Skip to contentSkip to conversation

status: add an empty line when there is no hint

PATCH In review John Lin 4 participants 16 messages Apr 30, 2019 – May 22, 2019

lore
Share & tools
lore

John Lin

author
#0 lore raw

PATCH v2 1/1 status: add an empty line when there is no hint

When typing "git status", there is an empty line between the "Changes not staged for commit:" block and the list of changed files. However, when typing "git commit" with no files added, there are no empty lines between them.

This patch adds empty lines in the above case and some similar cases.

---
 t/t7500-commit-template-squash-signoff.sh |  1 +
 t/t7508-status.sh                         |  5 +++++
 t/t7512-status-help.sh                    |  1 +
 wt-status.c                               | 12 ++++++++----
 4 files changed, 15 insertions(+), 4 deletions(-)
t/t7500-commit-template-squash-signoff.sh +1 -0
index 46a5cd4b73..0423e77d1d 100755
@@ -345,6 +345,7 @@ cat >expected-template <<EOF
 #
 # On branch commit-template-check
 # Changes to be committed:
+#
 #	new file:   commit-template-check
 #
 # Untracked files not listed
t/t7508-status.sh +5 -0
index e1f11293e2..949b1dbcc4 100755
@@ -204,12 +204,15 @@ Your branch and 'upstream' have diverged,
 and have 1 and 2 different commits each, respectively.
 
 Changes to be committed:
+
 	new file:   dir2/added
 
 Changes not staged for commit:
+
 	modified:   dir1/modified
 
 Untracked files:
+
 	dir1/untracked
 	dir2/modified
 	dir2/untracked
@@ -449,9 +452,11 @@ Your branch and '\''upstream'\'' have diverged,
 and have 1 and 2 different commits each, respectively.
 
 Changes to be committed:
+
 	new file:   dir2/added
 
 Changes not staged for commit:
+
 	modified:   dir1/modified
 
 Untracked files not listed
t/t7512-status-help.sh +1 -0
index 458608cc1e..0a29fa66a2 100755
@@ -714,6 +714,7 @@ rebase in progress; onto $ONTO
 You are currently rebasing branch '\''statushints_disabled'\'' on '\''$ONTO'\''.
 
 Unmerged paths:
+
 	both modified:   main.txt
 
 no changes added to commit
wt-status.c +8 -4
index 445a36204a..0766e3ee12 100644
@@ -175,7 +175,7 @@ static void wt_longstatus_print_unmerged_header(struct wt_status *s)
 	}
 
 	if (!s->hints)
-		return;
+		goto conclude;
 	if (s->whence != FROM_COMMIT)
 		;
 	else if (!s->is_initial)
@@ -193,6 +193,7 @@ static void wt_longstatus_print_unmerged_header(struct wt_status *s)
 	} else {
 		status_printf_ln(s, c, _("  (use \"git add/rm <file>...\" as appropriate to mark resolution)"));
 	}
+conclude:
 	status_printf_ln(s, c, "%s", "");
 }
 
@@ -202,13 +203,14 @@ static void wt_longstatus_print_cached_header(struct wt_status *s)
 
 	status_printf_ln(s, c, _("Changes to be committed:"));
 	if (!s->hints)
-		return;
+		goto conclude;
 	if (s->whence != FROM_COMMIT)
 		; /* NEEDSWORK: use "git reset --unresolve"??? */
 	else if (!s->is_initial)
 		status_printf_ln(s, c, _("  (use \"git reset %s <file>...\" to unstage)"), s->reference);
 	else
 		status_printf_ln(s, c, _("  (use \"git rm --cached <file>...\" to unstage)"));
+conclude:
 	status_printf_ln(s, c, "%s", "");
 }
 
@@ -220,7 +222,7 @@ static void wt_longstatus_print_dirty_header(struct wt_status *s,
 
 	status_printf_ln(s, c, _("Changes not staged for commit:"));
 	if (!s->hints)
-		return;
+		goto conclude;
 	if (!has_deleted)
 		status_printf_ln(s, c, _("  (use \"git add <file>...\" to update what will be committed)"));
 	else
@@ -228,6 +230,7 @@ static void wt_longstatus_print_dirty_header(struct wt_status *s,
 	status_printf_ln(s, c, _("  (use \"git checkout -- <file>...\" to discard changes in working directory)"));
 	if (has_dirty_submodules)
 		status_printf_ln(s, c, _("  (commit or discard the untracked or modified content in submodules)"));
+conclude:
 	status_printf_ln(s, c, "%s", "");
 }
 
