FindAdv: Special handling for \dot{i} and 'ß'

Different behaviour in regexp{..} for 'İ' and 'ß':
1.) lowercase routine for 'İ' gives 'İ', so that if we are searching
  while ignoring case, the string '\dot{I}' is converted to '\dot{i}'.
  In this case we have to change it to 'İ' (instead of 'i', as one would expect).

2.) If 'ß' is inserted via keybord on fresh created regexp box it appears as \lyxmathsym{ß},
  if pasted from the lyx-screen it appears as \text{ß}

diff --git a/src/lyxfind.cpp b/src/lyxfind.cpp
index 8c11bdbe8e84ab804ae0dee421fa4ff017970798..a2c498e9b901bfcb1e35ac6f9a99cee466ec4cc4 100644 (file)
--- a/src/lyxfind.cpp
+++ b/src/lyxfind.cpp
@@ -1263,8 +1263,13 @@ static void buildaccent(string n, string param, string values)
       // get the corresponding utf8-value
       if ((values[start] & 0xc0) != 0xc0) {
         // should not happen, utf8 encoding starts at least with 11xxxxxx
-        start++;
-        continue;
+       // but value for '\dot{i}' is 'i', which is ascii
+       if ((values[start] & 0x80) == 0) {
+         // is ascii
+         accents[key] = values.substr(start, 1);
+       }
+       start++;
+       continue;
       }
       for (int j = 1; ;j++) {
         if (start + j >= values.size()) {
@@ -1272,7 +1277,7 @@ static void buildaccent(string n, string param, string values)
           start = values.size() - 1;
           break;
         }
-        else if ((values[start+j] & 0xc0) == 0xc0) {
+        else if ((values[start+j] & 0xc0) != 0x80) {
           // This is the first byte of following utf8 char
           accents[key] = values.substr(start, j);
           start += j;
@@ -1289,11 +1294,12 @@ static void buildAccentsMap()
   accents["i"] = "ı";
   accents["jmath"] = "ȷ";
   accents["lyxmathsym{ß}"] = "ß";
+  accents["text{ß}"] = "ß";
   accents["ddot{\\imath}"] = "ï";
-  buildaccent("ddot", "aAeEiIoOuUyY",
-                      "äÃ\84ëÃ\8bïÃ\8föÃ\96üÃ\9cÿŸ");     // umlaut
-  buildaccent("dot|.", "cCeEgGiIzZaAoObBdDfFyY",
-                       "Ä\8bÄ\8aÄ\97Ä\96Ä¡Ä iİżŻȧȦȯȮá¸\83á¸\82á¸\8bá¸\8aá¸\9fá¸\9eáº\8fáº\8e");
+  buildaccent("ddot", "aAeEiIioOuUyY",
+                      "äÃ\84ëÃ\8bïÃ\8fïöÃ\96üÃ\9cÿŸ");   // umlaut
+  buildaccent("dot|.", "cCeEGgIizZaAoObBdDfFyY",
+                       "Ä\8bÄ\8aÄ\97Ä\96Ä Ä¡Ä°Ä°Å¼Å»È§È¦È¯È®á¸\83á¸\82á¸\8bá¸\8aá¸\9fá¸\9eáº\8fáº\8e"); // dot{i} can only happen if ignoring case, but there is no lowercase of 'Ä°'
   accents["acute{\\imath}"] = "í";
   buildaccent("acute", "aAcCeElLoOnNrRsSuUyYzZiI",
                        "áÁćĆéÉĺĹóÓńŃŕŔśŚúÚýÝźŹíÍ");
@@ -1302,8 +1308,8 @@ static void buildAccentsMap()
                             "åÅůŮẘẙ");  // ring
   accents["check{\\imath}"] = "ǐ";
   accents["check{\\jmath}"] = "ǰ";
-  buildaccent("check|v", "cCdDaAeEiIoOuUgGkKhHlLnNrRsSTzZ",
-                         "Ä\8dÄ\8cÄ\8fÄ\8eÇ\8eÇ\8dÄ\9bÄ\9aÇ\90Ç\8fÇ\92Ç\91Ç\94Ç\93ǧǦǩǨÈ\9fÈ\9eľĽÅ\88Å\87Å\99Å\98šŠŤžŽ");    // caron
+  buildaccent("check|v", "cCdDaAeEiIoOuUgGkKhHlLnNrRsSTtzZ",
+                         "Ä\8dÄ\8cÄ\8fÄ\8eÇ\8eÇ\8dÄ\9bÄ\9aÇ\90Ç\8fÇ\92Ç\91Ç\94Ç\93ǧǦǩǨÈ\9fÈ\9eľĽÅ\88Å\87Å\99Å\98šŠŤťžŽ");  // caron
   accents["hat{\\imath}"] = "î";
   accents["hat{\\jmath}"] = "ĵ";
   buildaccent("hat|^", "aAeEiIcCgGhHjJsSwWyYzZoOuU",
@@ -1332,7 +1338,7 @@ void Intervall::removeAccents()
 {
   if (accents.empty())
     buildAccentsMap();
-  static regex const accre("\\\\((.|grave|breve|lyxmathsym|ddot|dot|acute|dacute|mathring|check|hat|bar|tilde|subdot)\\{[^\\{\\}]+\\}|(i|imath|jmath)(?![a-zA-Z]))");
+  static regex const accre("\\\\((.|grave|breve|lyxmathsym|text|ddot|dot|acute|dacute|mathring|check|hat|bar|tilde|subdot)\\{[^\\{\\}]+\\}|(i|imath|jmath)(?![a-zA-Z]))");
   smatch sub;
   for (sregex_iterator itacc(par.begin(), par.end(), accre), end; itacc != end; ++itacc) {
     sub = *itacc;