/*
* Returns a string describing the scope in 'list'.
* We record the current scope as a list of entered scopes.
* Scopes corresponding to 'if' statements and the like are
* represented by empty strings. Scopes corresponding to
* modules and classes are represented by the name of the
* module or class.
*/
static vString* stringListToScope (const stringList* list)
{
unsigned int i;
unsigned int chunks_output = 0;
vString* result = vStringNew ();
const unsigned int max = stringListCount (list);
for (i = 0; i < max; ++i)
{
vString* chunk = stringListItem (list, i);
if (vStringLength (chunk) > 0)
{
vStringCatS (result, (chunks_output++ > 0) ? "." : "");
vStringCatS (result, vStringValue (chunk));
}
}
return result;
}
static void makeVhdlTag (tokenInfo * const token, const vhdlKind kind)
{
if (VhdlKinds[kind].enabled)
{
/*
* If a scope has been added to the token, change the token
* string to include the scope when making the tag.
*/
if (vStringLength (token->scope) > 0)
{
vString *fulltag = vStringNew ();
vStringCopy (fulltag, token->scope);
vStringCatS (fulltag, ".");
vStringCatS (fulltag, vStringValue (token->string));
vStringTerminate (fulltag);
vStringCopy (token->string, fulltag);
vStringDelete (fulltag);
}
makeConstTag (token, kind);
}
}
CXXToken * cxxTokenCreateKeyword(int iLineNumber,MIOPos oFilePosition,enum CXXKeyword eKeyword)
{
CXXToken * pToken = cxxTokenCreate();
pToken->iLineNumber = iLineNumber;
pToken->oFilePosition = oFilePosition;
pToken->eType = CXXTokenTypeKeyword;
pToken->eKeyword = eKeyword;
pToken->bFollowedBySpace = TRUE;
vStringCatS(pToken->pszWord,cxxKeywordName(eKeyword));
return pToken;
}
static boolean recurseIntoDirectory (const char *const dirName)
{
boolean resize = FALSE;
if (isRecursiveLink (dirName))
verbose ("ignoring \"%s\" (recursive link)\n", dirName);
else if (! Option.recurse)
verbose ("ignoring \"%s\" (directory)\n", dirName);
else
{
verbose ("RECURSING into directory \"%s\"\n", dirName);
#if defined (HAVE_OPENDIR)
resize = recurseUsingOpendir (dirName);
#elif defined (HAVE_FINDFIRST) || defined (HAVE__FINDFIRST)
{
vString *const pattern = vStringNew ();
vStringCopyS (pattern, dirName);
vStringPut (pattern, OUTPUT_PATH_SEPARATOR);
vStringCatS (pattern, "*.*");
resize = createTagsForWildcardUsingFindfirst (vStringValue (pattern));
vStringDelete (pattern);
}
#elif defined (AMIGA)
{
vString *const pattern = vStringNew ();
if (*dirName != '\0' && strcmp (dirName, ".") != 0)
{
vStringCopyS (pattern, dirName);
if (dirName [strlen (dirName) - 1] != '/')
vStringPut (pattern, '/');
}
vStringCatS (pattern, "#?");
resize = createTagsForAmigaWildcard (vStringValue (pattern));
vStringDelete (pattern);
}
#endif
}
return resize;
}
static void addToScope (tokenInfo *const token, const vString *const extra,
int kindOfUpperScope)
{
if (vStringLength (token->scope) > 0)
{
const char* sep;
sep = phpScopeSeparatorFor(token->parentKind,
kindOfUpperScope);
vStringCatS (token->scope, sep);
}
vStringCat (token->scope, extra);
vStringTerminate(token->scope);
}
static void findAbcTags (void)
{
vString *name = vStringNew();
const unsigned char *line;
while ((line = readLineFromInputFile()) != NULL)
{
/*int name_len = vStringLength(name);*/
/* underlines must be the same length or more */
/*if (name_len > 0 && (line[0] == '=' || line[0] == '-') && issame((const char*) line))
{
makeAbcTag(name, TRUE);
}*/
/* if (line[1] == '%') {
vStringClear(name);
vStringCatS(name, (const char *) line);
vStringTerminate(name);
makeAbcTag(name, FALSE);
}*/
if (line[0] == 'T') {
/*vStringClear(name);*/
vStringCatS(name, " / ");
vStringCatS(name, (const char *) line);
vStringTerminate(name);
makeAbcTag(name, FALSE);
}
else {
vStringClear (name);
if (! isspace(*line))
vStringCatS(name, (const char*) line);
vStringTerminate(name);
}
}
vStringDelete (name);
}
static boolean createTagsForWildcardEntry (
const char *const pattern, const size_t dirLength,
const char *const entryName)
{
boolean resize = FALSE;
