Mercurial > projects > dil

changeset 722:ceaac6a24258

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Added isUnicodeAlpha() for DDocParser and MacroParser.
author Aziz K?ksal <aziz.koeksal@gmail.com>
date Fri, 01 Feb 2008 19:44:00 +0100
parents 8955296dd807
children 5dd17d4568ce
files trunk/src/dil/Unicode.d trunk/src/dil/doc/Doc.d trunk/src/dil/doc/Macro.d
diffstat 3 files changed, 84 insertions(+), 144 deletions(-) [+]
line wrap: on
line diff
--- a/trunk/src/dil/Unicode.d	Fri Feb 01 16:40:35 2008 +0100
+++ b/trunk/src/dil/Unicode.d	Fri Feb 01 19:44:00 2008 +0100
@@ -42,25 +42,51 @@
   return (b & 0xC0) == 0xC0; // 11xx_xxxx
 }
 
+/// Advances ref_p only if this is a valid Unicode alpha character.
+bool isUnicodeAlpha(ref char* ref_p, char* end)
+in { assert(ref_p && ref_p < end); }
+body
+{
+  if (*ref_p < 0x80)
+    return false;
+  auto p = ref_p;
+  auto c = decode(p, end);
+  if (!isUniAlpha(c))
+    return false;
+  ref_p = p;
+  return true;
+}
+
+/// index is set one past the last trail byte of the valid UTF-8 sequence.
 dchar decode(char[] str, ref size_t index)
-in { assert(str.length); }
+in { assert(str.length && index < str.length); }
 out(c) { assert(isValidChar(c)); }
 body
 {
   char* p = str.ptr + index;
   char* end = str.ptr + str.length;
-  dchar c = *p;
+  dchar c = decode(p, end);
+  if (c != ERROR_CHAR)
+    index = p - str.ptr + 1;
+  return c;
+}
 
-  if (!(p < end))
-    return ERROR_CHAR;
+/// ref_p is set to the last trail byte of the valid UTF-8 sequence.
+dchar decode(ref char* ref_p, char* end)
+in { assert(ref_p && ref_p < end); }
+out(c) { assert(isValidChar(c)); }
+body
+{
+  char* p = ref_p;
+  dchar c = *p;
 
   if (c < 0x80)
   {
-    ++index;
+    ref_p++;
     return c;
   }
 
-  ++p; // Move to second byte.
+  p++; // Move to second byte.
   if (!(p < end))
     return ERROR_CHAR;
 
@@ -82,18 +108,16 @@
       return ERROR_CHAR;
   }
 
-  const char[] checkNextByte = "if (++p < end && !isTrailByte(*p))"
+  const char[] checkNextByte = "if (!(++p < end && isTrailByte(*p)))"
                                 "  return ERROR_CHAR;";
   const char[] appendSixBits = "c = (c << 6) | *p & 0b0011_1111;";
 
-  auto next_index = index;
   // Decode
   if ((c & 0b1110_0000) == 0b1100_0000)
   {
     // 110xxxxx 10xxxxxx
     c &= 0b0001_1111;
     mixin(appendSixBits);
-    next_index += 2;
   }
   else if ((c & 0b1111_0000) == 0b1110_0000)
   {
@@ -101,7 +125,6 @@
     c &= 0b0000_1111;
     mixin(appendSixBits ~
           checkNextByte ~ appendSixBits);
-    next_index += 3;
   }
   else if ((c & 0b1111_1000) == 0b1111_0000)
   {
@@ -110,7 +133,6 @@
     mixin(appendSixBits ~
           checkNextByte ~ appendSixBits ~
           checkNextByte ~ appendSixBits);
-    next_index += 4;
   }
   else
     // 5 and 6 byte UTF-8 sequences are not allowed yet.
@@ -122,7 +144,7 @@
 
   if (!isValidChar(c))
     return ERROR_CHAR;
-  index = next_index;
+  ref_p = p;
   return c;
 }
 
--- a/trunk/src/dil/doc/Doc.d	Fri Feb 01 16:40:35 2008 +0100
+++ b/trunk/src/dil/doc/Doc.d	Fri Feb 01 19:44:00 2008 +0100
@@ -11,22 +11,36 @@
 
 class DDocComment
 {
-  string text;
   Section[] sections;
   Section summary; /// Optional summary section.
   Section description; /// Optional description section.
 
