static void eatComment (lexingState * st)
{
boolean unfinished = TRUE;
boolean lastIsStar = FALSE;
const unsigned char *c = st->cp + 2;
while (unfinished)
{
/* we've reached the end of the line..
* so we have to reload a line... */
if (c == NULL || *c == '\0')
{
st->cp = fileReadLine ();
/* WOOPS... no more input...
* we return, next lexing read
* will be null and ok */
if (st->cp == NULL)
return;
c = st->cp;
}
/* we've reached the end of the comment */
else if (*c == ')' && lastIsStar)
unfinished = FALSE;
/* here we deal with imbricated comment, which
* are allowed in OCaml */
else if (c[0] == '(' && c[1] == '*')
{
st->cp = c;
eatComment (st);
c = st->cp;
if (c == NULL)
return;
lastIsStar = FALSE;
c++;
}
/* OCaml has a rule which says :
*
* "Comments do not occur inside string or character literals.
* Nested comments are handled correctly."
*
* So if we encounter a string beginning, we must parse it to
* get a good comment nesting (bug ID: 3117537)
*/
else if (*c == '"')
{
st->cp = c;
eatString (st);
c = st->cp;
}
else
{
lastIsStar = '*' == *c;
c++;
}
}
st->cp = c;
}
static Value eatObject(std::istream& stream)
{
Value obj(Type::objectValue);
while (!stream.eof())
{
ltrim(stream);
char token = stream.get();
if (token == '}')
break; //empty object
if (token == '/')
{
eatComment(stream);
token = stream.get();
}
assert(token == '"');
Value key = eatString(stream);
ltrim(stream);
token = stream.get();
assert(token == ':');
ltrim(stream);
Value val = eatValue(stream);
obj[key.asString()] = val;
ltrim(stream);
token = stream.get();
if (token == '}')
break;
if (token != ',')
throw "arg";
assert(token == ',');
}
return obj;
};
void eatIgnored() {
bool stop(false);
while (!stop){
while (isIgnored(pickChar())) eatChar();
if (_ignore_flag == ignoreFlag::None)
return;
stop = true;
if (_ignore_flag == ignoreFlag::CppC) {
if (pickString(sizeof("//") - 1) == "//") { eatString(sizeof("//") - 1);
while (stream_cursor != stream_buffer.end() || pickChar() != '\n') eatChar();
stop = false;
}
if (pickString(sizeof("/*") - 1) == "/*") {eatString(sizeof("/*") - 1);
while (pickString(sizeof("*/")-1) != "*/") eatChar();
eatString(sizeof("/*") - 1);
stop = false;
}
}
}
}
std::string eatAndConvertString(void)
{
std::string res;
auto str = eatString();
if (str.second)
{
v8::String::Utf8Value val(str.first);
res = std::string(*val, val.length());
}
return res;
}
static Value eatValue(std::istream& stream)
{
ltrim(stream);
char token = stream.get();
if (token == '{')
return eatObject(stream);
if (token == '[')
return eatArray(stream);
if ((token >= '0' && token <= '9') || token == '.' || token == '-')
return eatNumeric(stream, token);
if (token == '"')
return eatString(stream);
if (token == 't' || token == 'f')
return eatBool(stream);
if (token == 'n')
return eatNull(stream);
if (token == '/')
{
eatComment(stream);
return eatValue(stream);
}
throw "Unable to parse json";
};
/* The lexer is in charge of reading the file.
* Some of sub-lexer (like eatComment) also read file.
* lexing is finished when the lexer return Tok_EOF */
static objcKeyword lex (lexingState * st)
{
int retType;
/* handling data input here */
while (st->cp == NULL || st->cp[0] == '\0')
{
st->cp = readLineFromInputFile ();
if (st->cp == NULL)
return Tok_EOF;
return Tok_EOL;
}
if (isAlpha (*st->cp) || (*st->cp == '_'))
{
readIdentifier (st);
retType = lookupKeyword (vStringValue (st->name), Lang_ObjectiveC);
if (retType == -1) /* If it's not a keyword */
{
return ObjcIDENTIFIER;
}
else
{
return retType;
}
}
else if (*st->cp == '@')
{
readIdentifierObjcDirective (st);
retType = lookupKeyword (vStringValue (st->name), Lang_ObjectiveC);
if (retType == -1) /* If it's not a keyword */
{
return Tok_any;
}
else
{
return retType;
}
}
else if (isSpace (*st->cp))
{
eatWhiteSpace (st);
return lex (st);
}
else
switch (*st->cp)
{
case '(':
st->cp++;
return Tok_PARL;
case '\\':
st->cp++;
return Tok_Backslash;
case '#':
st->cp++;
return Tok_Sharp;
case '/':
if (st->cp[1] == '*') /* ergl, a comment */
{
eatComment (st);
return lex (st);
}
else if (st->cp[1] == '/')
{
st->cp = NULL;
return lex (st);
}
else
{
st->cp++;
return Tok_any;
}
break;
case ')':
st->cp++;
return Tok_PARR;
case '{':
st->cp++;
return Tok_CurlL;
case '}':
st->cp++;
return Tok_CurlR;
case '[':
st->cp++;
return Tok_SQUAREL;
case ']':
st->cp++;
return Tok_SQUARER;
case ',':
st->cp++;
return Tok_COMA;
case ';':
st->cp++;
return Tok_semi;
case ':':
st->cp++;
return Tok_dpoint;
case '"':
eatString (st);
return Tok_any;
case '+':
st->cp++;
return Tok_PLUS;
case '-':
st->cp++;
return Tok_MINUS;
case '*':
st->cp++;
return Tok_Asterisk;
case '<':
st->cp++;
return Tok_ANGLEL;
case '>':
st->cp++;
return Tok_ANGLER;
default:
st->cp++;
break;
}
/* default return if nothing is recognized,
* shouldn't happen, but at least, it will
* be handled without destroying the parsing. */
return Tok_any;
}
/* The lexer is in charge of reading the file.
