe

On Tue, Apr 07, 2020 at 02:56:19PM -0700, Joe Perches wrote:

Oh, it's much more horrible than that. RFC822 contains this example:

  Muhammed.(I am  the greatest) Ali @(the)Vegas.WBA

which parses to:

  Muhammed.Ali@Vegas.WBA

Perl's Mail::Address does decipher that correctly.

TBH, so do I. That's all done by Mail::Address's format() method. We could probably convince it to be less magical, but perhaps it's best to just leave it alone. Presumably that logic has some historical basis, and as you note, it's a mistake to be passing these fields into send-email in the first place.

-Peff