Rewrite fl_utf.c under the driver model: the file disappears and its content is moved to fl_utf8.cxx

All functions of fl_utf.c keep their C API in fl_utf8.cxx

git-svn-id: file:///fltk/svn/fltk/branches/branch-1.3-porting@11472 ea41ed52-d2ee-0310-a9c1-e6b18d33e121

1 files changed, 149 insertions, 0 deletions

diff --git a/src/Fl_System_Driver.cxx b/src/Fl_System_Driver.cxx
index 3fab6bf68..4907c9c2d 100644
--- a/src/Fl_System_Driver.cxx
+++ b/src/Fl_System_Driver.cxx
@@ -18,9 +18,11 @@
 
 
 #include <FL/Fl_System_Driver.H>
+#include <FL/fl_utf8.h>
 #include <stdlib.h>
 #include <stdio.h>
 #include <stdarg.h>
+#include <string.h>
 
 const int Fl_System_Driver::flNoValue =	    0x0000;
 const int Fl_System_Driver::flWidthValue =  0x0004;
@@ -218,6 +220,153 @@ char *Fl_System_Driver::getenv(const char *v) {
   return ::getenv(v);
 }
 
+unsigned Fl_System_Driver::utf8towc(const char* src, unsigned srclen, wchar_t* dst, unsigned dstlen) {
+  const char* p = src;
+  const char* e = src+srclen;
+  unsigned count = 0;
+  if (dstlen) for (;;) {
+    if (p >= e) {
+      dst[count] = 0;
+      return count;
+    }
+    if (!(*p & 0x80)) { /* ascii */
+      dst[count] = *p++;
+    } else {
+      int len; unsigned ucs = fl_utf8decode(p,e,&len);
+      p += len;
+      dst[count] = (wchar_t)ucs;
+    }
+    if (++count == dstlen) {dst[count-1] = 0; break;}
+  }
+  /* we filled dst, measure the rest: */
+  while (p < e) {
+    if (!(*p & 0x80)) p++;
+    else {
+      int len; fl_utf8decode(p,e,&len);
+      p += len;
+    }
+    ++count;
+  }
+  return count;
+}
+
+unsigned Fl_System_Driver::utf8fromwc(char* dst, unsigned dstlen, const wchar_t* src, unsigned srclen)
+{
+  unsigned i = 0;
+  unsigned count = 0;
+  if (dstlen) for (;;) {
+    unsigned ucs;
+    if (i >= srclen) {dst[count] = 0; return count;}
+    ucs = src[i++];
+    if (ucs < 0x80U) {
+      dst[count++] = ucs;
+      if (count >= dstlen) {dst[count-1] = 0; break;}
+    } else if (ucs < 0x800U) { /* 2 bytes */
+      if (count+2 >= dstlen) {dst[count] = 0; count += 2; break;}
+      dst[count++] = 0xc0 | (ucs >> 6);
+      dst[count++] = 0x80 | (ucs & 0x3F);
+    } else if (ucs >= 0x10000) {
+      if (ucs > 0x10ffff) {
+        ucs = 0xfffd;
+        goto J1;
+      }
+      if (count+4 >= dstlen) {dst[count] = 0; count += 4; break;}
+      dst[count++] = 0xf0 | (ucs >> 18);
+      dst[count++] = 0x80 | ((ucs >> 12) & 0x3F);
+      dst[count++] = 0x80 | ((ucs >> 6) & 0x3F);
+      dst[count++] = 0x80 | (ucs & 0x3F);
+    } else {
+J1:
+      /* all others are 3 bytes: */
+      if (count+3 >= dstlen) {dst[count] = 0; count += 3; break;}
+      dst[count++] = 0xe0 | (ucs >> 12);
+      dst[count++] = 0x80 | ((ucs >> 6) & 0x3F);
+      dst[count++] = 0x80 | (ucs & 0x3F);
+    }
+  }
+  /* we filled dst, measure the rest: */
+  while (i < srclen) {
+    unsigned ucs = src[i++];
+    if (ucs < 0x80U) {
+      count++;
+    } else if (ucs < 0x800U) { /* 2 bytes */
+      count += 2;
+    } else if (ucs >= 0x10000 && ucs <= 0x10ffff) {
+      count += 4;
+    } else {
+      count += 3;
+    }
+  }
+  return count;
+}
+
+int Fl_System_Driver::utf8locale() {
+  static int ret = 2;
+  if (ret == 2) {
+    char* s;
+    ret = 1; // assume UTF-8 if no locale
+    if (((s = getenv("LC_CTYPE")) && *s) ||
+        ((s = getenv("LC_ALL"))   && *s) ||
+        ((s = getenv("LANG"))     && *s)) {
+      ret = (strstr(s,"utf") || strstr(s,"UTF"));
+    }
+  }
+  return ret;
+}
+
+unsigned Fl_System_Driver::utf8to_mb(const char* src, unsigned srclen, char* dst, unsigned dstlen) {
+  wchar_t lbuf[1024];
+  wchar_t* buf = lbuf;
+  unsigned length = fl_utf8towc(src, srclen, buf, 1024);
+  int ret; // note: wcstombs() returns unsigned(length) or unsigned(-1)
+  if (length >= 1024) {
+    buf = (wchar_t*)(malloc((length+1)*sizeof(wchar_t)));
+    fl_utf8towc(src, srclen, buf, length+1);
+  }
+  if (dstlen) {
+    ret = wcstombs(dst, buf, dstlen);
+    if (ret >= (int)dstlen-1) ret = wcstombs(0,buf,0);
+  } else {
+    ret = wcstombs(0,buf,0);
+  }
+  if (buf != lbuf) free(buf);
+  if (ret >= 0) return (unsigned)ret;
+  // on any errors we return the UTF-8 as raw text...
+  if (srclen < dstlen) {
+    memcpy(dst, src, srclen);
+    dst[srclen] = 0;
+  } else {
+    // Buffer insufficent or buffer query
+  }
+  return srclen;
+}
+
+unsigned Fl_System_Driver::utf8from_mb(char* dst, unsigned dstlen, const char* src, unsigned srclen) {
+  wchar_t lbuf[1024];
+  wchar_t* buf = lbuf;
+  int length;
+  unsigned ret;
+  length = mbstowcs(buf, src, 1024);
+  if (length >= 1024) {
+    length = mbstowcs(0, src, 0)+1;
+    buf = (wchar_t*)(malloc(length*sizeof(wchar_t)));
+    mbstowcs(buf, src, length);
+  }
+  if (length >= 0) {
+    ret = fl_utf8fromwc(dst, dstlen, buf, length);
+    if (buf != lbuf) free(buf);
+    return ret;
+  }
+  // errors in conversion return the UTF-8 unchanged
+  if (srclen < dstlen) {
+    memcpy(dst, src, srclen);
+    dst[srclen] = 0;
+  } else {
+    // Buffer insufficent or buffer query
+  }
+  return srclen;
+}
+
 //
 // End of "$Id$".
 //