[PATCH 1/3] builtin-status: submodule summary support

[PATCH 1/3] builtin-status: submodule summary support

[Ping Yin]

This commit teaches 'git commit/status' show 'Modified submodules'
section given by 'git submodule summary --cached --for-status'
just before 'Untracked files' section.

Signed-off-by: Ping Yin <pkufranky@gmail.com>
---
 wt-status.c |   31 +++++++++++++++++++++++++++++++
 1 files changed, 31 insertions(+), 0 deletions(-)

diff --git a/wt-status.c b/wt-status.c
index 701d13d..6d36e54 100644
--- a/wt-status.c
+++ b/wt-status.c
@@ -8,6 +8,7 @@
 #include "revision.h"
 #include "diffcore.h"
 #include "quote.h"
+#include "run-command.h"
 
 int wt_status_relative_paths = 1;
 int wt_status_use_color = -1;
@@ -219,6 +220,33 @@ static void wt_status_print_changed(struct wt_status *s)
 	rev.diffopt.format_callback_data = s;
 	run_diff_files(&rev, 0);
 }
+static void wt_status_print_submodule_summary(struct wt_status *s)
+{
+	struct child_process sm_summary;
+	const char *argv[] = {
+		"submodule",
+		"summary",
+		"--cached",
+		"--for-status",
+		"--summary-limit",
+		summary_limit,
+		s->amend ? "HEAD^" : "HEAD",
+		NULL
+	};
+	char index[PATH_MAX];
+	const char *env[2] = { index, NULL };
+	snprintf(index, sizeof(index), "GIT_INDEX_FILE=%s", s->index_file);
+
+	memset(&sm_summary, 0, sizeof(sm_summary));
+	sm_summary.argv = argv;
+	sm_summary.env = env;
+	sm_summary.git_cmd = 1;
+	sm_summary.no_stdin = 1;
+	fflush(s->fp);
+	sm_summary.out = dup(fileno(s->fp));    /* run_command closes it */
+	sm_summary.no_stderr = 1;
+	run_command(&sm_summary);
+}
 
 static void wt_status_print_untracked(struct wt_status *s)
 {
@@ -321,6 +349,9 @@ void wt_status_print(struct wt_status *s)
 	}
 
 	wt_status_print_changed(s);
+	// must flush s->fp since following call will write to s->fp in a child process
+	fflush(s->fp);
+	wt_status_print_submodule_summary(s);
 	wt_status_print_untracked(s);
 
 	if (s->verbose && !s->is_initial)
-- 
1.5.4.3.347.g5314c

[PATCH 2/3] builtin-status: configurable submodule summary size

[Ping Yin]

Add config variable status.submodulesummary which is passed as
arg for '--summary-limit' of 'git submodule summary' to limit the
submodule summary size.

status.submodulesummary is 0 by default which disables the summary.

Signed-off-by: Ping Yin <pkufranky@gmail.com>
---
 wt-status.c |    7 +++++++
 1 files changed, 7 insertions(+), 0 deletions(-)

diff --git a/wt-status.c b/wt-status.c
index 6d36e54..6a58057 100644
--- a/wt-status.c
+++ b/wt-status.c
@@ -12,6 +12,7 @@
 
 int wt_status_relative_paths = 1;
 int wt_status_use_color = -1;
+int wt_status_submodule_summary = 0;
 static char wt_status_colors[][COLOR_MAXLEN] = {
 	"",         /* WT_STATUS_HEADER: normal */
 	"\033[32m", /* WT_STATUS_UPDATED: green */
@@ -223,6 +224,8 @@ static void wt_status_print_changed(struct wt_status *s)
 static void wt_status_print_submodule_summary(struct wt_status *s)
 {
 	struct child_process sm_summary;
+	char summary_limit[64];
+	sprintf(summary_limit, "%d", wt_status_submodule_summary);
 	const char *argv[] = {
 		"submodule",
 		"summary",
@@ -374,6 +377,10 @@ void wt_status_print(struct wt_status *s)
 
 int git_status_config(const char *k, const char *v)
 {
+	if (!strcmp(k, "status.submodulesummary")) {
+		wt_status_submodule_summary = atoi(v);
+		return 0;
+	}
 	if (!strcmp(k, "status.color") || !strcmp(k, "color.status")) {
 		wt_status_use_color = git_config_colorbool(k, v, -1);
 		return 0;
-- 
1.5.4.3.347.g5314c

[PATCH 3/3] buitin-status: Add tests for submodule summary

[Ping Yin]

