# [PATCH] format-patch: allow --no-patch to disable patch output 5 messages from 2016-04-07 to 2016-04-08. Participants: Jacob Keller, Eric Sunshine, Junio C Hamano. Thread: https://gitlist.dev/t/41954 ## Jacob Keller, 2016-04-07 16:46 Subject: [PATCH] format-patch: allow --no-patch to disable patch output Message-ID: <1460047599-9267-1-git-send-email-jacob.e.keller@intel.com> URL: https://gitlist.dev/e/1460047599-9267-1-git-send-email-jacob.e.keller%40intel.com ``` The documentation for format-patch indicates that --no-patch wilL suppress patch output. It also incorrectly mentions that -s will also suppress the patch output, but this is incorrect because -s is used to add the sign-off line first. Fix the documentation to remove the indication about -s. Fix builtin format-patch to prevent always adding the patch incase --no-patch is given. An alternative approach would be to remove --no-patch from the documentation for git-format-patch, and add an error check to indicate that --no-patch is not supported. Add a test for --no-patch to ensure it functions as expected. Signed-off-by: Jacob Keller --- Documentation/diff-options.txt | 2 ++ builtin/log.c | 4 ++-- t/t4014-format-patch.sh | 26 ++++++++++++++++++++++++++ 3 files changed, 30 insertions(+), 2 deletions(-) diff --git a/Documentation/diff-options.txt b/Documentation/diff-options.txt index 4b0318e2ac15..471ee47d0b07 100644 --- a/Documentation/diff-options.txt +++ b/Documentation/diff-options.txt @@ -27,7 +27,9 @@ ifdef::git-diff[] This is the default. endif::git-diff[] +ifndef::git-format-patch[] -s:: +endif::git-format-patch[] --no-patch:: Suppress diff output. Useful for commands like `git show` that show the patch by default, or to cancel the effect of `--patch`. diff --git a/builtin/log.c b/builtin/log.c index dff3fbbb437c..10f74837945b 100644 --- a/builtin/log.c +++ b/builtin/log.c @@ -1390,8 +1390,8 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix) rev.diffopt.output_format == DIFF_FORMAT_PATCH)) rev.diffopt.output_format = DIFF_FORMAT_DIFFSTAT | DIFF_FORMAT_SUMMARY; - /* Always generate a patch */ - rev.diffopt.output_format |= DIFF_FORMAT_PATCH; + if (!(rev.diffopt.output_format & DIFF_FORMAT_NO_OUTPUT)) + rev.diffopt.output_format |= DIFF_FORMAT_PATCH; rev.zero_commit = zero_commit; diff --git a/t/t4014-format-patch.sh b/t/t4014-format-patch.sh index eed2981b96df..ef9cfca30a9e 100755 --- a/t/t4014-format-patch.sh +++ b/t/t4014-format-patch.sh @@ -618,6 +618,32 @@ test_expect_success 'format-patch -p suppresses stat' ' ' +cat >expect << EOF +From 422768ba4254e7abbeb467b6fe254472207e3d1f Mon Sep 17 00:00:00 2001 +From: A U Thor +Date: Thu, 7 Apr 2005 15:20:13 -0700 +Subject: [PATCH 1/2] This is an excessively long subject line for a message + due to the habit some projects have of not having a short, one-line subject + at the start of the commit message, but rather sticking a whole paragraph + right at the start as the only thing in the commit message. It had better not + become the filename for the patch. +Header1: B E Cipient +To: Someone +Cc: C E Cipient + +-- +2.8.1.102.ga49ec4a.dirty + +EOF + +test_expect_success 'format-patch --no-patch suppresses patch' ' + + git format-patch --no-patch -2 && + cat 0001-This-is-an-excessively-long-subject-line-for-a-messa.patch > output && + test_cmp expect output + +' + test_expect_success 'format-patch from a subdirectory (1)' ' filename=$( rm -rf sub && -- 2.8.1.102.ga49ec4a ``` ## Eric Sunshine, 2016-04-07 21:30 Subject: Re: [PATCH] format-patch: allow --no-patch to disable patch output Message-ID: URL: https://gitlist.dev/e/CAPig%2BcRNCwFmGW-Zo%2BosjceA6nb%3D6t3%2BrJ2uEbnyYxGfSWQ-Uw%40mail.gmail.com In-Reply-To: <1460047599-9267-1-git-send-email-jacob.e.keller@intel.com> ``` On Thu, Apr 7, 2016 at 12:46 PM, Jacob Keller wrote: > The documentation for format-patch indicates that --no-patch wilL > suppress patch output. It also incorrectly mentions that -s will also > suppress the patch output, but this is incorrect because -s is used to > add the sign-off line first. Fix the documentation to remove the > indication about -s. Fix builtin format-patch to prevent always adding > the patch incase --no-patch is given. An alternative approach would be > to remove --no-patch from the documentation for git-format-patch, and > add an