Commit 7b3bd11f92f0bed76f25eab8c79c0eca21e037bc

David Herrmann 2012-10-16T16:05:34

Add xkb_keysym_from_name() flags argument for case-insensitive search This adds a flags argument to xkb_keysym_from_name() so we can perform a case-insensitive search. This should really be supported as many keysyms have really weird capitalization-rules. However, as this may produce conflicts, users must be warned to only use this for fallback paths or error-recovery. This is also the reason why the internal XKB parsers still use the case-sensitive search. This also adds some test-cases so the expected results are really produced. The binary-size does _not_ change with this patch. However, case-sensitive search may be slightly slower with this patch. But this is barely measurable. [ran: use bool instead of int for icase, add a recommendation to the doc, and test a couple "thorny" cases.] Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>

diff --git a/makekeys.py b/makekeys.py
index 94885c0..996ff3d 100644
--- a/makekeys.py
+++ b/makekeys.py
@@ -12,7 +12,7 @@ print('''struct name_keysym {
 };\n''')
 
 print('static const struct name_keysym name_to_keysym[] = {');
-for (name, _) in sorted(entries, key=lambda e: e[0]):
+for (name, _) in sorted(entries, key=lambda e: e[0].lower()):
     print('    {{ "{name}", XKB_KEY_{name} }},'.format(name=name))
 print('};\n')
 
diff --git a/src/keysym.c b/src/keysym.c
index cff0b15..4f6285b 100644
--- a/src/keysym.c
+++ b/src/keysym.c
@@ -64,7 +64,7 @@ static int
 compare_by_name(const void *a, const void *b)
 {
     const struct name_keysym *key = a, *entry = b;
-    return strcmp(key->name, entry->name);
+    return strcasecmp(key->name, entry->name);
 }
 
 XKB_EXPORT int
@@ -93,22 +93,80 @@ xkb_keysym_get_name(xkb_keysym_t ks, char *buffer, size_t size)
     return snprintf(buffer, size, "0x%08x", ks);
 }
 
+/*
+ * Find the correct keysym if one case-insensitive match is given.
+ *
+ * The name_to_keysym table is sorted by strcasecmp(). So bsearch() may return
+ * _any_ of all possible case-insensitive duplicates. This function searches the
+ * returned entry @entry, all previous and all next entries that match by
+ * case-insensitive comparison and returns the exact match to @name. If @icase
+ * is true, then this returns the best case-insensitive match instead of a
+ * correct match.
+ * The "best" case-insensitive match is the lower-case keysym which we find with
+ * the help of xkb_keysym_is_lower().
+ * The only keysyms that only differ by letter-case are keysyms that are
+ * available as lower-case and upper-case variant (like KEY_a and KEY_A). So
+ * returning the first lower-case match is enough in this case.
+ */
+static const struct name_keysym *
+find_sym(const struct name_keysym *entry, const char *name, bool icase)
+{
+    const struct name_keysym *iter, *last;
+    size_t len = sizeof(name_to_keysym) / sizeof(*name_to_keysym);
+
+    if (!entry)
+        return NULL;
+
+    if (!icase && strcmp(entry->name, name) == 0)
+        return entry;
+    if (icase && xkb_keysym_is_lower(entry->keysym))
+        return entry;
+
+    for (iter = entry - 1; iter >= name_to_keysym; --iter) {
+        if (!icase && strcmp(iter->name, name) == 0)
+            return iter;
+        if (strcasecmp(iter->name, entry->name) != 0)
+            break;
+        if (icase && xkb_keysym_is_lower(iter->keysym))
+            return iter;
+    }
+
+    last = name_to_keysym + len;
+    for (iter = entry + 1; iter < last; --iter) {
+        if (!icase && strcmp(iter->name, name) == 0)
+            return iter;
+        if (strcasecmp(iter->name, entry->name) != 0)
+            break;
+        if (icase && xkb_keysym_is_lower(iter->keysym))
+            return iter;
+    }
+
+    if (icase)
+        return entry;
+    return NULL;
+}
+
 XKB_EXPORT xkb_keysym_t
-xkb_keysym_from_name(const char *s)
+xkb_keysym_from_name(const char *s, enum xkb_keysym_flags flags)
 {
     const struct name_keysym search = { .name = s, .keysym = 0 };
     const struct name_keysym *entry;
     char *tmp;
     xkb_keysym_t val;
+    bool icase = !!(flags & XKB_KEYSYM_CASE_INSENSITIVE);
+
+    if (flags & ~XKB_KEYSYM_CASE_INSENSITIVE)
+        return XKB_KEY_NoSymbol;
 