-  this(string text)
+  this(Section[] sections, Section summary, Section description)
   {
-    assert(text.length && text[$-1] == '\0');
-    this.text = text;
+    this.sections = sections;
+    this.summary = summary;
+    this.description = description;
   }
+}
+
+struct DDocParser
+{
+  char* p;
+  char* textEnd;
+  Section[] sections;
+  Section summary; /// Optional summary section.
+  Section description; /// Optional description section.
 
   /// Parses the DDoc text into sections.
-  void parseSections()
+  Section[] parse(string text)
   {
-    char* p = text.ptr;
-    char* textEnd = p + text.length;
+    if (!text.length)
+      return null;
+    if (text[$-1] != '\0')
+      text ~= '\0';
+    p = text.ptr;
+    textEnd = p + text.length;
+
     char* summaryBegin;
     char* idBegin, idEnd;
     char* nextIdBegin, nextIdEnd;
@@ -34,16 +48,19 @@
     skipWhitespace(p);
     summaryBegin = p;
 
-    if (findNextIdColon(p, idBegin, idEnd))
+    if (findNextIdColon(idBegin, idEnd))
     { // Check that this is not an explicit section.
       if (summaryBegin != idBegin)
         scanSummaryAndDescription(summaryBegin, idBegin);
     }
     else // There are no explicit sections.
-      return scanSummaryAndDescription(summaryBegin, textEnd);
+    {
+      scanSummaryAndDescription(summaryBegin, textEnd);
+      return null;
+    }
 
     assert(idBegin && idEnd);
-    while (findNextIdColon(p, nextIdBegin, nextIdEnd))
+    while (findNextIdColon(nextIdBegin, nextIdEnd))
     {
       sections ~= new Section(makeString(idBegin, idEnd), makeString(idEnd+1, nextIdBegin));
       idBegin = nextIdBegin;
@@ -51,14 +68,15 @@
     }
     // Add last section.
     sections ~= new Section(makeString(idBegin, idEnd), makeString(idEnd+1, textEnd));
+    return sections;
   }
 
   void scanSummaryAndDescription(char* p, char* end)
   {
-    assert(p != end && p < end);
+    assert(p < end);
     char* sectionBegin = p;
     // Search for the end of the first paragraph.
-    while (p != end && !(*p == '\n' && p[1] == '\n'))
+    while (p < end && !(*p == '\n' && p[1] == '\n'))
       p++;
     // The first paragraph is the summary.
     summary = new Section("", makeString(sectionBegin, p));
@@ -88,23 +106,23 @@
   ///   idBegin = set to the first character of the Identifier
   ///   idEnd   = set to the colon following the Identifier
   /// Returns: true if found
-  bool findNextIdColon(ref char* ref_p, ref char* ref_idBegin, ref char* ref_idEnd)
+  bool findNextIdColon(ref char* ref_idBegin, ref char* ref_idEnd)
   {
-    auto p = ref_p;
+    auto p = this.p;
     while (*p != '\0')
     {
       auto idBegin = p;
       assert(isascii(*p) || isLeadByte(*p));
-      if (isidbeg(*p) || isUnicodeAlpha(p)) // IdStart
+      if (isidbeg(*p) || isUnicodeAlpha(p, textEnd)) // IdStart
       {
         do // IdChar*
           p++;
-        while (isident(*p) || isUnicodeAlpha(p))
+        while (isident(*p) || isUnicodeAlpha(p, textEnd))
         if (*p == ':') // :
         {
           ref_idBegin = idBegin;
           ref_idEnd = p;
-          ref_p = p;
+          this.p = p;
           return true;
         }
       }
@@ -118,65 +136,6 @@
     }
     return false;
   }
-
-  /// This function assumes that there are no invalid
-  /// UTF-8 sequences in the string.
-  bool isUnicodeAlpha(ref char* ref_p)
-  {
-    char* p = ref_p; // Copy.
-    if (isascii(*p))
-      return false;
-
-    dchar d = *p;
-    p++; // Move to second byte.
-    // Error if second byte is not a trail byte.
-    assert(isTrailByte(*p));
-    // Check for overlong sequences.
-    assert(delegate () {
-      switch (d)
-      {
-      case 0xE0, 0xF0, 0xF8, 0xFC:
-        if ((*p & d) == 0x80)
-          return false;
-      default:
-        if ((d & 0xFE) == 0xC0) // 1100000x
-          return false;
-        return true;
-      }
-    }() == true
-    );
-    const char[] checkNextByte = "p++;"
-                                 "assert(isTrailByte(*p));";
-    const char[] appendSixBits = "d = (d << 6) | *p & 0b0011_1111;";
-    // Decode
-    if ((d & 0b1110_0000) == 0b1100_0000)
-    {
-      d &= 0b0001_1111;
-      mixin(appendSixBits);
-    }
-    else if ((d & 0b1111_0000) == 0b1110_0000)
-    {
-      d &= 0b0000_1111;
-      mixin(appendSixBits ~
-            checkNextByte ~ appendSixBits);
-    }
-    else if ((d & 0b1111_1000) == 0b1111_0000)
-    {
-      d &= 0b0000_0111;
-      mixin(appendSixBits ~
-            checkNextByte ~ appendSixBits ~
-            checkNextByte ~ appendSixBits);
-    }
-    else
-      return false;
-
-    assert(isTrailByte(*p) && isValidChar(d));
-    if (!isUniAlpha(d))
-      return false;
-    // Only advance pointer if this is a Unicode alpha character.
-    ref_p = p;
-    return true;
-  }
 }
 