* Some of sub-lexer (like eatComment) also read file.
* lexing is finished when the lexer return Tok_EOF */
static ocamlKeyword lex (lexingState * st)
{
int retType;
/* handling data input here */
while (st->cp == NULL || st->cp[0] == '\0')
{
st->cp = fileReadLine ();
if (st->cp == NULL)
return Tok_EOF;
}
if (isAlpha (*st->cp))
{
readIdentifier (st);
retType = lookupKeyword (vStringValue (st->name), Lang_Ocaml);
if (retType == -1) /* If it's not a keyword */
{
return OcaIDENTIFIER;
}
else
{
return retType;
}
}
else if (isNum (*st->cp))
return eatNumber (st);
else if (isSpace (*st->cp))
{
eatWhiteSpace (st);
return lex (st);
}
/* OCaml permit the definition of our own operators
* so here we check all the consecuting chars which
* are operators to discard them. */
else if (isOperator[*st->cp])
return eatOperator (st);
else
switch (*st->cp)
{
case '(':
if (st->cp[1] == '*') /* ergl, a comment */
{
eatComment (st);
return lex (st);
}
else
{
st->cp++;
return Tok_PARL;
}
case ')':
st->cp++;
return Tok_PARR;
case '[':
st->cp++;
return Tok_BRL;
case ']':
st->cp++;
return Tok_BRR;
case '{':
st->cp++;
return Tok_CurlL;
case '}':
st->cp++;
return Tok_CurlR;
case '\'':
st->cp++;
return Tok_Prime;
case ',':
st->cp++;
return Tok_comma;
case '=':
st->cp++;
return Tok_EQ;
case ';':
st->cp++;
return Tok_semi;
case '"':
eatString (st);
return Tok_Val;
case '_':
st->cp++;
return Tok_Val;
case '#':
st->cp++;
return Tok_Sharp;
case '\\':
st->cp++;
return Tok_Backslash;
default:
st->cp++;
break;
}
/* default return if nothing is recognized,
* shouldn't happen, but at least, it will
* be handled without destroying the parsing. */
return Tok_Val;
}
String cursorToString(CXCursor cursor, unsigned flags)
{
const CXCursorKind kind = clang_getCursorKind(cursor);
String ret;
ret.reserve(256);
ret += eatString(clang_getCursorKindSpelling(kind));
if (clang_isInvalid(kind))
return ret;
switch (RTags::cursorType(kind)) {
case Reference:
ret += " r";
break;
case Cursor:
ret += " c";
break;
case Other:
ret += " o";
break;
case Include:
ret += " i";
break;
}
const String name = eatString(clang_getCursorDisplayName(cursor));
const String other = eatString(clang_getCursorSpelling(cursor));
if (!name.isEmpty())
ret += " " + name;
if (other != name && !other.isEmpty())
ret += " " + other;
if (clang_isCursorDefinition(cursor))
ret += " def";
if (flags & IncludeUSR)
ret += " " + eatString(clang_getCursorUSR(cursor));
CXFile file;
unsigned off, line, col; //presumedLine, presumedCol, instantiationLoc, expansionLoc;
CXSourceLocation location = clang_getCursorLocation(cursor);
clang_getSpellingLocation(location, &file, &line, &col, &off);
// clang_getPresumedLocation(location, 0, &presumedLine, &presumedCol);
// clang_getInstantiationLocation(location, 0, 0, 0, &instantiationLoc);
// clang_getExpansionLocation(location, 0, 0, 0, &expansionLoc);
const Str fileName(clang_getFileName(file));
if (fileName.data() && *fileName.data()) {
ret += ' ';
ret += fileName.data();
ret += ',';
ret += String::number(off);
if (flags & IncludeRange) {
ret += " (";
CXSourceRange range = clang_getCursorExtent(cursor);
unsigned start, end;
clang_getSpellingLocation(clang_getRangeStart(range), 0, 0, 0, &start);
clang_getSpellingLocation(clang_getRangeEnd(range), 0, 0, 0, &end);
ret += String::number(start);
ret += '-';
ret += String::number(end);
ret += ')';
}
// if (presumedLine != line || presumedCol != col)
// ret += String::snprintf<32>("presumed: %d:%d", presumedLine, presumedCol);
// if (instantiationLoc != off)
// ret += String::snprintf<32>("instantiation: %d", instantiationLoc);
// if (expansionLoc != off)
// ret += String::snprintf<32>("expansion: %d", expansionLoc);
}
return ret;
}