error check to indicate that --no-patch is not supported. See [1] and [2]. [1]: http://thread.gmane.org/gmane.comp.version-control.git/290023 [2]: b73a1bc (git-format-patch.txt: don't show -s as shorthand for multiple options, 2016-03-27) > Add a test for --no-patch to ensure it functions as expected. > > Signed-off-by: Jacob Keller ``` ## Jacob Keller, 2016-04-08 18:17 Subject: Re: [PATCH] format-patch: allow --no-patch to disable patch output Message-ID: URL: https://gitlist.dev/e/CA%2BP7%2BxoUi5H19zOCAiiwQZ0-j8yNZS9-2DzgoMmLSVOrC8Nbpw%40mail.gmail.com In-Reply-To: ``` On Thu, Apr 7, 2016 at 2:30 PM, Eric Sunshine wrote: > On Thu, Apr 7, 2016 at 12:46 PM, Jacob Keller wrote: >> The documentation for format-patch indicates that --no-patch wilL >> suppress patch output. It also incorrectly mentions that -s will also >> suppress the patch output, but this is incorrect because -s is used to >> add the sign-off line first. Fix the documentation to remove the >> indication about -s. Fix builtin format-patch to prevent always adding >> the patch incase --no-patch is given. An alternative approach would be >> to remove --no-patch from the documentation for git-format-patch, and >> add an error check to indicate that --no-patch is not supported. > > See [1] and [2]. > > [1]: http://thread.gmane.org/gmane.comp.version-control.git/290023 > [2]: b73a1bc (git-format-patch.txt: don't show -s as shorthand for > multiple options, 2016-03-27) > >> Add a test for --no-patch to ensure it functions as expected. >> >> Signed-off-by: Jacob Keller > -- > To unsubscribe from this list: send the line "unsubscribe git" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html So if I want to get the summary of a patch out of a commit without the actual diff what would you suggest? Basically, I have a process where we have post processed code, and I want to be able to generate a "patch" that shows the diff only of the post-processed code, ie: checkout commit^ to get the parent commit, process the code, then checkout the commit and process that code, and run a diff between, then spit out the equivalent of "git format-patch commit --no-patch" + the diff between the post processed code. I think maybe git show --pretty=email might work? Thanks, Jake ``` ## Junio C Hamano, 2016-04-08 18:30 Subject: Re: [PATCH] format-patch: allow --no-patch to disable patch output Message-ID: URL: https://gitlist.dev/e/xmqqh9fcvtqg.fsf%40gitster.mtv.corp.google.com In-Reply-To: ``` Jacob Keller writes: > So if I want to get the summary of a patch out of a commit without the > actual diff what would you suggest? "git log --stat"? If you have a set of (discrete) commits, "git show --stat A B C"? > Basically, I have a process where we have post processed code, and I > want to be able to generate a "patch" that shows the diff only of the > post-processed code, ie: If you want to script and depend on the exact output, you cannot depend on "log" or "show", so you would likely be doing "cat-file commit" for log message part and "diff-tree" (with options like -p and --stat) for the patch part, I guess. ``` ## Jacob Keller, 2016-04-08 18:34 Subject: Re: [PATCH] format-patch: allow --no-patch to disable patch output Message-ID: URL: https://gitlist.dev/e/CA%2BP7%2BxrYMqcohYpWUpgAu1K%3D0N-3TgQi0_1sQ%2Bz4v7nCPGcYSQ%40mail.gmail.com In-Reply-To: ``` On Fri, Apr 8, 2016 at 11:30 AM, Junio C Hamano wrote: > Jacob Keller writes: > >> So if I want to get the summary of a patch out of a commit without the >> actual diff what would you suggest? > > "git log --stat"? > > If you have a set of (discrete) commits, "git show --stat A B C"? > >> Basically, I have a process where we have post processed code, and I >> want to be able to generate a "patch" that shows the diff only of the >> post-processed code, ie: > > If you want to script and depend on the exact output, you cannot > depend on "log" or "show", so you would likely be doing "cat-file > commit" for log message part and "diff-tree" (with options like -p > and --stat) for the patch part, I guess. > The diff is being done on separate post-processed code, I can do that part. I just couldn't figure out how to get the same "header" as format-patch, but without the diff, which is when I saw no-patch, and then saw that it didn't work. Thanks, Jake ```