git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [PATCH] gitignore(5): Allow "foo/" in ignore list to match directory "foo"

From
Junio C Hamano <gitster@pobox.com>
Date
Jan 31, 2008, 21:51 UTC
Message-ID
<7vr6fxbr5a.fsf@gitster.siamese.dyndns.org>
In-Reply-To
<6bc632150801310356w1b2fa019n87d92986aed807c5@mail.gmail.com>
"pradeep singh rautela" <rautelap@gmail.com> writes:
Show 5 quoted lines
> How about something like,
>                   warning("Ignoring ignore entry because of trailing
> slash: %s\n Remove the trailing slash from the directory name to
> ignore it", string);
> May be this will help absolute git newbies.
I am afraid that this is leading us in the wrong direction.

What would be the first reaction if somebody sees such a message?

    The message implies that the user said "foo/" which would be
    ignored and the right substitution is "foo".  If that is the
    right substitution, why doesn't the stupid "git" program do
    that for the user automatically?!?!?!?!
See?

"Remove the trailing" suggestion assumes that we would want "foo/" and "foo" to mean the same thing.

Maybe we do, but we usually match both directory "foo/" and regular file "foo" when you say "foo", and we match only directory "foo/" when you say "foo/", as you saw in the ls-files example.

While I am not 100% convinced that we want to keep the distinction between these two forms, I am far from thinking that the existing distinction in other parts of the system is useless and should be removed.

Maybe we would want to drop this distinction in the gitignore entries, and the apparent inconsistency may not hurt in reality. If that is what we would want, that is fine, but then we shouldn't give a warning with a stupid piece of advice, but instead just do it ourselves.

Like this on top of 'master' (i.e. discarding all the previous patches), perhaps...

-- >8 -- [PATCH] gitignore(5): Allow "foo/" in ignore list to match directory "foo"

A pattern "foo/" in the exclude list did not match directory "foo", but a pattern "foo" did. This just strips the trailing slash from such input.

This makes the behaviour slightly inconsistent with that of pathspecs, where "foo/" only matches directory "foo" and not regular file "foo" and make "foo/" in the ignore list match regular file "foo" happily. This may hopefully does not matter in practice.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
---
 Documentation/gitignore.txt        |    3 +++
 dir.c                              |   22 ++++++++++++++++++----
 t/t3001-ls-files-others-exclude.sh |   26 ++++++++++++++++++++++++++
 3 files changed, 47 insertions(+), 4 deletions(-)
diff --git a/Documentation/gitignore.txt b/Documentation/gitignore.txt
index 08373f5..081a4df 100644
--- a/Documentation/gitignore.txt
+++ b/Documentation/gitignore.txt
@@ -57,6 +57,9 @@ Patterns have the following format:
    included again.  If a negated pattern matches, this will
    override lower precedence patterns sources.
 
+ - If the pattern ends with a slash, it is removed for the
+   purpose of the following description.
+
  - If the pattern does not contain a slash '/', git treats it as
    a shell glob pattern and checks for a match against the
    pathname without leading directories.
diff --git a/dir.c b/dir.c
index 3e345c2..fe51829 100644
--- a/dir.c
+++ b/dir.c
@@ -126,14 +126,28 @@ static int no_wildcard(const char *string)
 void add_exclude(const char *string, const char *base,
 		 int baselen, struct exclude_list *which)
 {
-	struct exclude *x = xmalloc(sizeof (*x));
+	struct exclude *x;
+	size_t len;
+	int to_exclude = 1;
 
-	x->to_exclude = 1;
 	if (*string == '!') {
-		x->to_exclude = 0;
+		to_exclude = 0;
 		string++;
 	}
-	x->pattern = string;
+	len = strlen(string);
+	if (len && string[len - 1] == '/') {
+		char *s;
+		x = xmalloc(sizeof(*x) + len);
+		s = (char*)(x+1);
+		memcpy(s, string, len - 1);
+		s[len - 1] = '\0';
+		string = s;
+		x->pattern = s;
+	} else {
+		x = xmalloc(sizeof(*x));
+		x->pattern = string;
+	}
+	x->to_exclude = to_exclude;
 	x->patternlen = strlen(string);
 	x->base = base;
 	x->baselen = baselen;
diff --git a/t/t3001-ls-files-others-exclude.sh b/t/t3001-ls-files-others-exclude.sh
index e25b255..5bc4885 100755
--- a/t/t3001-ls-files-others-exclude.sh
+++ b/t/t3001-ls-files-others-exclude.sh
@@ -99,4 +99,30 @@ EOF
 test_expect_success 'git-status honours core.excludesfile' \
 	'diff -u expect output'
 
+test_expect_success 'trailing slash in exclude allows directory match(1)' '
+
+	git ls-files --others --exclude=one/ >output &&
+	if grep "^one/" output
+	then
+		echo Ooops
+		false
+	else
+		: happy
+	fi
+
+'
+
+test_expect_success 'trailing slash in exclude allows directory match (2)' '
+
+	git ls-files --others --exclude=one/two/ >output &&
+	if grep "^one/two/" output
+	then
+		echo Ooops
+		false
+	else
+		: happy
+	fi
+
+'
+
 test_done
-- 
1.5.4.rc5.16.gc0279
Previous: pradeep singh rautelaNext: Adam Piatyszek
Message 16 of 20 in “Why does git track directory listed in .gitignore/".git/info/exclude"?”
  1. pradeep singh rautelaJan 23, 2008
  2. pradeep singh rautelaJan 23, 2008
  3. Linus TorvaldsJan 23, 2008
  4. pradeep singh rautelaJan 24, 2008
  5. Adam PiatyszekJan 30, 2008
  6. Junio C HamanoJan 30, 2008
  7. Junio C HamanoJan 30, 2008
  8. Adam PiatyszekJan 31, 2008
  9. *Re: Why does git track directory listed in .gitignore/".git/info/exclude"?Junio C Hamano, Jan 31, 2008
  10. gitignore(5): Allow "foo/" in ignore list to match directory "foo"Junio C Hamano, Jan 31, 2008
  11. Jeff KingJan 31, 2008
  12. Junio C HamanoJan 31, 2008
  13. Jeff KingJan 31, 2008
  14. Johannes SchindelinJan 31, 2008
  15. pradeep singh rautelaJan 31, 2008
  16. Junio C HamanoJan 31, 2008
  17. Adam PiatyszekJan 31, 2008
  18. Andreas EricssonFeb 1, 2008
  19. Adam PiatyszekJan 31, 2008
  20. Wayne DavisonJan 23, 2008

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.