@@ -238,8 +241,9 @@ static void wt_longstatus_print_other_header(struct wt_status *s,
 	const char *c = color(WT_STATUS_HEADER, s);
 	status_printf_ln(s, c, "%s:", what);
 	if (!s->hints)
-		return;
+		goto conclude;
 	status_printf_ln(s, c, _("  (use \"git %s <file>...\" to include in what will be committed)"), how);
+conclude:
 	status_printf_ln(s, c, "%s", "");
 }
 
Signature · 1 line
2.21.0
#1 lore raw

Hi John

Quoted 3 lines · John Lin

When typing "git status", there is an empty line between the "Changes not staged for commit:" block and the list of changed files.

I'm a bit confused by this as you change a status test below by inserting these blank lines into the expected output, implying they are not there now. I think maybe the blank line is only shown when status prints advice.

Quoted 2 lines

However, when typing "git commit" with
no files added, there are no empty lines between them.

I have to say looking at the changes to the output I prefer the original, the lists are nicely indented so there is no need for a blank line to separate the header from the list and having the header immediately before the list means the blank line at the end of the block makes the extent of the block clear. It also saves screen space which is useful for small laptop screens. I can see why one might want a blank line to separate the advice and list of changes (though even there the indention of the list and advice is different) but for a one line header I think it is better to start the list on the next line.

Best Wishes

Phillip

Quoted 133 lines

This patch adds empty lines in the above case and some similar cases.

---
  t/t7500-commit-template-squash-signoff.sh |  1 +
  t/t7508-status.sh                         |  5 +++++
  t/t7512-status-help.sh                    |  1 +
  wt-status.c                               | 12 ++++++++----
  4 files changed, 15 insertions(+), 4 deletions(-)
t/t7500-commit-template-squash-signoff.sh +1 -0
index 46a5cd4b73..0423e77d1d 100755
@@ -345,6 +345,7 @@ cat >expected-template <<EOF
  #
  # On branch commit-template-check
  # Changes to be committed:
+#
  #	new file:   commit-template-check
  #
  # Untracked files not listed
t/t7508-status.sh +5 -0
index e1f11293e2..949b1dbcc4 100755
@@ -204,12 +204,15 @@ Your branch and 'upstream' have diverged,
  and have 1 and 2 different commits each, respectively.
 
  Changes to be committed:
+
  	new file:   dir2/added
 
  Changes not staged for commit:
+
  	modified:   dir1/modified
 
  Untracked files:
+
  	dir1/untracked
  	dir2/modified
  	dir2/untracked
@@ -449,9 +452,11 @@ Your branch and '\''upstream'\'' have diverged,
  and have 1 and 2 different commits each, respectively.
 
  Changes to be committed:
+
  	new file:   dir2/added
 
  Changes not staged for commit:
+
  	modified:   dir1/modified
 
  Untracked files not listed
t/t7512-status-help.sh +1 -0
index 458608cc1e..0a29fa66a2 100755
@@ -714,6 +714,7 @@ rebase in progress; onto $ONTO
  You are currently rebasing branch '\''statushints_disabled'\'' on '\''$ONTO'\''.
 
  Unmerged paths:
+
  	both modified:   main.txt
 
  no changes added to commit
wt-status.c +8 -4
index 445a36204a..0766e3ee12 100644
@@ -175,7 +175,7 @@ static void wt_longstatus_print_unmerged_header(struct wt_status *s)
  	}
 
  	if (!s->hints)
-		return;
+		goto conclude;
  	if (s->whence != FROM_COMMIT)
  		;
  	else if (!s->is_initial)
@@ -193,6 +193,7 @@ static void wt_longstatus_print_unmerged_header(struct wt_status *s)
  	} else {
  		status_printf_ln(s, c, _("  (use \"git add/rm <file>...\" as appropriate to mark resolution)"));
  	}
+conclude:
  	status_printf_ln(s, c, "%s", "");
  }
 
