issue #8375: Lowercase search does not find non-ASCII uppercase pages and vice versa

1 files changed, 56 insertions, 0 deletions

diff --git a/src/caseconvert.py b/src/caseconvert.py
new file mode 100644
index 0000000..0b0b57b
--- /dev/null
+++ b/src/caseconvert.py
@@ -0,0 +1,56 @@
+# python3 script to generate caseconvert.h.
+# It uses difference in lower() and upper() on a character to make a mapping
+# that maps a given unicode point to either a lower or upper case UTF-8 character.
+# this also include multi-byte characters.
+
+import codecs
+
+toupper = {}
+tolower = {}
+
+def writeMapping(file,mapping):
+    for k,v in sorted(mapping.items()):
+        file.write(u"    case %s /* %s */: return u8\"%s\" /* %s */;\n" %
+                (hex(ord(k[0])),k,v,",".join("0x{:02x}".format(ord(c)) for c in v)))
+
+# create mappings of characters whose upper and lower case differs
+for codeValue in range(0,0x1FFFF):
+        s = chr(codeValue)
+        sl = s.lower()
+        su = s.upper()
+        if ord(s[0])!=ord(sl[0]):
+            tolower[s]=sl
+        if ord(s[0])!=ord(su[0]):
+            toupper[s]=su
+
+file = codecs.open("caseconvert.h", "w", "utf-8")
+file.write(r'''/** This file is generated by python3 caseconvert.py. DO NOT EDIT! */
+
+#ifndef CASECONVERT_H
+#define CASECONVERT_H
+
+#include <cstdint>
+#include <string>
+
+inline const char *convertUnicodeToUpper(uint32_t code)
+{
+  switch(code)
+  {
+''');
+writeMapping(file,toupper);
+file.write(r'''    default: return nullptr;
+  }
+}
+
+inline const char *convertUnicodeToLower(uint32_t code)
+{
+  switch(code)
+  {
+''');
+writeMapping(file,tolower);
+file.write(r'''    default: return nullptr;
+  }
+}
+
+#endif
+''');