Thread (17 messages) 17 messages, 2 authors, 2023-12-12

Re: [PATCH 1/7] config: handle NULL value when parsing non-bools

From: Patrick Steinhardt <hidden>
Date: 2023-12-07 08:14:46

On Thu, Dec 07, 2023 at 02:11:14AM -0500, Jeff King wrote:
quoted hunk ↗ jump to hunk
When the config parser sees an "implicit" bool like:

  [core]
  someVariable

it passes NULL to the config callback. Any callback code which expects a
string must check for NULL. This usually happens via helpers like
git_config_string(), etc, but some custom code forgets to do so and will
segfault.

These are all fairly vanilla cases where the solution is just the usual
pattern of:

  if (!value)
        return config_error_nonbool(var);

though note that in a few cases we have to split initializers like:

  int some_var = initializer();

into:

  int some_var;
  if (!value)
        return config_error_nonbool(var);
  some_var = initializer();

There are still some broken instances after this patch, which I'll
address on their own in individual patches after this one.

Reported-by: Carlos Andrés Ramírez Cataño <redacted>
Signed-off-by: Jeff King <redacted>
---
 builtin/blame.c        |  2 ++
 builtin/checkout.c     |  2 ++
 builtin/clone.c        |  2 ++
 builtin/log.c          |  5 ++++-
 builtin/pack-objects.c |  6 +++++-
 compat/mingw.c         |  2 ++
 config.c               |  8 ++++++++
 diff.c                 | 19 ++++++++++++++++---
 mailinfo.c             |  2 ++
 notes-utils.c          |  2 ++
 trailer.c              |  2 ++
 11 files changed, 47 insertions(+), 5 deletions(-)
diff --git a/builtin/blame.c b/builtin/blame.c
index 9c987d6567..2433b7da5c 100644
--- a/builtin/blame.c
+++ b/builtin/blame.c
@@ -748,6 +748,8 @@ static int git_blame_config(const char *var, const char *value,
 	}
 
 	if (!strcmp(var, "blame.coloring")) {
+		if (!value)
+			return config_error_nonbool(var);
In the `else` statement where we fail to parse the value we only
generate a warning and return successfully regardless. Should we do the
same here?
quoted hunk ↗ jump to hunk
 		if (!strcmp(value, "repeatedLines")) {
 			coloring_mode |= OUTPUT_COLOR_LINE;
 		} else if (!strcmp(value, "highlightRecent")) {
diff --git a/builtin/checkout.c b/builtin/checkout.c
index f02434bc15..d5c784854f 100644
--- a/builtin/checkout.c
+++ b/builtin/checkout.c
@@ -1202,6 +1202,8 @@ static int git_checkout_config(const char *var, const char *value,
 	struct checkout_opts *opts = cb;
 
 	if (!strcmp(var, "diff.ignoresubmodules")) {
+		if (!value)
+			return config_error_nonbool(var);
 		handle_ignore_submodules_arg(&opts->diff_options, value);
 		return 0;
 	}
This one is fine, `handle_ignore_submodules_arg()` dies if it gets an
unknown value and `git_config()` will die when the callback function
returns an error. The same is true for many other cases you've
converted.

[snip]
quoted hunk ↗ jump to hunk
diff --git a/builtin/pack-objects.c b/builtin/pack-objects.c
index 89a8b5a976..62c540b4db 100644
--- a/builtin/pack-objects.c
+++ b/builtin/pack-objects.c
@@ -3204,14 +3204,18 @@ static int git_pack_config(const char *k, const char *v,
 		return 0;
 	}
 	if (!strcmp(k, "uploadpack.blobpackfileuri")) {
-		struct configured_exclusion *ex = xmalloc(sizeof(*ex));
+		struct configured_exclusion *ex;
 		const char *oid_end, *pack_end;
 		/*
 		 * Stores the pack hash. This is not a true object ID, but is
 		 * of the same form.
 		 */
 		struct object_id pack_hash;
 
+		if (!v)
+			return config_error_nonbool(k);
+
+		ex = xmalloc(sizeof(*ex));
 		if (parse_oid_hex(v, &ex->e.oid, &oid_end) ||
 		    *oid_end != ' ' ||
 		    parse_oid_hex(oid_end + 1, &pack_hash, &pack_end) ||
This isn't part of the diff and not a new issue, but why don't we
`return 0` when parsing this config correctly? We fall through to
`git_default_config()` even if we've successfully parsed the config key,
which seems like a bug to me.