@@ -202,13 +203,14 @@ static void wt_longstatus_print_cached_header(struct wt_status *s)
 
  	status_printf_ln(s, c, _("Changes to be committed:"));
  	if (!s->hints)
-		return;
+		goto conclude;
  	if (s->whence != FROM_COMMIT)
  		; /* NEEDSWORK: use "git reset --unresolve"??? */
  	else if (!s->is_initial)
  		status_printf_ln(s, c, _("  (use \"git reset %s <file>...\" to unstage)"), s->reference);
  	else
  		status_printf_ln(s, c, _("  (use \"git rm --cached <file>...\" to unstage)"));
+conclude:
  	status_printf_ln(s, c, "%s", "");
  }
 
@@ -220,7 +222,7 @@ static void wt_longstatus_print_dirty_header(struct wt_status *s,
 
  	status_printf_ln(s, c, _("Changes not staged for commit:"));
  	if (!s->hints)
-		return;
+		goto conclude;
  	if (!has_deleted)
  		status_printf_ln(s, c, _("  (use \"git add <file>...\" to update what will be committed)"));
  	else
@@ -228,6 +230,7 @@ static void wt_longstatus_print_dirty_header(struct wt_status *s,
  	status_printf_ln(s, c, _("  (use \"git checkout -- <file>...\" to discard changes in working directory)"));
  	if (has_dirty_submodules)
  		status_printf_ln(s, c, _("  (commit or discard the untracked or modified content in submodules)"));
+conclude:
  	status_printf_ln(s, c, "%s", "");
  }
 
@@ -238,8 +241,9 @@ static void wt_longstatus_print_other_header(struct wt_status *s,
  	const char *c = color(WT_STATUS_HEADER, s);
  	status_printf_ln(s, c, "%s:", what);
  	if (!s->hints)
-		return;
+		goto conclude;
  	status_printf_ln(s, c, _("  (use \"git %s <file>...\" to include in what will be committed)"), how);
+conclude:
  	status_printf_ln(s, c, "%s", "");
  }
 
author
#2 lore raw

Hi Phillip,

Phillip Wood <[email protected]> 於 2019年4月30日 週二 下午7:15寫道:

Quoted 16 lines

Hi John

Quoted 3 lines · John Lin

When typing "git status", there is an empty line between the "Changes not staged for commit:" block and the list of changed files.

I'm a bit confused by this as you change a status test below by inserting these blank lines into the expected output, implying they are not there now. I think maybe the blank line is only shown when status prints advice.

Quoted 2 lines

However, when typing "git commit" with
no files added, there are no empty lines between them.

Quoted 9 lines

I have to say looking at the changes to the output I prefer the original, the lists are nicely indented so there is no need for a blank line to separate the header from the list and having the header immediately before the list means the blank line at the end of the block makes the extent of the block clear. It also saves screen space which is useful for small laptop screens. I can see why one might want a blank line to separate the advice and list of changes (though even there the indention of the list and advice is different) but for a one line header I think it is better to start the list on the next line.

Thanks for the review. Your opinion does make sense to me. Please consider this patch dropped. Thank you.

Best,
John Lin

Quoted 138 lines

Best Wishes

Phillip

Quoted 133 lines

This patch adds empty lines in the above case and some similar cases.

---
  t/t7500-commit-template-squash-signoff.sh |  1 +
  t/t7508-status.sh                         |  5 +++++
  t/t7512-status-help.sh                    |  1 +
  wt-status.c                               | 12 ++++++++----
  4 files changed, 15 insertions(+), 4 deletions(-)
t/t7500-commit-template-squash-signoff.sh +1 -0
index 46a5cd4b73..0423e77d1d 100755
@@ -345,6 +345,7 @@ cat >expected-template <<EOF
  #
  # On branch commit-template-check
  # Changes to be committed:
+#
  #   new file:   commit-template-check
  #
  # Untracked files not listed
t/t7508-status.sh +5 -0
index e1f11293e2..949b1dbcc4 100755
@@ -204,12 +204,15 @@ Your branch and 'upstream' have diverged,
  and have 1 and 2 different commits each, respectively.
 
