# [PATCH] git-send-email: Fix authenticating on some servers when using TLS. 2 messages from 2008-07-09 to 2008-07-09. Participants: Robert Shearman, Junio C Hamano. Thread: https://gitlist.dev/t/14386 ## Robert Shearman, 2008-07-09 21:39 Subject: [PATCH] git-send-email: Fix authenticating on some servers when using TLS. Message-ID: <1215639580-25040-1-git-send-email-robertshearman@gmail.com> URL: https://gitlist.dev/e/1215639580-25040-1-git-send-email-robertshearman%40gmail.com ``` Send hello after a successful STARTTLS command to refresh the list of extensions. These may be different to what is returned over a clear connection (for example the AUTH command may be accepted over a secure connection, but not over a clear connection). Furthermore, this behaviour is recommended by RFC 2487 (http://www.ietf.org/rfc/rfc2487.txt). Signed-off-by: Robert Shearman --- git-send-email.perl | 3 +++ 1 files changed, 3 insertions(+), 0 deletions(-) diff --git a/git-send-email.perl b/git-send-email.perl index 3564419..6adb669 100755 --- a/git-send-email.perl +++ b/git-send-email.perl @@ -773,6 +773,9 @@ X-Mailer: git-send-email $gitversion $smtp = Net::SMTP::SSL->start_SSL($smtp) or die "STARTTLS failed! ".$smtp->message; $smtp_encryption = ''; + # Send EHLO again to receive fresh + # supported commands + $smtp->hello(); } else { die "Server does not support STARTTLS! ".$smtp->message; } -- 1.5.6.2.224.g26efb.dirty ``` ## Junio C Hamano, 2008-07-09 22:51 Subject: Re: [PATCH] git-send-email: Fix authenticating on some servers when using TLS. Message-ID: <7v1w227jtu.fsf@gitster.siamese.dyndns.org> URL: https://gitlist.dev/e/7v1w227jtu.fsf%40gitster.siamese.dyndns.org In-Reply-To: <1215639580-25040-1-git-send-email-robertshearman@gmail.com> ``` Robert Shearman writes: > Send hello after a successful STARTTLS command to refresh the list of extensions. These may be different to what is returned over a clear connection (for example the AUTH command may be accepted over a secure connection, but not over a clear connection). > > Furthermore, this behaviour is recommended by RFC 2487 (http://www.ietf.org/rfc/rfc2487.txt). Very true. Thanks. ```