[Helper] Add -normalize-match option.

Decomposes string and remove accent characters before matching.
This makes o match ö, é match e and more.
It is not a perfect implementation but works.
Currently disables the match highlighting.

Fixes: #1119

6 files changed, 70 insertions, 3 deletions

diff --git a/config/config.c b/config/config.c
index 05723e5a..111ec742 100644
--- a/config/config.c
+++ b/config/config.c
@@ -170,5 +170,8 @@ Settings config = {
     .drun_reload_desktop_cache = FALSE,
 
     /** Benchmarks */
-    .benchmark_ui              = FALSE
+    .benchmark_ui              = FALSE,
+
+    /** normalize match */
+    .normalize_match 	       = FALSE
 };
diff --git a/doc/rofi.1 b/doc/rofi.1
index 8edd8174..f527bffb 100644
--- a/doc/rofi.1
+++ b/doc/rofi.1
@@ -373,6 +373,14 @@ Default: *true*
 .RE
 
 .PP
+\fB\fC\-normalize\-match\fR
+
+.PP
+Normalize the string before matching, so o will match ö, and é matches e.
+This is not a perfect implementation, but works.
+For now it disabled highlighting of the matched part.
+
+.PP
 \fB\fC\-no\-lazy\-grab\fR
 
 .PP
diff --git a/doc/rofi.1.markdown b/doc/rofi.1.markdown
index 4325ce2b..3c287361 100644
--- a/doc/rofi.1.markdown
+++ b/doc/rofi.1.markdown
@@ -214,6 +214,12 @@ Show the indicator that shows what part of the string is matched.
 
     Default: *true*
 
+`-normalize-match`
+
+Normalize the string before matching, so o will match ö, and é matches e.
+This is not a perfect implementation, but works.
+For now it disabled highlighting of the matched part. 
+
 `-no-lazy-grab`
 
 Disables lazy grab, this forces the keyboard being grabbed before gui is shown.
diff --git a/include/settings.h b/include/settings.h
index cecb9c34..491036b5 100644
--- a/include/settings.h
+++ b/include/settings.h
@@ -203,6 +203,8 @@ typedef struct
 
     /** Benchmark */
     gboolean       benchmark_ui;
+
+    gboolean       normalize_match;
 } Settings;
 /** Global Settings structure. */
 extern Settings config;
diff --git a/source/helper.c b/source/helper.c
index a7bfa3e8..0a5e86ca 100644
--- a/source/helper.c
+++ b/source/helper.c
@@ -180,10 +180,43 @@ static gchar *fuzzy_to_regex ( const char * input )
     return retv;
 }
 
+static char *utf8_helper_simplify_string ( const char *s)
+{
+  gunichar buf2[G_UNICHAR_MAX_DECOMPOSITION_LENGTH] = {0,};
+  char buf[6] = {0,};
+  // Compose the string in maximally composed form.
+  char * str = g_malloc0((g_utf8_strlen(s,0)*6+2));
+  char *striter = str;
+  for ( const char *iter = s; iter && *iter; iter = g_utf8_next_char ( iter )) {
+    gunichar uc = g_utf8_get_char ( iter );
+    int l = 0;
+    gsize dl = g_unichar_fully_decompose ( uc, FALSE, buf2, G_UNICHAR_MAX_DECOMPOSITION_LENGTH) ;
+    if ( dl ) {
+      l = g_unichar_to_utf8 ( buf2[0], buf);
+    } else {
+      l = g_unichar_to_utf8 ( uc, buf);
+    }
+    memcpy(striter, buf, l);
+    striter+=l;
+  }
+
+  return str;
+}
+
 // Macro for quickly generating regex for matching.
 static inline GRegex * R ( const char *s, int case_sensitive  )
 {
+  if ( config.normalize_match ) {
+    char *str = utf8_helper_simplify_string ( s );
+
+    GRegex *r = g_regex_new ( str, G_REGEX_OPTIMIZE | ( ( case_sensitive ) ? 0 : G_REGEX_CASELESS ), 0, NULL );
+
+    g_free ( str );
+    return r;
+  } else {
     return g_regex_new ( s, G_REGEX_OPTIMIZE | ( ( case_sensitive ) ? 0 : G_REGEX_CASELESS ), 0, NULL );
+
+  }
 }
 
 static rofi_int_matcher * create_regex ( const char *input, int case_sensitive )
@@ -380,6 +413,10 @@ int find_arg_char ( const char * const key, char *val )
 
 PangoAttrList *helper_token_match_get_pango_attr ( RofiHighlightColorStyle th, rofi_int_matcher**tokens, const char *input, PangoAttrList *retv )
 {
+    // Disable highlighting for normalize match, not supported atm.
+    if ( config.normalize_match ) {
+      return retv;
+    }
     // Do a tokenized match.
     if ( tokens ) {
         for ( int j = 0; tokens[j]; j++ ) {
@@ -453,10 +490,19 @@ int helper_token_match ( rofi_int_matcher* const *tokens, const char *input )
     int match = TRUE;
     // Do a tokenized match.
     if ( tokens ) {
+      if ( config.normalize_match ) {
+        char *r = utf8_helper_simplify_string(input);
+        for ( int j = 0; match && tokens[j]; j++ ) {
+          match  = g_regex_match ( tokens[j]->regex, r, 0, NULL );
+          match ^= tokens[j]->invert;
+        }
+        g_free(r);
+      } else {
         for ( int j = 0; match && tokens[j]; j++ ) {
-            match  = g_regex_match ( tokens[j]->regex, input, 0, NULL );
-            match ^= tokens[j]->invert;
+          match  = g_regex_match ( tokens[j]->regex, input, 0, NULL );
+          match ^= tokens[j]->invert;
         }
+      }
     }
     return match;
 }
diff --git a/source/xrmoptions.c b/source/xrmoptions.c
index aff2ec9b..9e4289d0 100644
--- a/source/xrmoptions.c
+++ b/source/xrmoptions.c
@@ -231,6 +231,8 @@ static XrmOption xrmOptions[] = {
       "DRUN: build and use a cache with desktop file content.", CONFIG_DEFAULT },
     { xrm_Boolean, "drun-reload-desktop-cache", { .snum  = &config.drun_reload_desktop_cache            }, NULL,
       "DRUN: If enabled, reload the cache with desktop file content.", CONFIG_DEFAULT },
+    { xrm_Boolean, "normalize-match", 		      { .snum  = &config.normalize_match                      }, NULL,
+            "Normalize string when matching (implies -no-show-match).", CONFIG_DEFAULT },
 };
 
 /** Dynamic array of extra options */