--- trunk/instrument-js.c	2008/09/20 23:27:14	174
+++ trunk/instrument-js.c	2008/10/03 02:26:04	214
@@ -35,6 +35,8 @@
 #include <jsstr.h>
 
 #include "encoding.h"
+#include "global.h"
+#include "highlight.h"
 #include "resource-manager.h"
 #include "util.h"
 
@@ -48,6 +50,7 @@
 */
 static const char * file_id = NULL;
 static char * lines = NULL;
+static uint16_t num_lines = 0;
 
 void jscoverage_init(void) {
   runtime = JS_NewRuntime(8L * 1024L * 1024L);
@@ -872,7 +875,11 @@
 */
 static void instrument_statement(JSParseNode * node, Stream * f, int indent) {
   if (node->pn_type != TOK_LC) {
-    int line = node->pn_pos.begin.lineno;
+    uint16_t line = node->pn_pos.begin.lineno;
+    if (line > num_lines) {
+      fatal("%s: script contains more than 65,535 lines", file_id);
+    }
+
     /* the root node has line number 0 */
     if (line != 0) {
       Stream_printf(f, "%*s", indent, "");
@@ -883,20 +890,30 @@
   output_statement(node, f, indent);
 }
 
-void jscoverage_instrument_js(const char * id, const char * encoding, Stream * input, Stream * output) {
+static bool characters_start_with(const jschar * characters, size_t line_start, size_t line_end, const char * prefix) {
+  const jschar * characters_end = characters + line_end;
+  const jschar * cp = characters + line_start;
+  const char * bp = prefix;
+  for (;;) {
+    if (*bp == '\0') {
+      return true;
+    }
+    else if (cp == characters_end) {
+      return false;
+    }
+    else if (*cp != *bp) {
+      return false;
+    }
+    bp++;
+    cp++;
+  }
+}
+
+void jscoverage_instrument_js(const char * id, const uint16_t * characters, size_t num_characters, Stream * output) {
   file_id = id;
 
   /* scan the javascript */
-  size_t input_length = input->length;
-  jschar * base = NULL;
-  int result = jscoverage_bytes_to_characters(encoding, input->data, input->length, &base, &input_length);
-  if (result == JSCOVERAGE_ERROR_ENCODING_NOT_SUPPORTED) {
-    fatal("encoding %s not supported in file %s", encoding, id);
-  }
-  else if (result == JSCOVERAGE_ERROR_INVALID_BYTE_SEQUENCE) {
-    fatal("error decoding %s in file %s", encoding, id);
-  }
-  JSTokenStream * token_stream = js_NewTokenStream(context, base, input_length, NULL, 1, NULL);
+  JSTokenStream * token_stream = js_NewTokenStream(context, characters, num_characters, NULL, 1, NULL);
   if (token_stream == NULL) {
     fatal("cannot create token stream from file: %s", file_id);
   }
@@ -906,7 +923,7 @@
   if (node == NULL) {
     fatal("parse error in file: %s", file_id);
   }
-  int num_lines = node->pn_pos.end.lineno;
+  num_lines = node->pn_pos.end.lineno;
   lines = xmalloc(num_lines);
   for (int i = 0; i < num_lines; i++) {
     lines[i] = 0;
@@ -939,89 +956,209 @@
 
   /* copy the instrumented source code to the output */
   Stream_write(output, instrumented->data, instrumented->length);
-  Stream_write_char(output, '\n');
 