  Changes to be committed:
+
      new file:   dir2/added
 
  Changes not staged for commit:
+
      modified:   dir1/modified
 
  Untracked files:
+
      dir1/untracked
      dir2/modified
      dir2/untracked
@@ -449,9 +452,11 @@ Your branch and '\''upstream'\'' have diverged,
  and have 1 and 2 different commits each, respectively.
 
  Changes to be committed:
+
      new file:   dir2/added
 
  Changes not staged for commit:
+
      modified:   dir1/modified
 
  Untracked files not listed
t/t7512-status-help.sh +1 -0
index 458608cc1e..0a29fa66a2 100755
@@ -714,6 +714,7 @@ rebase in progress; onto $ONTO
  You are currently rebasing branch '\''statushints_disabled'\'' on '\''$ONTO'\''.
 
  Unmerged paths:
+
      both modified:   main.txt
 
  no changes added to commit
wt-status.c +8 -4
index 445a36204a..0766e3ee12 100644
@@ -175,7 +175,7 @@ static void wt_longstatus_print_unmerged_header(struct wt_status *s)
      }
 
      if (!s->hints)
-             return;
+             goto conclude;
      if (s->whence != FROM_COMMIT)
              ;
      else if (!s->is_initial)
@@ -193,6 +193,7 @@ static void wt_longstatus_print_unmerged_header(struct wt_status *s)
      } else {
              status_printf_ln(s, c, _("  (use \"git add/rm <file>...\" as appropriate to mark resolution)"));
      }
+conclude:
      status_printf_ln(s, c, "%s", "");
  }
 
@@ -202,13 +203,14 @@ static void wt_longstatus_print_cached_header(struct wt_status *s)
 
      status_printf_ln(s, c, _("Changes to be committed:"));
      if (!s->hints)
-             return;
+             goto conclude;
      if (s->whence != FROM_COMMIT)
              ; /* NEEDSWORK: use "git reset --unresolve"??? */
      else if (!s->is_initial)
              status_printf_ln(s, c, _("  (use \"git reset %s <file>...\" to unstage)"), s->reference);
      else
              status_printf_ln(s, c, _("  (use \"git rm --cached <file>...\" to unstage)"));
+conclude:
      status_printf_ln(s, c, "%s", "");
  }
 
@@ -220,7 +222,7 @@ static void wt_longstatus_print_dirty_header(struct wt_status *s,
 
      status_printf_ln(s, c, _("Changes not staged for commit:"));
      if (!s->hints)
-             return;
+             goto conclude;
      if (!has_deleted)
              status_printf_ln(s, c, _("  (use \"git add <file>...\" to update what will be committed)"));
      else
@@ -228,6 +230,7 @@ static void wt_longstatus_print_dirty_header(struct wt_status *s,
      status_printf_ln(s, c, _("  (use \"git checkout -- <file>...\" to discard changes in working directory)"));
      if (has_dirty_submodules)
              status_printf_ln(s, c, _("  (commit or discard the untracked or modified content in submodules)"));
+conclude:
      status_printf_ln(s, c, "%s", "");
  }
 
@@ -238,8 +241,9 @@ static void wt_longstatus_print_other_header(struct wt_status *s,
      const char *c = color(WT_STATUS_HEADER, s);
      status_printf_ln(s, c, "%s:", what);
      if (!s->hints)
-             return;
+             goto conclude;
      status_printf_ln(s, c, _("  (use \"git %s <file>...\" to include in what will be committed)"), how);
+conclude:
      status_printf_ln(s, c, "%s", "");
  }
 
#3 lore raw
In reply to Phillip Wood
Quoted 14 lines · Phillip Wood

Hi John

Quoted 3 lines · John Lin

When typing "git status", there is an empty line between the "Changes not staged for commit:" block and the list of changed files.

I'm a bit confused by this as you change a status test below by inserting these blank lines into the expected output, implying they are not there now. I think maybe the blank line is only shown when status prints advice.

Quoted 2 lines

However, when typing "git commit" with
no files added, there are no empty lines between them.

Quoted 9 lines

I have to say looking at the changes to the output I prefer the original, the lists are nicely indented so there is no need for a blank line to separate the header from the list and having the header immediately before the list means the blank line at the end of the block makes the extent of the block clear. It also saves screen space which is useful for small laptop screens. I can see why one might want a blank line to separate the advice and list of changes (though even there the indention of the list and advice is different) but for a one line header I think it is better to start the list on the next line.