Anyway, this case looks fine.

[snip]
quoted hunk ↗ jump to hunk
diff --git a/config.c b/config.c
index b330c7adb4..18085c7e38 100644
--- a/config.c
+++ b/config.c
@@ -1386,6 +1386,8 @@ static int git_default_core_config(const char *var, const char *value,
 		return 0;
 	}
 	if (!strcmp(var, "core.checkstat")) {
+		if (!value)
+			return config_error_nonbool(var);
 		if (!strcasecmp(value, "default"))
 			check_stat = 1;
 		else if (!strcasecmp(value, "minimal"))
We would ignore `true` here, so should we ignore implicit `true`, as
well?
quoted hunk ↗ jump to hunk
@@ -1547,11 +1549,15 @@ static int git_default_core_config(const char *var, const char *value,
 	}
 
 	if (!strcmp(var, "core.checkroundtripencoding")) {
+		if (!value)
+			return config_error_nonbool(var);
 		check_roundtrip_encoding = xstrdup(value);
 		return 0;
 	}
 
 	if (!strcmp(var, "core.notesref")) {
+		if (!value)
+			return config_error_nonbool(var);
 		notes_ref_name = xstrdup(value);
 		return 0;
 	}
I wonder the same here. We might as well use `xstrdup_or_null()`, but it
feels like the right thing to do to convert these to actual errors.
quoted hunk ↗ jump to hunk
@@ -426,10 +429,15 @@ int git_diff_ui_config(const char *var, const char *value,
 	if (!strcmp(var, "diff.orderfile"))
 		return git_config_pathname(&diff_order_file_cfg, var, value);
 
-	if (!strcmp(var, "diff.ignoresubmodules"))
+	if (!strcmp(var, "diff.ignoresubmodules")) {
+		if (!value)
+			return config_error_nonbool(var);
 		handle_ignore_submodules_arg(&default_diff_options, value);
+	}
 
 	if (!strcmp(var, "diff.submodule")) {
+		if (!value)
+			return config_error_nonbool(var);
 		if (parse_submodule_params(&default_diff_options, value))
 			warning(_("Unknown value for 'diff.submodule' config variable: '%s'"),
 				value);
Should we generate a warning instead according to the preexisting code
for "diff.submodule"?
quoted hunk ↗ jump to hunk
@@ -490,6 +501,8 @@ int git_diff_basic_config(const char *var, const char *value,
 
 	if (!strcmp(var, "diff.dirstat")) {
 		struct strbuf errmsg = STRBUF_INIT;
+		if (!value)
+			return config_error_nonbool(var);
 		default_diff_options.dirstat_permille = diff_dirstat_permille_default;
 		if (parse_dirstat_params(&default_diff_options, value, &errmsg))
 			warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
Same here, should we generate a warning instead?
quoted hunk ↗ jump to hunk
diff --git a/notes-utils.c b/notes-utils.c
index 97c031c26e..01f4f5b424 100644
--- a/notes-utils.c
+++ b/notes-utils.c
@@ -112,6 +112,8 @@ static int notes_rewrite_config(const char *k, const char *v,
 		}
 		return 0;
 	} else if (!c->refs_from_env && !strcmp(k, "notes.rewriteref")) {
+		if (!v)
+			return config_error_nonbool(k);
 		/* note that a refs/ prefix is implied in the
 		 * underlying for_each_glob_ref */
 		if (starts_with(v, "refs/notes/"))
Here, as well.
quoted hunk ↗ jump to hunk
diff --git a/trailer.c b/trailer.c
index b6de5d9cb2..b0e2ec224a 100644
--- a/trailer.c
+++ b/trailer.c
@@ -507,6 +507,8 @@ static int git_trailer_default_config(const char *conf_key, const char *value,
 				warning(_("unknown value '%s' for key '%s'"),
 					value, conf_key);
 		} else if (!strcmp(trailer_item, "separators")) {
+			if (!value)
+				return config_error_nonbool(conf_key);
 			separators = xstrdup(value);
 		}
 	}
And here.

Patrick

Attachments

Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help