   /* conditionals */
   bool has_conditionals = false;
   size_t line_number = 0;
   size_t i = 0;
-  while (i < input_length) {
+  while (i < num_characters) {
     line_number++;
     size_t line_start = i;
-    while (i < input_length && base[i] != '\r' && base[i] != '\n') {
-      i++;
-    }
-    size_t line_end = i;
-    if (i < input_length) {
-      if (base[i] == '\r') {
-        line_end = i;
+    jschar c;
+    bool done = false;
+    while (! done && i < num_characters) {
+      c = characters[i];
+      switch (c) {
+      case '\r':
+      case '\n':
+      case 0x2028:
+      case 0x2029:
+        done = true;
+        break;
+      default:
         i++;
-        if (i < input_length && base[i] == '\n') {
-          i++;
-        }
+        break;
       }
-      else if (base[i] == '\n') {
-        line_end = i;
+    }
+    size_t line_end = i;
+    if (i < num_characters) {
+      i++;
+      if (c == '\r' && i < num_characters && characters[i] == '\n') {
         i++;
       }
-      else {
-        abort();
-      }
     }
-    char * line = js_DeflateString(context, base + line_start, line_end - line_start);
-    if (str_starts_with(line, "//#JSCOVERAGE_IF")) {
+    if (characters_start_with(characters, line_start, line_end, "//#JSCOVERAGE_IF")) {
       if (! has_conditionals) {
         has_conditionals = true;
         Stream_printf(output, "_$jscoverage['%s'].conditionals = [];\n", file_id);
       }
-      Stream_printf(output, "if (!%s) {\n", line + 16);
+      Stream_write_string(output, "if (!(");
+      for (size_t j = line_start + 16; j < line_end; j++) {
+        jschar c = characters[j];
+        if (c == '\t' || (32 <= c && c <= 126)) {
+          Stream_write_char(output, c);
+        }
+        else {
+          Stream_printf(output, "\\u%04x", c);
+        }
+      }
+      Stream_write_string(output, ")) {\n");
       Stream_printf(output, "  _$jscoverage['%s'].conditionals[%d] = ", file_id, line_number);
     }
-    else if (str_starts_with(line, "//#JSCOVERAGE_ENDIF")) {
+    else if (characters_start_with(characters, line_start, line_end, "//#JSCOVERAGE_ENDIF")) {
       Stream_printf(output, "%d;\n", line_number);
       Stream_printf(output, "}\n");
     }
-    JS_free(context, line);
   }
 
   /* copy the original source to the output */
-  i = 0;
-  while (i < input_length) {
-    Stream_write_string(output, "// ");
-    size_t line_start = i;
-    while (i < input_length && base[i] != '\r' && base[i] != '\n') {
-      i++;
-    }
+  Stream_printf(output, "_$jscoverage['%s'].source = ", file_id);
+  jscoverage_write_source(id, characters, num_characters, output);
+  Stream_printf(output, ";\n");
 
