de-escape backslashes in depfiles while parsing

1 files changed, 62 insertions, 34 deletions

diff --git a/src/depfile_parser.in.cc b/src/depfile_parser.in.cc
index 7ac95c6..c469a2c 100644
--- a/src/depfile_parser.in.cc
+++ b/src/depfile_parser.in.cc
@@ -23,51 +23,79 @@
 // How do you end a line with a backslash?  The netbsd Make docs suggest
 // reading the result of a shell command echoing a backslash!
 //
-// Rather than implement the above, we do the simpler thing here.
+// Rather than implement all of above, we do a simpler thing here:
+// Backslashes escape a set of characters (see "escapes" defined below),
+// otherwise they are passed through verbatim.
 // If anyone actually has depfiles that rely on the more complicated
 // behavior we can adjust this.
 bool DepfileParser::Parse(string* content, string* err) {
-  char* p = &(*content)[0];
-  char* end = p + content->size();
-  for (;;) {
-    const char* start = p;
-    char yych;
-    /*!re2c
-    re2c:define:YYCTYPE = "char";
-    re2c:define:YYCURSOR = p;
-    re2c:define:YYLIMIT = end;
+  // in: current parser input point.
+  // end: end of input.
+  char* in = &(*content)[0];
+  char* end = in + content->size();
+  while (in < end) {
+    // out: current output point (typically same as in, but can fall behind
+    // as we de-escape backslashes).
+    char* out = in;
+    // filename: start of the current parsed filename.
+    char* filename = out;
+    for (;;) {
+      // start: beginning of the current parsed span.
+      const char* start = in;
+      char yych;
+      /*!re2c
+      re2c:define:YYCTYPE = "char";
+      re2c:define:YYCURSOR = in;
+      re2c:define:YYLIMIT = end;
 
-    re2c:yyfill:parameter = 0;
-    re2c:define:YYFILL = break;
+      re2c:yyfill:enable = 0;
 
-    re2c:indent:top = 2;
-    re2c:indent:string = "  ";
+      re2c:indent:top = 2;
+      re2c:indent:string = "  ";
 
-    re2c:yych:emit = 0;
+      re2c:yych:emit = 0;
 
-    '\\\n' { continue; }
-    [ \n]+ { continue; }
-    [a-zA-Z0-9+,/\\_:.-]+ {
-      // Got a filename.
-      int len = p - start;
-      if (start[len - 1] == ':')
-        len--;  // Strip off trailing colon, if any.
+      escape = [ \\#*$[|];
 
-      if (len == 0)
-        continue;  // Drop isolated colons.
-
-      if (!out_.str_) {
-        out_ = StringPiece(start, len);
-      } else {
-        ins_.push_back(StringPiece(start, len));
+      '\\' escape {
+        // De-escape backslashed character.
+        *out++ = yych;
+        continue;
       }
-      continue;
+      '\\'. {
+        // Let backslash before other characters through verbatim.
+        *out++ = '\\';
+        *out++ = yych;
+        continue;
+      }
+      [a-zA-Z0-9+,/_:.-]+ {
+        // Got a span of plain text.  Copy it to out if necessary.
+        int len = in - start;
+        if (out < start)
+          memmove(out, start, len);
+        out += len;
+        continue;
+      }
+      [^] {
+        // For any other character (e.g. whitespace), swallow it here,
+        // allowing the outer logic to loop around again.
+        break;
+      }
+      */
     }
-    [^] {
-      *err = "BUG: depfile lexer encountered unknown state";
-      return false;
+
+    int len = out - filename;
+    if (len > 0 && filename[len - 1] == ':')
+      len--;  // Strip off trailing colon, if any.
+
+    if (len == 0)
+      continue;
+
+    if (!out_.str_) {
+      out_ = StringPiece(filename, len);
+    } else {
+      ins_.push_back(StringPiece(filename, len));
     }
-    */
   }
   return true;
 }