---
 t/t7502-status.sh |   76 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 files changed, 76 insertions(+), 0 deletions(-)

diff --git a/t/t7502-status.sh b/t/t7502-status.sh
index e006074..d48c0fe 100755
--- a/t/t7502-status.sh
+++ b/t/t7502-status.sh
@@ -149,4 +149,80 @@ test_expect_success 'status of partial commit excluding new file in index' '
 	diff -u expect output
 '
 
+test_expect_success "setup status submodule summary" '
+	test_create_repo sm &&
+	cd sm &&
+	: >foo &&
+	git add foo &&
+	git commit -m "Add foo" &&
+	cd .. &&
+	git add sm
+'
+
+cat > expect << \EOF
+# On branch master
+# Changes to be committed:
+#   (use "git reset HEAD <file>..." to unstage)
+#
+#	new file:   dir2/added
+#	new file:   sm
+#
+# Changed but not updated:
+#   (use "git add <file>..." to update what will be committed)
+#
+#	modified:   dir1/modified
+#
+# Untracked files:
+#   (use "git add <file>..." to include in what will be committed)
+#
+#	dir1/untracked
+#	dir2/modified
+#	dir2/untracked
+#	expect
+#	output
+#	untracked
+EOF
+test_expect_success "status submodule summary is disabled by default" '
+	git status > output &&
+	git diff expect output
+'
+
+cd sm &&
+head=$(git rev-parse --verify HEAD | cut -c1-7) &&
+cd ..
+
+cat > expect <<-EOF
+# On branch master
+# Changes to be committed:
+#   (use "git reset HEAD <file>..." to unstage)
+#
+#	new file:   dir2/added
+#	new file:   sm
+#
+# Changed but not updated:
+#   (use "git add <file>..." to update what will be committed)
+#
+#	modified:   dir1/modified
+#
+# Modified submodules:
+#
+# * sm 0000000...$head (1):
+#   > Add foo
+#
+# Untracked files:
+#   (use "git add <file>..." to include in what will be committed)
+#
+#	dir1/untracked
+#	dir2/modified
+#	dir2/untracked
+#	expect
+#	output
+#	untracked
+EOF
+test_expect_success "status submodule summary" '
+	git config status.submodulesummary 10 &&
+	git status > output &&
+	git diff expect output
+'
+
 test_done
-- 
1.5.4.3.347.g5314c

Re: [PATCH 1/3] builtin-status: submodule summary support

[Johannes Sixt]

Ping Yin schrieb:
> +static void wt_status_print_submodule_summary(struct wt_status *s)
> +{
> +	struct child_process sm_summary;
> +	const char *argv[] = {
> +		"submodule",
> +		"summary",
> +		"--cached",
> +		"--for-status",
> +		"--summary-limit",
> +		summary_limit,

Where is summary_limit? Did you split this patch series incorrectly?

> +		s->amend ? "HEAD^" : "HEAD",

Any chance that we avoid "HEAD^" here? Perhaps we have the SHA1 around
somewhere? I fear that our MSYS bash mangles it into "HEAD". I shall test it.

BTW, you don't mention the prerequisites of this series. I assume it
builds on top of your "[PATCH v5 0/5] git-submodule summary" series.

> +	sm_summary.no_stderr = 1;

Why this? If the submodule summary has errors we certainly want to see them.

> @@ -321,6 +349,9 @@ void wt_status_print(struct wt_status *s)
>  	}
>  
>  	wt_status_print_changed(s);
> +	// must flush s->fp since following call will write to s->fp in a child process
> +	fflush(s->fp);
> +	wt_status_print_submodule_summary(s);

Hmm. Aren't you unconditionally spawning "git submodule summary" for each
git-status/git-commit?

-- Hannes

Re: [PATCH 3/3] buitin-status: Add tests for submodule summary

[Johannes Sixt]

Ping Yin schrieb:
> ---
>  t/t7502-status.sh |   76 +++++++++++++++++++++++++++++++++++++++++++++++++++++
>  1 files changed, 76 insertions(+), 0 deletions(-)

Signed-off-by?

Would you please add a test case that uses "git status --amend"? It would
help diagnose failures with regard to "HEAD^" on Windows greatly.

-- Hannes

Re: [PATCH 1/3] builtin-status: submodule summary support

[Ping Yin]

>
>  Where is summary_limit? Did you split this patch series incorrectly?
>
>  > +             s->amend ? "HEAD^" : "HEAD",
>
>  Any chance that we avoid "HEAD^" here? Perhaps we have the SHA1 around
>  somewhere? I fear that our MSYS bash mangles it into "HEAD". I shall test it.
>
I am not familiar with git core code, and don't find a way to avoid 'HEAD^' yet



-- 
Ping Yin