     entry = bsearch(&search, name_to_keysym,
                     sizeof(name_to_keysym) / sizeof(*name_to_keysym),
                     sizeof(*name_to_keysym),
                     compare_by_name);
+    entry = find_sym(entry, s, icase);
     if (entry)
         return entry->keysym;
 
-    if (*s == 'U') {
+    if (*s == 'U' || (icase && *s == 'u')) {
         val = strtoul(&s[1], &tmp, 16);
         if (tmp && *tmp != '\0')
             return XKB_KEY_NoSymbol;
@@ -121,7 +179,7 @@ xkb_keysym_from_name(const char *s)
             return XKB_KEY_NoSymbol;
         return val | 0x01000000;
     }
-    else if (s[0] == '0' && s[1] == 'x') {
+    else if (s[0] == '0' && (s[1] == 'x' || (icase && s[1] == 'X'))) {
         val = strtoul(&s[2], &tmp, 16);
         if (tmp && *tmp != '\0')
             return XKB_KEY_NoSymbol;
@@ -132,13 +190,14 @@ xkb_keysym_from_name(const char *s)
     /* Stupid inconsistency between the headers and XKeysymDB: the former has
      * no separating underscore, while some XF86* syms in the latter did.
      * As a last ditch effort, try without. */
-    if (strncmp(s, "XF86_", 5) == 0) {
+    if (strncmp(s, "XF86_", 5) == 0 ||
+        (icase && strncasecmp(s, "XF86_", 5) == 0)) {
         xkb_keysym_t ret;
         tmp = strdup(s);
         if (!tmp)
             return XKB_KEY_NoSymbol;
         memmove(&tmp[4], &tmp[5], strlen(s) - 5 + 1);
-        ret = xkb_keysym_from_name(tmp);
+        ret = xkb_keysym_from_name(tmp, flags);
         free(tmp);
         return ret;
     }
diff --git a/src/xkbcomp/expr.c b/src/xkbcomp/expr.c
index eb043e1..dc64d78 100644
--- a/src/xkbcomp/expr.c
+++ b/src/xkbcomp/expr.c
@@ -641,7 +641,7 @@ ExprResolveKeySym(struct xkb_context *ctx, const ExprDef *expr,
     if (expr->op == EXPR_IDENT) {
         const char *str;
         str = xkb_atom_text(ctx, expr->value.str);
-        *sym_rtrn = xkb_keysym_from_name(str);
+        *sym_rtrn = xkb_keysym_from_name(str, 0);
         if (*sym_rtrn != XKB_KEY_NoSymbol)
             return true;
     }
diff --git a/src/xkbcomp/symbols.c b/src/xkbcomp/symbols.c
index 10efef0..9900469 100644
--- a/src/xkbcomp/symbols.c
+++ b/src/xkbcomp/symbols.c
@@ -652,7 +652,7 @@ LookupKeysym(const char *str, xkb_keysym_t *sym_rtrn)
         return 1;
     }
 
-    sym = xkb_keysym_from_name(str);
+    sym = xkb_keysym_from_name(str, 0);
     if (sym != XKB_KEY_NoSymbol) {
         *sym_rtrn = sym;
         return 1;
diff --git a/test/keyseq.c b/test/keyseq.c
index d5993ec..cb53739 100644
--- a/test/keyseq.c
+++ b/test/keyseq.c
@@ -305,7 +305,7 @@ main(void)
                         KEY_RIGHTSHIFT,  UP,    XKB_KEY_Shift_R,         NEXT,
                         KEY_V,           BOTH,  XKB_KEY_Cyrillic_ZHE,    FINISH));
 
-#define KS(name) xkb_keysym_from_name(name)
+#define KS(name) xkb_keysym_from_name(name, 0)
 