/* we must not recurse into the directories "." or ".." */
if (strcmp (entryName, ".") != 0 && strcmp (entryName, "..") != 0)
{
vString *const filePath = vStringNew ();
vStringNCopyS (filePath, pattern, dirLength);
vStringCatS (filePath, entryName);
resize = createTagsForEntry (vStringValue (filePath));
vStringDelete (filePath);
}
return resize;
}
static void initPhpEntry (tagEntryInfo *const e, const tokenInfo *const token,
const phpKind kind, const accessType access)
{
int parentKind = -1;
vStringClear (FullScope);
if (vStringLength (CurrentNamesapce) > 0)
{
parentKind = K_NAMESPACE;
vStringCat (FullScope, CurrentNamesapce);
}
initTagEntry (e, vStringValue (token->string), kind);
e->lineNumber = token->lineNumber;
e->filePosition = token->filePosition;
if (access != ACCESS_UNDEFINED)
e->extensionFields.access = accessToString (access);
if (vStringLength (token->scope) > 0)
{
parentKind = token->parentKind;
if (vStringLength (FullScope) > 0)
{
const char* sep;
sep = phpScopeSeparatorFor (parentKind,
K_NAMESPACE);
vStringCatS (FullScope, sep);
}
vStringCat (FullScope, token->scope);
}
if (vStringLength (FullScope) > 0)
{
Assert (parentKind >= 0);
e->extensionFields.scopeKindIndex = parentKind;
e->extensionFields.scopeName = vStringValue (FullScope);
}
if (token->anonymous)
markTagExtraBit (e, XTAG_ANONYMOUS);
}
/*
* Returns a string describing the scope in 'nls'.
* We record the current scope as a list of entered scopes.
* Scopes corresponding to 'if' statements and the like are
* represented by empty strings. Scopes corresponding to
* modules and classes are represented by the name of the
* module or class.
*/
static vString* nestingLevelsToScope (const NestingLevels* nls)
{
int i;
unsigned int chunks_output = 0;
vString* result = vStringNew ();
for (i = 0; i < nls->n; ++i)
{
NestingLevel *nl = nestingLevelsGetNth (nls, i);
tagEntryInfo *e = getEntryOfNestingLevel (nl);
if (e && strlen (e->name) > 0 && (!e->placeholder))
{
if (chunks_output++ > 0)
vStringPut (result, SCOPE_SEPARATOR);
vStringCatS (result, e->name);
}
}
return result;
}
/* TODO: parse overlining & underlining as distinct sections. */
static void findRstTags (void)
{
vString *name = vStringNew ();
fpos_t filepos;
const unsigned char *line;
memset(&filepos, 0, sizeof(fpos_t));
memset(kindchars, 0, sizeof kindchars);
nestingLevels = nestingLevelsNew();
while ((line = readLineFromInputFile ()) != NULL)
{
int line_len = strlen((const char*) line);
int name_len_bytes = vStringLength(name);
int name_len = utf8_strlen(vStringValue(name), name_len_bytes);
/* if the name doesn't look like UTF-8, assume one-byte charset */
if (name_len < 0)
name_len = name_len_bytes;
/* underlines must be the same length or more */
if (line_len >= name_len && name_len > 0 &&
ispunct(line[0]) && issame((const char*) line))
{
char c = line[0];
int kind = get_kind(c);
if (kind >= 0)
{
makeRstTag(name, kind, filepos);
continue;
}
}
vStringClear (name);
if (!isspace(*line))
{
vStringCatS(name, (const char*)line);
filepos = getInputFilePosition();
}
vStringTerminate(name);
}
vStringDelete (name);
nestingLevelsFree(nestingLevels);
}
static void createContext (tokenInfo *const scope)
{
if (scope)
{
vString *contextName = vStringNew ();
verbose ("Creating new context %s\n", vStringValue (scope->name));
/* Determine full context name */
if (currentContext->kind != K_UNDEFINED)
{
vStringCopy (contextName, currentContext->name);
vStringCatS (contextName, ".");
}
vStringCat (contextName, scope->name);
/* Create context */
currentContext = pushToken (currentContext, scope);
vStringCopy (currentContext->name, contextName);
vStringDelete (contextName);
}
}
static boolean createTagsForWildcardArg (const char *const arg)
{
boolean resize = FALSE;
vString *const pattern = vStringNewInit (arg);
char *patternS = vStringValue (pattern);
#if defined (HAVE_FINDFIRST) || defined (HAVE__FINDFIRST)
/* We must transform the "." and ".." forms into something that can
* be expanded by the findfirst/_findfirst functions.
