Detect gitlog and gitstash filetypes from first line of buffer

This works when opening files, but more importantly it also works when
piping to stdin at startup. For example:

    git log -p HEAD~8..HEAD | dte
    git stash list -p | dte

See also:

* Commit f75e62f
* Commit 03900b6

src/filetype.c

@@ -10,6 +10,7 @@
 #include "util/log.h"
 #include "util/path.h"
 #include "util/str-util.h"
+#include "util/strtonum.h"
 #include "util/xmalloc.h"
 #include "util/xmemmem.h"
 

src/filetype/names.c

@@ -33,7 +33,9 @@ typedef enum {
     GETTEXT,
     GITCOMMIT,
     GITIGNORE,
+    GITLOG,
     GITREBASE,
+    GITSTASH,
     GLSL,
     GNUPLOT,
     GO,
@@ -155,7 +157,9 @@ static const char builtin_filetype_names[NR_BUILTIN_FILETYPES][12] = {
     [GETTEXT] = "gettext",
     [GITCOMMIT] = "gitcommit",
     [GITIGNORE] = "gitignore",
+    [GITLOG] = "gitlog",
     [GITREBASE] = "gitrebase",
+    [GITSTASH] = "gitstash",
     [GLSL] = "glsl",
     [GNUPLOT] = "gnuplot",
     [GO] = "go",

src/filetype/signatures.c

@@ -65,6 +65,21 @@ static FileTypeEnum filetype_from_emacs_var(const StringView line)
     return (ft >= 0) ? (FileTypeEnum)ft : NONE;
 }
 
+static bool is_gitlog_commit_line(StringView line)
+{
+    if (!strview_remove_matching_prefix(&line, "commit ") || line.length < 40) {
+        return false;
+    }
+
+    size_t ndigits = ascii_hex_prefix_length(line.data, line.length);
+    if (ndigits < 40) {
+        return false;
+    }
+
+    strview_remove_prefix(&line, ndigits);
+    return line.length == 0 || strview_has_prefix_and_suffix(&line, " (", ")");
+}
+
 static FileTypeEnum filetype_from_signature(const StringView line)
 {
     if (line.length < 5) {
@@ -95,8 +110,12 @@ static FileTypeEnum filetype_from_signature(const StringView line)
         break;
     case 'I':
         return strview_has_prefix(&line, "ISO-10303-21;") ? STEP : NONE;
+    case 'c':
+        return is_gitlog_commit_line(line) ? GITLOG : NONE;
     case 'd':
         return strview_has_prefix(&line, "diff --git") ? DIFF : NONE;
+    case 's':
+        return strview_has_prefix(&line, "stash@{") ? GITSTASH : NONE;
     case '/':
     case '.':
     case ';':

src/util/string-view.h

@@ -112,6 +112,15 @@ static inline bool strview_has_suffix(const StringView *sv, const char *suffix)
     return strview_has_strn_suffix(sv, suffix, strlen(suffix));
 }
 
+NONNULL_ARGS
+static inline bool strview_has_prefix_and_suffix (
+    const StringView *sv,
+    const char *prefix,
+    const char *suffix
+) {
+    return strview_has_prefix(sv, prefix) && strview_has_suffix(sv, suffix);
+}
+
 NONNULL_ARGS
 static inline bool strview_isblank(const StringView *sv)
 {

src/util/strtonum.h

@@ -20,6 +20,20 @@ static inline unsigned int hex_decode(unsigned char c)
     return hex_decode_table[c];
 }
 
+static inline bool ascii_isxdigit(unsigned char c)
+{
+    return hex_decode(c) <= 0xF;
+}
+
+static inline size_t ascii_hex_prefix_length(const char *str, size_t len)
+{
+    size_t i = 0;
+    while (i < len && ascii_isxdigit(str[i])) {
+        i++;
+    }
+    return i;
+}
+
 WARN_UNUSED_RESULT
 static inline size_t buf_parse_hex_uint(const char *str, size_t len, unsigned int *valp)
 {

test/filetype.c

@@ -209,6 +209,12 @@ static void test_find_ft_firstline(TestContext *ctx)
         {"#autoload", "sh"},
         {"#compdef dte", "sh"},
         {"#compdef", NULL},
+        {"stash@{0}: WIP on master: ...", "gitstash"},
+        {"stash@", NULL},
+        {"commit 8c6db8e8f8fbf055633ffb7a8bb449bb177adf01 (...)", "gitlog"},
+        {"commit 8c6db8e8f8fbf055633ffb7a8bb449bb177adf01", "gitlog"},
+        {"commit 8c6db8e8f8fbf055633ffb7a8bb449bb177adf01 ", NULL},
+        {"commit 8c6db8e8f8fbf055633ffb7a8bb449bb177adf0 (...)", NULL},
 
         // Emacs style file-local variables
         {"<!--*-xml-*-->", "xml"},

test/util.c

@@ -428,8 +428,8 @@ static void test_ascii(TestContext *ctx)
         EXPECT_EQ(ascii_isdigit(i), !!isdigit(i));
         EXPECT_EQ(ascii_isblank(i), !!isblank(i));
         EXPECT_EQ(ascii_isprint(i), !!isprint(i));
+        EXPECT_EQ(ascii_isxdigit(i), !!isxdigit(i));
         EXPECT_EQ(u_is_ascii_upper(i), !!isupper(i));
-        EXPECT_EQ(hex_decode(i) <= 0xF, !!isxdigit(i));
         EXPECT_EQ(is_alpha_or_underscore(i), !!isalpha(i) || i == '_');
         EXPECT_EQ(is_alnum_or_underscore(i), !!isalnum(i) || i == '_');
         if (i != '\v' && i != '\f') {
@@ -1575,6 +1575,14 @@ static void test_u_is_cntrl(TestContext *ctx)
     EXPECT_FALSE(u_is_cntrl(0xFF));
 }
 
+static void test_u_is_unicode(TestContext *ctx)
+{
+    EXPECT_TRUE(u_is_unicode(0));
+    EXPECT_TRUE(u_is_unicode(1));
+    EXPECT_TRUE(u_is_unicode(UNICODE_MAX_VALID_CODEPOINT));
+    EXPECT_FALSE(u_is_unicode(UNICODE_MAX_VALID_CODEPOINT + 1));
+}
+
 static void test_u_is_zero_width(TestContext *ctx)
 {
     // Default ignorable codepoints:
@@ -3246,6 +3254,7 @@ static const TestEntry tests[] = {
     TEST(test_u_is_upper),
     TEST(test_u_is_ascii_upper),
     TEST(test_u_is_cntrl),
+    TEST(test_u_is_unicode),
     TEST(test_u_is_zero_width),
     TEST(test_u_is_special_whitespace),
     TEST(test_u_is_unprintable),