     /* Test that levels (1-5) in de(neo) symbols map work. */
     assert(test_key_seq(keymap,
diff --git a/test/keysym.c b/test/keysym.c
index ef934ab..595f829 100644
--- a/test/keysym.c
+++ b/test/keysym.c
@@ -29,7 +29,7 @@ test_string(const char *string, xkb_keysym_t expected)
 {
     xkb_keysym_t keysym;
 
-    keysym = xkb_keysym_from_name(string);
+    keysym = xkb_keysym_from_name(string, 0);
 
     fprintf(stderr, "Expected string %s -> %x\n", string, expected);
     fprintf(stderr, "Received string %s -> %x\n\n", string, keysym);
@@ -38,6 +38,19 @@ test_string(const char *string, xkb_keysym_t expected)
 }
 
 static int
+test_casestring(const char *string, xkb_keysym_t expected)
+{
+    xkb_keysym_t keysym;
+
+    keysym = xkb_keysym_from_name(string, XKB_KEYSYM_CASE_INSENSITIVE);
+
+    fprintf(stderr, "Expected casestring %s -> %x\n", string, expected);
+    fprintf(stderr, "Received casestring %s -> %x\n\n", string, keysym);
+
+    return keysym == expected;
+}
+
+static int
 test_keysym(xkb_keysym_t keysym, const char *expected)
 {
     char s[16];
@@ -75,12 +88,39 @@ main(void)
     assert(test_string("VoidSymbol", 0xFFFFFF));
     assert(test_string("U4567", 0x1004567));
     assert(test_string("0x10203040", 0x10203040));
+    assert(test_string("a", 0x61));
+    assert(test_string("A", 0x41));
+    assert(test_string("ch", 0xfea0));
+    assert(test_string("Ch", 0xfea1));
+    assert(test_string("CH", 0xfea2));
+    assert(test_string("THORN", 0x00de));
+    assert(test_string("Thorn", 0x00de));
+    assert(test_string("thorn", 0x00fe));
 
     assert(test_keysym(0x1008FF56, "XF86Close"));
     assert(test_keysym(0x0, "NoSymbol"));
     assert(test_keysym(0x1008FE20, "XF86Ungrab"));
     assert(test_keysym(0x01001234, "U1234"));
 
+    assert(test_casestring("Undo", 0xFF65));
+    assert(test_casestring("UNDO", 0xFF65));
+    assert(test_casestring("A", 0x61));
+    assert(test_casestring("a", 0x61));
+    assert(test_casestring("ThisKeyShouldNotExist", XKB_KEY_NoSymbol));
+    assert(test_casestring("XF86_Switch_vT_5", 0x1008FE05));
+    assert(test_casestring("xF86_SwitcH_VT_5", 0x1008FE05));
+    assert(test_casestring("xF86SwiTch_VT_5", 0x1008FE05));
+    assert(test_casestring("xF86Switch_vt_5", 0x1008FE05));
+    assert(test_casestring("VoidSymbol", 0xFFFFFF));
+    assert(test_casestring("vOIDsymBol", 0xFFFFFF));
+    assert(test_casestring("U4567", 0x1004567));
+    assert(test_casestring("u4567", 0x1004567));
+    assert(test_casestring("0x10203040", 0x10203040));
+    assert(test_casestring("0X10203040", 0x10203040));
+    assert(test_casestring("THORN", 0x00fe));
+    assert(test_casestring("Thorn", 0x00fe));
+    assert(test_casestring("thorn", 0x00fe));
+
     assert(test_utf8(XKB_KEY_y, "y"));
     assert(test_utf8(XKB_KEY_u, "u"));
     assert(test_utf8(XKB_KEY_m, "m"));
@@ -100,13 +140,13 @@ main(void)
 
     assert(xkb_keysym_is_lower(XKB_KEY_a));
     assert(xkb_keysym_is_lower(XKB_KEY_Greek_lambda));
-    assert(xkb_keysym_is_lower(xkb_keysym_from_name("U03b1"))); /* GREEK SMALL LETTER ALPHA */
-    assert(xkb_keysym_is_lower(xkb_keysym_from_name("U03af"))); /* GREEK SMALL LETTER IOTA WITH TONOS */
+    assert(xkb_keysym_is_lower(xkb_keysym_from_name("U03b1", 0))); /* GREEK SMALL LETTER ALPHA */
+    assert(xkb_keysym_is_lower(xkb_keysym_from_name("U03af", 0))); /* GREEK SMALL LETTER IOTA WITH TONOS */
 
