Hi,
I'm using git version 1.5.4.rc3.37.gfdcf3 and when I use
git-send-email with the --compose option like this:
git send-email --compose -- a_dir_of_patches/
The intro email gets sent, but not any of the actual patches. If I
ommit the --compose option, then the patches do get sent. Pretty sure
this worked for me until a recent update so I wonder if there's a
buglet in the recent validation patches (or I'm in error and missing
some needed config?).
+ seth
--
Seth Falcon | seth@userprimary.net | blog: http://userprimary.net/user/
From: Gustaf Hendeby <hidden> Date: 2016-06-15 22:44:07
This fixes the subtile bug in git send-email that was introduced into
git send-email with
commit aa54892f5ada8282643dc7387b33261c7135d784
send-email: detect invocation errors earlier
causing no patches to be sent out if the --compose flag was used.
Signed-off-by: Gustaf Hendeby <redacted>
---
git-send-email.perl | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
@@ -462,7 +462,7 @@ EOTexit(0);}-@files=($compose_filename.".final");+@files=($compose_filename.".final",@files);}# Variables we set as part of the loop over files
From: Jeff King <hidden> Date: 2016-06-15 22:44:07
On Mon, Jan 21, 2008 at 08:57:46PM +0100, Gustaf Hendeby wrote:
This fixes the subtile bug in git send-email that was introduced into
git send-email with
commit aa54892f5ada8282643dc7387b33261c7135d784
send-email: detect invocation errors earlier
Sorry, this was my fault. I have looked over the code again to make sure
there are not any other places where the movement of @files should
matter, but this seems to be the only one.
As penance, I'll send some tests in a moment.
I think this might be more readable as
unshift @files, "$compose_filename.final";
but either way it fixes the bug, so
Acked-by: Jeff King <redacted>
-Peff
From: Jeff King <hidden> Date: 2016-06-15 22:44:07
The exit value of some commands was not being used for the
test output.
Signed-off-by: Jeff King <redacted>
---
On Mon, Jan 21, 2008 at 10:19:32PM -0500, Jeff King wrote:
As penance, I'll send some tests in a moment.
This turned into a 3-patch series because of some cleanup needed. This
first one is just tightening up the setup tests, which will be modified
in the next patch.
t/t9001-send-email.sh | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)
From: Jeff King <hidden> Date: 2016-06-15 22:44:07
Previously, the fake.sendmail test harness would write its
output to a hardcoded file, allowing only a single message
to be tested. Instead, let's have it save the messages for
all of its invocations so that we can see which messages
which were sent, and in which order.
Signed-off-by: Jeff King <redacted>
---
This shouldn't change the test behavior at all, but is needed for the
test in 3/3.
t/t9001-send-email.sh | 16 +++++++++++-----
1 files changed, 11 insertions(+), 5 deletions(-)
@@ -82,7 +88,7 @@ z8=zzzzzzzzz64=$z8$z8$z8$z8$z8$z8$z8$z8z512=$z64$z64$z64$z64$z64$z64$z64$z64 test_expect_success'reject long lines''-rm-fcommandline&&+clean_fake_sendmail&&cp$patcheslongline.patch&&echo$z512$z512>>longline.patch&&!gitsend-email\
@@ -95,7 +101,7 @@ test_expect_success 'reject long lines' '' test_expect_success'no patch was sent''-!test-ecommandline+!test-ecommandline1' test_expect_success'allow long lines with --no-validate''
From: Jeff King <hidden> Date: 2016-06-15 22:44:07
This is just a basic sanity check that --compose works at
all. Unfortunately, we have to add a --no-confirm-send
option to git-send-email to make this work, since it
otherwise unconditionally tries to talk to the terminal.
Signed-off-by: Jeff King <redacted>
---
Note that the change to git-send-email depends textually on Gustaf's
fix (actually, on my suggested version). I am not happy with adding the
--no-confirm-send option, but there just isn't any way to avoid the
readline call with the existing code short of providing a dummy perl
module in the test harness.
git-send-email.perl | 16 ++++++++++------
t/t9001-send-email.sh | 27 +++++++++++++++++++++++++++
2 files changed, 37 insertions(+), 6 deletions(-)