[PATCH v1 2/2] convert: add alias support for 'working-tree-encoding' attributes

[larsxschneider@gmail.com]

From: Lars Schneider <larsxschneider@gmail.com>

In 107642fe26 ("convert: add 'working-tree-encoding' attribute",
2018-04-15) we added an attribute which defines the working tree
encoding of a file.

Some platforms might spell the name of a certain encoding differently or
some users might want to use different encodings on different platforms.
Add the Git config "encoding.<iconv-name>.insteadOf = <alias-name>" to
support these use-cases with a user specific mapping. If the alias
matches an existing encoding name, then the alias will take precedence.
The alias is case insensitive.

Example:

	(in .gitattributes)
	*.c	working-tree-encoding=foo

	(in config)
	[encoding "UTF-16"]
		insteadOf = foo

Signed-off-by: Lars Schneider <larsxschneider@gmail.com>
---
 Documentation/gitattributes.txt  | 19 ++++++++++++
 convert.c                        | 50 ++++++++++++++++++++++++++++++++
 t/t0028-working-tree-encoding.sh | 28 ++++++++++++++++++
 3 files changed, 97 insertions(+)

diff --git a/Documentation/gitattributes.txt b/Documentation/gitattributes.txt
index 92010b062e..3628f0e5cf 100644
--- a/Documentation/gitattributes.txt
+++ b/Documentation/gitattributes.txt
@@ -366,6 +366,25 @@ command to guess the encoding:
 file foo.ps1
 ------------------------
 
+The encoding in all examples above was directly defined in the Git
+attributes. In addition, it is possible to define encodings indirectly
+using aliases set via Git config:
+
+------------------------
+[encoding "UTF-16"]
+    insteadOf = my-custom-alias
+------------------------
+
+The alias name can be used in the Git attributes instead of the actual
+encoding name:
+
+------------------------
+*.ps1   text working-tree-encoding=my-custom-alias
+------------------------
+
+This mapping can be useful if equivalent encodings are spelled
+differently across platforms. It can also be useful if a user wants to
+use different encodings on different platforms for the same file.
 
 `ident`
 ^^^^^^^
diff --git a/convert.c b/convert.c
index 949bc783e4..4f19ce1a04 100644
--- a/convert.c
+++ b/convert.c
@@ -997,6 +997,15 @@ static int apply_filter(const char *path, const char *src, size_t len,
 	return 0;
 }
 
+struct alias2enc {
+	struct hashmap_entry ent; /* must be the first member! */
+	const char *alias;
+	const char *encoding;
+};
+
+static int encoding_aliases_initialized;
+static struct hashmap encoding_map;
+
 static int read_convert_config(const char *var, const char *value, void *cb)
 {
 	const char *key, *name;
@@ -1040,6 +1049,36 @@ static int read_convert_config(const char *var, const char *value, void *cb)
 			drv->required = git_config_bool(var, value);
 			return 0;
 		}
+	} else if (
+		parse_config_key(var, "encoding", &name, &namelen, &key) >= 0 &&
+		name &&	!strcmp(key, "insteadof")) {
+		/*
+		 * Encoding aliases are configured using
+		 * "encoding.<iconv-name>.insteadOf = <alias-name>".
+		 */
+		struct alias2enc *entry;
+		if (!value)
+			return config_error_nonbool(key);
+
+		if (!encoding_aliases_initialized) {
+			encoding_aliases_initialized = 1;
+			hashmap_init(&encoding_map, NULL, NULL, 0);
+			entry = NULL;
+		} else {
+			struct alias2enc hashkey;
+			hashmap_entry_init(&hashkey, strihash(value));
+			hashkey.alias = value;
+			entry = hashmap_get(&encoding_map, &hashkey, NULL);
+		}
+
+		if (!entry) {
+			entry = xmalloc(sizeof(*entry));
+			entry->encoding = xstrndup(name, namelen);
+			entry->alias = xstrdup(value);
+
+			hashmap_entry_init(entry, strihash(value));
+			hashmap_add(&encoding_map, entry);
+		}
 	}
 
 	return 0;
@@ -1225,6 +1264,17 @@ static const char *git_path_check_encoding(struct attr_check_item *check)
 		die(_("true/false are no valid working-tree-encodings"));
 	}
 
+	/* Check if an alias was defined for the encoding in the Git config */
+	if (encoding_aliases_initialized) {
+		struct alias2enc hashkey;
+		struct alias2enc *entry;
+		hashmap_entry_init(&hashkey, strihash(value));
+		hashkey.alias = value;
+		entry = hashmap_get(&encoding_map, &hashkey, NULL);
+		if (entry)
+			value = entry->encoding;
+	}
+
 	/* Don't encode to the default encoding */
 	if (same_encoding(value, default_encoding))
 		return NULL;
diff --git a/t/t0028-working-tree-encoding.sh b/t/t0028-working-tree-encoding.sh
index 12b8eb963a..d803e00cbe 100755
--- a/t/t0028-working-tree-encoding.sh
+++ b/t/t0028-working-tree-encoding.sh
@@ -242,4 +242,32 @@ test_expect_success 'check roundtrip encoding' '
 	git reset
 '
 
+test_expect_success 'encoding alias' '
+	test_when_finished "rm -f test.foo16.git test.foo8" &&
+	test_when_finished "git reset --hard HEAD" &&
+
+	test_config encoding.UTF-16.InsteadOf foo16 &&
+
+	text="hallo there!\ncan you read me with an alias?" &&
+	printf "$text" >test.foo8 &&
+	printf "$text" | iconv -f UTF-8 -t UTF-16 >test.foo16 &&
+
+	echo "*.foo16 text working-tree-encoding=fOO16" >.gitattributes &&
+	git add test.foo16 .gitattributes &&
+
+	git cat-file -p :test.foo16 >test.foo16.git &&
+	test_cmp_bin test.foo8 test.foo16.git
+'
+
+test_expect_success 'encoding alias overwrites existing encoding' '
+	test_when_finished "git reset --hard HEAD" &&
+
+	test_config encoding.CONFUSE.insteadOf UTF-16 &&
+
+	echo "*.garbage text working-tree-encoding=UTF-16" >.gitattributes &&
+	printf "garbage" >t.garbage &&
+	test_must_fail git add t.garbage 2>err.out &&
+	test_i18ngrep "failed to encode .* from CONFUSE to UTF-8" err.out
+'
+
 test_done
-- 
2.18.0