I actually was going to submit an equivalent patch eventually. The inconsistency between status and commit is bothersome to me and I think that adding the whitespace improves readability.

I'd love to see this picked up.

Signature · 2 lines
brian m. carlson: Houston, Texas, US
OpenPGP: https://keybase.io/bk2204
author
#4 lore raw

Hi Brian,

brian m. carlson <[email protected]> 於 2019年5月2日 週四 上午7:45寫道:

Quoted 26 lines · Phillip Wood
Quoted 14 lines

Hi John

Quoted 3 lines · John Lin

When typing "git status", there is an empty line between the "Changes not staged for commit:" block and the list of changed files.

I'm a bit confused by this as you change a status test below by inserting these blank lines into the expected output, implying they are not there now. I think maybe the blank line is only shown when status prints advice.

Quoted 2 lines

However, when typing "git commit" with
no files added, there are no empty lines between them.

Quoted 9 lines

I have to say looking at the changes to the output I prefer the original, the lists are nicely indented so there is no need for a blank line to separate the header from the list and having the header immediately before the list means the blank line at the end of the block makes the extent of the block clear. It also saves screen space which is useful for small laptop screens. I can see why one might want a blank line to separate the advice and list of changes (though even there the indention of the list and advice is different) but for a one line header I think it is better to start the list on the next line.

Quoted 5 lines

I actually was going to submit an equivalent patch eventually. The inconsistency between status and commit is bothersome to me and I think that adding the whitespace improves readability.

I'd love to see this picked up.

Thank you for the support! I'll be very happy if my patch can get into the code base. Is there any improvement I can make for this patch?

Best,
John Lin

Quoted 3 lines

--
brian m. carlson: Houston, Texas, US
OpenPGP: https://keybase.io/bk2204

#5 lore raw
Quoted 2 lines · 林自均

Thank you for the support! I'll be very happy if my patch can get into the code base. Is there any improvement I can make for this patch?

The patch seems sane to me. There's less duplicate code than the patch I started writing; the goto conclude seems to avoid most of it.

Signature · 2 lines
brian m. carlson: Houston, Texas, US
OpenPGP: https://keybase.io/bk2204
author
#6 lore raw

Hi Brian,

brian m. carlson <[email protected]> 於 2019年5月3日 週五 上午7:15寫道:

Quoted 7 lines · 林自均
Quoted 2 lines

Thank you for the support! I'll be very happy if my patch can get into the code base. Is there any improvement I can make for this patch?

The patch seems sane to me. There's less duplicate code than the patch I started writing; the goto conclude seems to avoid most of it.

I see. Let's wait for more comments on my patch. Thank you.

Best,
John Lin

Quoted 3 lines

--
brian m. carlson: Houston, Texas, US
OpenPGP: https://keybase.io/bk2204

author
#7 lore raw

Hi Junio,

林自均 <[email protected]> 於 2019年5月3日 週五 下午12:15寫道:

Quoted 9 lines

Hi Brian,

brian m. carlson <[email protected]> 於 2019年5月3日 週五 上午7:15寫道:

Quoted 5 lines · 林自均
Quoted 2 lines

Thank you for the support! I'll be very happy if my patch can get into the code base. Is there any improvement I can make for this patch?

Quoted 7 lines
Quoted 2 lines

The patch seems sane to me. There's less duplicate code than the patch I started writing; the goto conclude seems to avoid most of it.

I see. Let's wait for more comments on my patch. Thank you.

Best,
John Lin

Could you please give me some advice on my patch? Thank you.

Best,
John Lin

Quoted 5 lines
Quoted 5 lines

--
brian m. carlson: Houston, Texas, US
OpenPGP: https://keybase.io/bk2204

#8 lore raw
Quoted 3 lines · 林自均

Hi Junio,

Quoted 1 line

...

Could you please give me some advice on my patch? Thank you.

I tend to agree with what Phillip said in
<[email protected]>.

If the difference between "status" and "commit" bothers you so much, i.e.