-    size_t line_end = i;
-    if (i < input_length) {
-      if (base[i] == '\r') {
-        line_end = i;
+  Stream_delete(instrumented);
+
+  file_id = NULL;
+}
+
+void jscoverage_write_source(const char * id, const jschar * characters, size_t num_characters, Stream * output) {
+  Stream_write_string(output, "[");
+  if (jscoverage_highlight) {
+    Stream * highlighted_stream = Stream_new(num_characters);
+    jscoverage_highlight_js(context, id, characters, num_characters, highlighted_stream);
+    size_t i = 0;
+    while (i < highlighted_stream->length) {
+      if (i > 0) {
+        Stream_write_char(output, ',');
+      }
+
+      Stream_write_char(output, '"');
+      bool done = false;
+      while (! done) {
+        char c = highlighted_stream->data[i];
+        switch (c) {
+        case 0x8:
+          /* backspace */
+          Stream_write_string(output, "\\b");
+          break;
+        case 0x9:
+          /* horizontal tab */
+          Stream_write_string(output, "\\t");
+          break;
+        case 0xa:
+          /* line feed (new line) */
+          done = true;
+          break;
+        case 0xb:
+          /* vertical tab */
+          Stream_write_string(output, "\\v");
+          break;
+        case 0xc:
+          /* form feed */
+          Stream_write_string(output, "\\f");
+          break;
+        case 0xd:
+          /* carriage return */
+          done = true;
+          if (i + 1 < highlighted_stream->length && highlighted_stream->data[i + 1] == '\n') {
+            i++;
+          }
+          break;
+        case '"':
+          Stream_write_string(output, "\\\"");
+          break;
+        case '\\':
+          Stream_write_string(output, "\\\\");
+          break;
+        default:
+          Stream_write_char(output, c);
+          break;
+        }
         i++;
-        if (i < input_length && base[i] == '\n') {
-          i++;
+        if (i >= highlighted_stream->length) {
+          done = true;
         }
       }
-      else if (base[i] == '\n') {
-        line_end = i;
+      Stream_write_char(output, '"');
+    }
+    Stream_delete(highlighted_stream);
+  }
+  else {
+    size_t i = 0;
+    while (i < num_characters) {
+      if (i > 0) {
+        Stream_write_char(output, ',');
+      }
+
+      Stream_write_char(output, '"');
+      bool done = false;
+      while (! done) {
+        jschar c = characters[i];
+        switch (c) {
+        case 0x8:
+          /* backspace */
+          Stream_write_string(output, "\\b");
+          break;
+        case 0x9:
+          /* horizontal tab */
+          Stream_write_string(output, "\\t");
+          break;
+        case 0xa:
+          /* line feed (new line) */
+          done = true;
+          break;
+        case 0xb:
+          /* vertical tab */
+          Stream_write_string(output, "\\v");
+          break;
+        case 0xc:
+          /* form feed */
+          Stream_write_string(output, "\\f");
+          break;
+        case 0xd:
+          /* carriage return */
+          done = true;
+          if (i + 1 < num_characters && characters[i + 1] == '\n') {
+            i++;
+          }
+          break;
+        case '"':
+          Stream_write_string(output, "\\\"");
+          break;
+        case '\\':
+          Stream_write_string(output, "\\\\");
+          break;
+        case '&':
+          Stream_write_string(output, "&amp;");
+          break;
+        case '<':
+          Stream_write_string(output, "&lt;");
+          break;
+        case '>':
+          Stream_write_string(output, "&gt;");
+          break;
+        case 0x2028:
+        case 0x2029:
+          done = true;
+          break;
+        default:
+          if (32 <= c && c <= 126) {
+            Stream_write_char(output, c);
+          }
+          else {
+            Stream_printf(output, "&#%d;", c);
+          }
+          break;
+        }
         i++;
+        if (i >= num_characters) {
+          done = true;
+        }
       }
-      else {
-        abort();
-      }
+      Stream_write_char(output, '"');
     }
-
-    char * line = js_DeflateString(context, base + line_start, line_end - line_start);
-    Stream_write_string(output, line);
-    Stream_write_char(output, '\n');
-    JS_free(context, line);
   }
-
-  Stream_delete(instrumented);
-
-  JS_free(context, base);
-
-  file_id = NULL;
+  Stream_write_string(output, "]");
 }
 
 void jscoverage_copy_resources(const char * destination_directory) {
@@ -1030,9 +1167,7 @@
   copy_resource("jscoverage.js", destination_directory);
   copy_resource("jscoverage-ie.css", destination_directory);
   copy_resource("jscoverage-throbber.gif", destination_directory);
-  copy_resource("jscoverage-sh_main.js", destination_directory);
-  copy_resource("jscoverage-sh_javascript.js", destination_directory);
-  copy_resource("jscoverage-sh_nedit.css", destination_directory);
+  copy_resource("jscoverage-highlight.css", destination_directory);
 }
 