 class Section
--- a/trunk/src/dil/doc/Macro.d	Fri Feb 01 16:40:35 2008 +0100
+++ b/trunk/src/dil/doc/Macro.d	Fri Feb 01 19:44:00 2008 +0100
@@ -94,11 +94,11 @@
     {
       skipWhitespace(p);
       auto idBegin = p;
-      if (isidbeg(*p) || isUnicodeAlpha(p)) // IdStart
+      if (isidbeg(*p) || isUnicodeAlpha(p, textEnd)) // IdStart
       {
         do // IdChar*
           p++;
-        while (isident(*p) || isUnicodeAlpha(p))
+        while (isident(*p) || isUnicodeAlpha(p, textEnd))
         auto idEnd = p;
 
         skipWhitespace(p);
@@ -121,63 +121,22 @@
       p++;
     p++;
   }
-
-  bool isUnicodeAlpha(ref char* ref_p)
-  {
-    char* p = ref_p; // Copy.
-    if (isascii(*p))
-      return false;
-
-    dchar d = *p;
-    p++; // Move to second byte.
-    // Error if second byte is not a trail byte.
-    if (!isTrailByte(*p))
-      return false;
-    // Check for overlong sequences.
-    switch (d)
-    {
-    case 0xE0, 0xF0, 0xF8, 0xFC:
-      if ((*p & d) == 0x80)
-        return false;
-    default:
-      if ((d & 0xFE) == 0xC0) // 1100000x
-        return false;
-    }
-    const char[] checkNextByte = "if (!isTrailByte(*++p))"
-                                 "  return false;";
-    const char[] appendSixBits = "d = (d << 6) | *p & 0b0011_1111;";
-    // Decode
-    if ((d & 0b1110_0000) == 0b1100_0000)
-    {
-      d &= 0b0001_1111;
-      mixin(appendSixBits);
-    }
-    else if ((d & 0b1111_0000) == 0b1110_0000)
-    {
-      d &= 0b0000_1111;
-      mixin(appendSixBits ~
-            checkNextByte ~ appendSixBits);
-    }
-    else if ((d & 0b1111_1000) == 0b1111_0000)
-    {
-      d &= 0b0000_0111;
-      mixin(appendSixBits ~
-            checkNextByte ~ appendSixBits ~
-            checkNextByte ~ appendSixBits);
-    }
-    else
-      return false;
-
-    assert(isTrailByte(*p));
-    if (!isValidChar(d) || !isUniAlpha(d))
-      return false;
-    // Only advance pointer if this is a Unicode alpha character.
-    ref_p = p;
-    return true;
-  }
 }
 
 char[] makeString(char* begin, char* end)
 {
   return begin[0 .. end - begin];
 }
+
+char[] expandMacros(MacroTable table, char[] text)
+{
+  char[] result;
+  char* p = text.ptr;
+  char* textEnd = p + text.length;
+//   while (p < text.length)
+//   {
+
+//   }
+  return result;
+}
+