     assert(xkb_keysym_is_upper(XKB_KEY_A));
     assert(xkb_keysym_is_upper(XKB_KEY_Greek_LAMBDA));
-    assert(xkb_keysym_is_upper(xkb_keysym_from_name("U0391"))); /* GREEK CAPITAL LETTER ALPHA */
-    assert(xkb_keysym_is_upper(xkb_keysym_from_name("U0388"))); /* GREEK CAPITAL LETTER EPSILON WITH TONOS */
+    assert(xkb_keysym_is_upper(xkb_keysym_from_name("U0391", 0))); /* GREEK CAPITAL LETTER ALPHA */
+    assert(xkb_keysym_is_upper(xkb_keysym_from_name("U0388", 0))); /* GREEK CAPITAL LETTER EPSILON WITH TONOS */
 
     assert(!xkb_keysym_is_upper(XKB_KEY_a));
     assert(!xkb_keysym_is_lower(XKB_KEY_A));
@@ -114,8 +154,8 @@ main(void)
     assert(!xkb_keysym_is_upper(XKB_KEY_Return));
     assert(!xkb_keysym_is_lower(XKB_KEY_hebrew_aleph));
     assert(!xkb_keysym_is_upper(XKB_KEY_hebrew_aleph));
-    assert(!xkb_keysym_is_upper(xkb_keysym_from_name("U05D0"))); /* HEBREW LETTER ALEF */
-    assert(!xkb_keysym_is_lower(xkb_keysym_from_name("U05D0"))); /* HEBREW LETTER ALEF */
+    assert(!xkb_keysym_is_upper(xkb_keysym_from_name("U05D0", 0))); /* HEBREW LETTER ALEF */
+    assert(!xkb_keysym_is_lower(xkb_keysym_from_name("U05D0", 0))); /* HEBREW LETTER ALEF */
     assert(!xkb_keysym_is_lower(XKB_KEY_8));
     assert(!xkb_keysym_is_upper(XKB_KEY_8));
 
diff --git a/xkbcommon/xkbcommon.h b/xkbcommon/xkbcommon.h
index 5d6a295..ddc715c 100644
--- a/xkbcommon/xkbcommon.h
+++ b/xkbcommon/xkbcommon.h
@@ -319,18 +319,34 @@ struct xkb_rule_names {
 int
 xkb_keysym_get_name(xkb_keysym_t keysym, char *buffer, size_t size);
 
+/** Flags for xkb_keysym_from_name(). */
+enum xkb_keysym_flags {
+    /** Find keysym by case-insensitive search. */
+    XKB_KEYSYM_CASE_INSENSITIVE = (1 << 0),
+};
+
 /**
  * Get a keysym from its name.
  *
  * @param name The name of a keysym. See remarks in xkb_keysym_get_name();
  * this function will accept any name returned by that function.
+ * @param flags A set of flags controlling how the search is done. If
+ * invalid flags are passed, this will fail with XKB_KEY_NoSymbol.
+ *
+ * If you use the XKB_KEYSYM_CASE_INSENSITIVE flag and two keysym names
+ * differ only by case, then the lower-case keysym is returned.  For
+ * instance, for KEY_a and KEY_A, this function would return KEY_a for the
+ * case-insensitive search.  If this functionality is needed, it is
+ * recommended to first call this function without this flag; and if that
+ * fails, only then to try with this flag, while possibly warning the user
+ * he had misspelled the name, and might get wrong results.
  *
  * @returns The keysym. If the name is invalid, returns XKB_KEY_NoSymbol.
  *
  * @sa xkb_keysym_t
  */
 xkb_keysym_t
-xkb_keysym_from_name(const char *name);
+xkb_keysym_from_name(const char *name, enum xkb_keysym_flags flags);
 
 /**
  * Get the Unicode/UTF-8 representation of a keysym.