 /*
@@ -1067,8 +1202,13 @@
   JS_HashTableDestroy(coverage->coverage_table);
   struct FileCoverageList * p = coverage->coverage_list;
   while (p != NULL) {
-    free(p->file_coverage->lines);
-    free(p->file_coverage->source);
+    free(p->file_coverage->coverage_lines);
+    if (p->file_coverage->source_lines != NULL) {
+      for (uint32 i = 0; i < p->file_coverage->num_source_lines; i++) {
+        free(p->file_coverage->source_lines[i]);
+      }
+      free(p->file_coverage->source_lines);
+    }
     free(p->file_coverage->id);
     free(p->file_coverage);
     struct FileCoverageList * q = p;
@@ -1190,7 +1330,7 @@
         }
         else if (strcmp(s, "source") == 0) {
           source = element->pn_right;
-          if (source->pn_type != TOK_STRING || ! ATOM_IS_STRING(source->pn_atom)) {
+          if (source->pn_type != TOK_RB) {
             return -1;
           }
         }
@@ -1214,23 +1354,32 @@
       char * id = xstrdup(id_bytes);
       file_coverage = xmalloc(sizeof(FileCoverage));
       file_coverage->id = id;
-      file_coverage->num_lines = array->pn_count - 1;
-      file_coverage->lines = xnew(int, array->pn_count);
+      file_coverage->num_coverage_lines = array->pn_count;
+      file_coverage->coverage_lines = xnew(int, array->pn_count);
       if (source == NULL) {
-        file_coverage->source = NULL;
+        file_coverage->source_lines = NULL;
       }
       else {
-        file_coverage->source = xstrdup(JS_GetStringBytes(ATOM_TO_STRING(source->pn_atom)));
+        file_coverage->num_source_lines = source->pn_count;
+        file_coverage->source_lines = xnew(char *, source->pn_count);
+        uint32 i = 0;
+        for (JSParseNode * element = source->pn_head; element != NULL; element = element->pn_next, i++) {
+          if (element->pn_type != TOK_STRING) {
+            return -1;
+          }
+          file_coverage->source_lines[i] = xstrdup(JS_GetStringBytes(ATOM_TO_STRING(element->pn_atom)));
+        }
+        assert(i == source->pn_count);
       }
 
       /* set coverage for all lines */
       uint32 i = 0;
       for (JSParseNode * element = array->pn_head; element != NULL; element = element->pn_next, i++) {
         if (element->pn_type == TOK_NUMBER) {
-          file_coverage->lines[i] = (int) element->pn_dval;
+          file_coverage->coverage_lines[i] = (int) element->pn_dval;
         }
         else if (element->pn_type == TOK_PRIMARY && element->pn_op == JSOP_NULL) {
-          file_coverage->lines[i] = -1;
+          file_coverage->coverage_lines[i] = -1;
         }
         else {
           return -1;
@@ -1248,7 +1397,7 @@
     else {
       /* sanity check */
       assert(strcmp(file_coverage->id, id_bytes) == 0);
-      if (file_coverage->num_lines != array->pn_count - 1) {
+      if (file_coverage->num_coverage_lines != array->pn_count) {
         return -2;
       }
 
@@ -1256,13 +1405,13 @@
       uint32 i = 0;
       for (JSParseNode * element = array->pn_head; element != NULL; element = element->pn_next, i++) {
         if (element->pn_type == TOK_NUMBER) {
-          if (file_coverage->lines[i] == -1) {
+          if (file_coverage->coverage_lines[i] == -1) {
             return -2;
           }
-          file_coverage->lines[i] += (int) element->pn_dval;
+          file_coverage->coverage_lines[i] += (int) element->pn_dval;
         }
         else if (element->pn_type == TOK_PRIMARY && element->pn_op == JSOP_NULL) {
-          if (file_coverage->lines[i] != -1) {
+          if (file_coverage->coverage_lines[i] != -1) {
             return -2;
           }
         }
@@ -1273,8 +1422,17 @@
       assert(i == array->pn_count);
 
       /* if this JSON file has source, use it */
-      if (file_coverage->source == NULL && source != NULL) {
-        file_coverage->source = xstrdup(JS_GetStringBytes(ATOM_TO_STRING(source->pn_atom)));
+      if (file_coverage->source_lines == NULL && source != NULL) {
+        file_coverage->num_source_lines = source->pn_count;
+        file_coverage->source_lines = xnew(char *, source->pn_count);
+        uint32 i = 0;
+        for (JSParseNode * element = source->pn_head; element != NULL; element = element->pn_next, i++) {
+          if (element->pn_type != TOK_STRING) {
+            return -1;
+          }
+          file_coverage->source_lines[i] = xstrdup(JS_GetStringBytes(ATOM_TO_STRING(element->pn_atom)));
+        }
+        assert(i == source->pn_count);
       }
     }
   }