[liblouis-liblouisxml] [liblouis] r718 committed - UTF-8 supported; individual tables work, but make check doesn't
- From: liblouis@xxxxxxxxxxxxxx
- To: liblouis-liblouisxml@xxxxxxxxxxxxx
- Date: Thu, 05 Jul 2012 02:40:18 +0000
Revision: 718
Author: john.boyer@xxxxxxxxxxxxxxxxx
Date: Wed Jul 4 19:40:04 2012
Log: UTF-8 supported; individual tables work, but make check doesn't
http://code.google.com/p/liblouis/source/detail?r=718
Modified:
/trunk/liblouis/compileTranslationTable.c
=======================================
--- /trunk/liblouis/compileTranslationTable.c Mon Jul 2 13:04:46 2012
+++ /trunk/liblouis/compileTranslationTable.c Wed Jul 4 19:40:04 2012
@@ -223,7 +223,7 @@
CharsString;
static int errorCount;
-
+static int warningCount;
static TranslationTableHeader *table;
static TranslationTableOffset tableSize;
static TranslationTableOffset tableUsed;
@@ -688,12 +688,35 @@
#endif
va_end (arguments);
if (nested)
- lou_logPrint ("%s:%d: %s", nested->fileName, nested->lineNumber,
buffer);
+ lou_logPrint ("%s:%d: error: %s", nested->fileName,
+ nested->lineNumber, buffer);
else
- lou_logPrint ("%s", buffer);
+ lou_logPrint ("error: %s", buffer);
errorCount++;
#endif
}
+
+static void
+compileWarning (FileInfo * nested, char *format, ...)
+{
+#ifndef __SYMBIAN32__
+ char buffer[MAXSTRING];
+ va_list arguments;
+ va_start (arguments, format);
+#ifdef _WIN32
+ (void) _vsnprintf (buffer, sizeof (buffer), format, arguments);
+#else
+ (void) vsnprintf (buffer, sizeof (buffer), format, arguments);
+#endif
+ va_end (arguments);
+ if (nested)
+ lou_logPrint ("%s:%d: warning: %s", nested->fileName,
+ nested->lineNumber, buffer);
+ else
+ lou_logPrint ("warning: %s", buffer);
+ warningCount++;
+#endif
+}
static int
allocateSpaceInTable (FileInfo * nested, TranslationTableOffset * offset,
@@ -1436,70 +1459,65 @@
}
return (widechar) binaryValue;
}
+
+#define MAXBYTES 7
+static int first0Bit[MAXBYTES] = { 0x80, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC,
0XFE };
static int
parseChars (FileInfo * nested, CharsString * result, CharsString * token)
{
-/*interpret ruleChars string */
- int count = 0;
- int index;
- for (index = 0; index < token->length; index++)
- {
- widechar character = token->chars[index];
- if (character == '\\')
- { /* escape sequence */
- int ok = 0;
- if (++index < token->length)
- {
- switch (character = token->chars[index])
+ int in = 0;
+ int out = 0;
+ int lastOutSize = 0;
+ unsigned int ch = 0;
+ int numBytes = 0;
+ unsigned int utf32 = 0;
+ int k;
+ while (in < token->length)
+ {
+ ch = token->chars[in++] & 0xff;
+ if (ch < 128)
+ {
+ if (ch == '\\')
+ { /* escape sequence */
+ int ok = 1;
+ switch (ch = token->chars[in])
{
case '\\':
- ok = 1;
break;
case 'e':
- character = 0x1b;
- ok = 1;
+ ch = 0x1b;
break;
case 'f':
- character = 12;
- ok = 1;
+ ch = 12;
break;
case 'n':
- character = 10;
- ok = 1;
+ ch = 10;
break;
case 'r':
- character = 13;
- ok = 1;
+ ch = 13;
break;
case 's':
- character = ' ';
- ok = 1;
+ ch = ' ';
break;
case 't':
- character = 9;
- ok = 1;
+ ch = 9;
break;
case 'v':
- character = 22;
- ok = 1;
+ ch = 22;
break;
case 'w':
- character = ENDSEGMENT;
- ok = 1;
+ ch = ENDSEGMENT;
break;
case 34:
- character = QUOTESUB;
- ok = 1;
+ ch = QUOTESUB;
break;
case 'X':
case 'x':
- if (token->length - index > 4)
- {
- character =
- hexValue (nested, &token->chars[index + 1], 4);
- index += 4;
- ok = 1;
+ if (token->length - in > 4)
+ {
+ ch = hexValue (nested, &token->chars[in + 1], 4);
+ in += 4;
}
break;
case 'y':
@@ -1509,40 +1527,62 @@
not32:
compileError (nested,
"liblouis has not been compiled for 32-bit
Unicode");
+ ok = 0;
break;
}
- if (token->length - index > 5)
- {
- character =
- hexValue (nested, &token->chars[index + 1], 5);
- index += 5;
- ok = 1;
+ if (token->length - in > 5)
+ {
+ ch = hexValue (nested, &token->chars[in + 1], 5);
+ in += 5;
}
break;
case 'z':
case 'Z':
if (CHARSIZE == 2)
goto not32;
- if (token->length - index > 8)
- {
- character =
- hexValue (nested, &token->chars[index + 1], 8);
- index += 8;
- ok = 1;
- }
+ if (token->length - in > 8)
+ {
+ ch = hexValue (nested, &token->chars[in + 1], 8);
+ in += 8;
+ }
+ break;
+ default:
+ compileError (nested,
+ "invalid escape sequence '\\%c'", ch);
+ ok = 0;
break;
}
- }
- if (!ok)
- {
- index++;
- compileError (nested, "invalid escape sequence.");
- return 0;
- }
- }
- result->chars[count++] = character;
- }
- result->length = count;
+ in++;
+ }
+ result->chars[out++] = (widechar) ch;
+ if (out >= MAXSTRING)
+ {
+ result->length = out;
+ return 1;
+ }
+ continue;
+ }
+ lastOutSize = out;
+ for (numBytes = MAXBYTES - 1; numBytes >= 0; numBytes--)
+ if (ch >= first0Bit[numBytes])
+ break;
+ utf32 = ch & (0XFF - first0Bit[numBytes]);
+ for (k = 0; k < numBytes; k++)
+ {
+ if (in >= MAXSTRING)
+ break;
+ utf32 = (utf32 << 6) + (token->chars[in++] & 0x3f);
+ }
+ if (CHARSIZE == 2 && utf32 > 0xffff)
+ utf32 = 0xffff;
+ result->chars[out++] = (widechar) utf32;
+ if (out >= MAXSTRING)
+ {
+ result->length = lastOutSize;
+ return 1;
+ }
+ }
+ result->length = out;
return 1;
}
@@ -4684,6 +4724,7 @@
int listLength;
int currentListPos = 0;
errorCount = 0;
+ warningCount = 0;
fileCount = 0;
table = NULL;
characterClasses = NULL;
For a description of the software, to download it and links to
project pages go to http://www.abilitiessoft.com
Other related posts:
- » [liblouis-liblouisxml] [liblouis] r718 committed - UTF-8 supported; individual tables work, but make check doesn't - liblouis