When typing "git status", there is an empty line between the "Changes not staged for commit:" block and the list of changed files. However, when typing "git commit" with no files added, there are no empty lines between them.

it may not be a bad idea to try making them consistent by removing the blank line that is given after the advice messages, perhaps?

author
#9 lore raw

Hi Junio,

Junio C Hamano <[email protected]> 於 2019年5月13日 週一 下午1:51寫道:

Quoted 13 lines
Quoted 3 lines · 林自均

Hi Junio,

Quoted 1 line

...

Could you please give me some advice on my patch? Thank you.

I tend to agree with what Phillip said in
<[email protected]>.

If the difference between "status" and "commit" bothers you so much, i.e.

Quoted 7 lines

When typing "git status", there is an empty line between the "Changes not staged for commit:" block and the list of changed files. However, when typing "git commit" with no files added, there are no empty lines between them.

it may not be a bad idea to try making them consistent by removing the blank line that is given after the advice messages, perhaps?

Thank you for the review. I guess removing the blank line also work for me. I'll submit another patch.

Best,
John Lin

#10 lore raw
In reply to Junio C Hamano
Quoted 10 lines · Junio C Hamano

If the difference between "status" and "commit" bothers you so much, i.e.

When typing "git status", there is an empty line between the "Changes not staged for commit:" block and the list of changed files. However, when typing "git commit" with no files added, there are no empty lines between them.

it may not be a bad idea to try making them consistent by removing the blank line that is given after the advice messages, perhaps?

I personally think the extra blank line aids readability, especially on screens with small text, but I'll defer to your decision on this.

Signature · 2 lines
brian m. carlson: Houston, Texas, US
OpenPGP: https://keybase.io/bk2204
#11 lore raw
Quoted 9 lines · brian m. carlson
Quoted 8 lines · Junio C Hamano

If the difference between "status" and "commit" bothers you so much, i.e.

When typing "git status", there is an empty line between the "Changes not staged for commit:" block and the list of changed files. However, when typing "git commit" with no files added, there are no empty lines between them.

Quoted 5 lines
Quoted 2 lines

it may not be a bad idea to try making them consistent by removing the blank line that is given after the advice messages, perhaps?

I personally think the extra blank line aids readability, especially on screens with small text, but I'll defer to your decision on this.

Heh, now we established that this is primarily of personal tastes, if you leave it up to me, my preference would be different from what was said in the thread so far.

- I do not mind having an extra blank line in the log message editor session "git commit" gives me, primarily because at that point I am in a full-screen editor that I can scroll up and down at ease.

- "git status" output, on the other hand, is shown in the context where vertical screen real estate is more precious (I do not say "git -p status"); I'd probably be happier without these empty lines.

But following the above two would mean that the result will still leave difference between the commands; the original justificaiton will not apply to such a change.

At the same time, I think I've been happy enough with the current output from both commands, so if you let me bikeshed freely, I'd probably pick "let's not change anything then" ;-)

author
#12 lore raw

Hi Junio,

Junio C Hamano <[email protected]> 於 2019年5月14日 週二 下午3:33寫道:

Quoted 18 lines
Quoted 9 lines · brian m. carlson
Quoted 8 lines · Junio C Hamano

If the difference between "status" and "commit" bothers you so much, i.e.

When typing "git status", there is an empty line between the "Changes not staged for commit:" block and the list of changed files. However, when typing "git commit" with no files added, there are no empty lines between them.

Quoted 5 lines
Quoted 2 lines

it may not be a bad idea to try making them consistent by removing the blank line that is given after the advice messages, perhaps?

I personally think the extra blank line aids readability, especially on screens with small text, but I'll defer to your decision on this.

Quoted 8 lines

Heh, now we established that this is primarily of personal tastes, if you leave it up to me, my preference would be different from what was said in the thread so far.

- I do not mind having an extra blank line in the log message editor session "git commit" gives me, primarily because at that point I am in a full-screen editor that I can scroll up and down at ease.

I was not talking about the messages in the editor session. I was talking about "git commit" without "git add" anything.

For example:

```
$ touch newfile.txt
$ git commit
On branch master
Untracked files:
        newfile.txt

nothing added to commit but untracked files present
```

My current patch is trying to add an empty line between "Untracked files:" and "newfile.txt".

Quoted 9 lines

- "git status" output, on the other hand, is shown in the context where vertical screen real estate is more precious (I do not say "git -p status"); I'd probably be happier without these empty lines.

