Function IsAnchorStart

extern/re2/re2/compile.cc:965–1007

Is this regexp required to start at the beginning of the text? Only approximate; can return false for complicated regexps like (\Aa|\Ab), but handles (\A(a|b)). Could use the Walker to write a more exact one.

// Only approximate; can return false for complicated regexps like (\Aa|\Ab),
// but handles (\A(a|b)).  Could use the Walker to write a more exact one.
static bool IsAnchorStart(Regexp** pre, int depth) {
  Regexp* re = *pre;
  Regexp* sub;
  // The depth limit makes sure that we don't overflow
  // the stack on a deeply nested regexp.  As the comment
  // above says, IsAnchorStart is conservative, so returning
  // a false negative is okay.  The exact limit is somewhat arbitrary.
  if (re == NULL || depth >= 4)
    return false;
  switch (re->op()) {
    default:
      break;
    case kRegexpConcat:
      if (re->nsub() > 0) {
        sub = re->sub()[0]->Incref();
        if (IsAnchorStart(&sub, depth+1)) {
          PODArray<Regexp*> subcopy(re->nsub());
          subcopy[0] = sub;  // already have reference
          for (int i = 1; i < re->nsub(); i++)
            subcopy[i] = re->sub()[i]->Incref();
          *pre = Regexp::Concat(subcopy.data(), re->nsub(), re->parse_flags());
          re->Decref();
          return true;
        }
        sub->Decref();
      }
      break;
    case kRegexpCapture:
      sub = re->sub()[0]->Incref();
      if (IsAnchorStart(&sub, depth+1)) {
        *pre = Regexp::Capture(sub, re->parse_flags(), re->cap());
        re->Decref();
        return true;
      }
      sub->Decref();
      break;
    case kRegexpBeginText:
      *pre = Regexp::LiteralString(NULL, 0, re->parse_flags());
      re->Decref();
      return true;
  }
  return false;
}

// Is this regexp required to start at the end of the text?
// Only approximate; can return false for complicated regexps like (a\z|b\z),