But following the above two would mean that the result will still leave difference between the commands; the original justificaiton will not apply to such a change.

With my above clarification, I guess your two use cases will be the
same: the messages will appear on the console screen. That leads to a conclusion: we should save vertical screen space. I am working on a new patch to achieve that.

Quoted 4 lines

At the same time, I think I've been happy enough with the current output from both commands, so if you let me bikeshed freely, I'd probably pick "let's not change anything then" ;-)

Best,
John Lin

#13 lore raw
Quoted 15 lines · 林自均

I was not talking about the messages in the editor session. I was talking about "git commit" without "git add" anything.

For example:

```
$ touch newfile.txt
$ git commit
On branch master
Untracked files:
        newfile.txt

nothing added to commit but untracked files present
```
Quoted 2 lines

My current patch is trying to add an empty line between "Untracked files:" and "newfile.txt".

I do not think that one is paged, so if you ask me, I'd say we shouldn't add an extra blank there. Is that message also reused in the editor session, or do two different codepaths produce a similar looking message, one for the above case direct to the terminal and the other for the editor session?

But again...

Quoted 3 lines
Quoted 3 lines

At the same time, I think I've been happy enough with the current output from both commands, so if you let me bikeshed freely, I'd probably pick "let's not change anything then" ;-)

author
#14 lore raw

Hi Junio,

Junio C Hamano <[email protected]> 於 2019年5月15日 週三 上午8:48寫道:

Quoted 18 lines
Quoted 15 lines · 林自均

I was not talking about the messages in the editor session. I was talking about "git commit" without "git add" anything.

For example:

```
$ touch newfile.txt
$ git commit
On branch master
Untracked files:
        newfile.txt

nothing added to commit but untracked files present
```
Quoted 8 lines
Quoted 2 lines

My current patch is trying to add an empty line between "Untracked files:" and "newfile.txt".

I do not think that one is paged, so if you ask me, I'd say we shouldn't add an extra blank there. Is that message also reused in the editor session, or do two different codepaths produce a similar looking message, one for the above case direct to the terminal and the other for the editor session?

The messages produced in wt-status.c seem to be reused in both terminal and editor session. When I tried to modify the messages in terminal, the ones in editor session will also be modified accordingly.

By the way, my new patch to remove extra blank line is here:
https://github.com/gitgitgadget/git/pull/196
I am still waiting for someone to comment "/allow johnlinp".

Quoted 7 lines

But again...

Quoted 3 lines
Quoted 3 lines

At the same time, I think I've been happy enough with the current output from both commands, so if you let me bikeshed freely, I'd probably pick "let's not change anything then" ;-)

author
#15 lore raw

Hi Junio,

林自均 <[email protected]> 於 2019年5月15日 週三 上午11:02寫道:

Quoted 31 lines

Hi Junio,

Junio C Hamano <[email protected]> 於 2019年5月15日 週三 上午8:48寫道:

Quoted 18 lines
Quoted 15 lines · 林自均

I was not talking about the messages in the editor session. I was talking about "git commit" without "git add" anything.

For example:

```
$ touch newfile.txt
$ git commit
On branch master
Untracked files:
        newfile.txt

nothing added to commit but untracked files present
```
Quoted 8 lines
Quoted 2 lines

My current patch is trying to add an empty line between "Untracked files:" and "newfile.txt".

I do not think that one is paged, so if you ask me, I'd say we shouldn't add an extra blank there. Is that message also reused in the editor session, or do two different codepaths produce a similar looking message, one for the above case direct to the terminal and the other for the editor session?

Quoted 8 lines

The messages produced in wt-status.c seem to be reused in both terminal and editor session. When I tried to modify the messages in terminal, the ones in editor session will also be modified accordingly.

By the way, my new patch to remove extra blank line is here:
https://github.com/gitgitgadget/git/pull/196
I am still waiting for someone to comment "/allow johnlinp".

I've submitted the patch with the title [PATCH 1/1] status: remove the empty line after hints Please review if you are available. Thank you.

Best,
John Lin

Quoted 8 lines
Quoted 8 lines

But again...

Quoted 3 lines
Quoted 3 lines

At the same time, I think I've been happy enough with the current output from both commands, so if you let me bikeshed freely, I'd probably pick "let's not change anything then" ;-)