*/
if (Option.recurse &&
(strcmp (patternS, ".") == 0 || strcmp (patternS, "..") == 0))
{
vStringPut (pattern, OUTPUT_PATH_SEPARATOR);
vStringCatS (pattern, "*.*");
}
resize |= createTagsForWildcardUsingFindfirst (patternS);
#endif
vStringDelete (pattern);
return resize;
}
void cxxTokenChainJoinInString(
CXXTokenChain * tc,
vString * s,
const char * szSeparator,
unsigned int uFlags
)
{
if(!tc)
return;
if(tc->iCount == 0)
return;
CXXToken * t = tc->pHead;
cxxTokenAppendToString(s,t);
if(
(!(uFlags & CXXTokenChainJoinNoTrailingSpaces)) &&
t->bFollowedBySpace
)
vStringPut (s, ' ');
t = t->pNext;
while(t)
{
if(szSeparator)
vStringCatS(s,szSeparator);
cxxTokenAppendToString(s,t);
if(
(!(uFlags & CXXTokenChainJoinNoTrailingSpaces)) &&
t->bFollowedBySpace
)
vStringPut (s, ' ');
t = t->pNext;
}
}
static void parseMethodsImplemName (vString * const ident, objcToken what)
{
switch (what)
{
case Tok_PARL:
toDoNext = &tillToken;
comeAfter = &parseMethodsImplemName;
waitedToken = Tok_PARR;
break;
case Tok_dpoint:
vStringCat (fullMethodName, prevIdent);
vStringCatS (fullMethodName, ":");
vStringClear (prevIdent);
break;
case ObjcIDENTIFIER:
vStringCopy (prevIdent, ident);
break;
case Tok_CurlL:
case Tok_semi:
/* method name is not simple */
if (vStringLength (fullMethodName) != '\0')
{
addTag (fullMethodName, methodKind);
vStringClear (fullMethodName);
}
else
addTag (prevIdent, methodKind);
toDoNext = &parseImplemMethods;
parseImplemMethods (ident, what);
vStringClear (prevIdent);
break;
default:
break;
}
}
static void makeTag (tokenInfo *const token, const goKind kind,
tokenInfo *const parent_token, const goKind parent_kind,
const char *argList, const char *varType)
{
const char *const name = vStringValue (token->string);
tagEntryInfo e;
initTagEntry (&e, name);
if (!GoKinds [kind].enabled)
return;
e.lineNumber = token->lineNumber;
e.filePosition = token->filePosition;
e.kindName = GoKinds [kind].name;
e.kind = GoKinds [kind].letter;
if (argList)
e.extensionFields.arglist = argList;
if (varType)
e.extensionFields.varType = varType;
if (parent_kind != GOTAG_UNDEFINED && parent_token != NULL)
{
e.extensionFields.scope[0] = GoKinds[parent_kind].name;
e.extensionFields.scope[1] = vStringValue (parent_token->string);
}
makeTagEntry (&e);
if (scope && Option.include.qualifiedTags)
{
vString *qualifiedName = vStringNew ();
vStringCopy (qualifiedName, scope);
vStringCatS (qualifiedName, ".");
vStringCat (qualifiedName, token->string);
e.name = vStringValue (qualifiedName);
makeTagEntry (&e);
vStringDelete (qualifiedName);
}
}
static void dropContext (tokenInfo *const token)
{
verbose ("current context %s; context kind %0d; nest level %0d\n", vStringValue (currentContext->name), currentContext->kind, currentContext->nestLevel);
vString *endTokenName = vStringNewInit("end");
if ((currentContext->kind == K_COVERGROUP && strcmp (vStringValue (token->name), "endgroup") == 0) ||
(currentContext->kind == K_BLOCK && currentContext->nestLevel == 0 && strcmp (vStringValue (token->name), vStringValue (endTokenName)) == 0)
)
{
verbose ("Dropping context %s\n", vStringValue (currentContext->name));
currentContext = popToken (currentContext);
}
else
{
vStringCatS (endTokenName, kindFromKind (currentContext->kind)->name);
if (strcmp (vStringValue (token->name), vStringValue (endTokenName)) == 0)
{
verbose ("Dropping context %s\n", vStringValue (currentContext->name));
currentContext = popToken (currentContext);
}
}
vStringDelete(endTokenName);
}
static void parseFunction (const unsigned char *line)
{
vString *name = vStringNew ();
/* boolean inFunction = FALSE; */
int scope;
const unsigned char *cp = line;
int index = CORK_NIL;
if (*cp == '!')
++cp;
if (isspace ((int) *cp))
{
while (*cp && isspace ((int) *cp))
++cp;
if (*cp)
{
cp = skipPrefix (cp, &scope);
if (isupper ((int) *cp) ||
scope == 's' || /* script scope */
scope == '<' || /* script scope */
scope == 'd' || /* dictionary */
scope == 'a') /* autoload */
{
char prefix[3] = { [0] = (char)scope, [1] = ':', [2] = '\0' };
if (scope == 's')
vStringCatS (name, prefix);
do
{
vStringPut (name, (int) *cp);
++cp;
} while (isalnum ((int) *cp) || *cp == '_' || *cp == '.' || *cp == '#');
vStringTerminate (name);
index = makeSimpleTag (name, VimKinds, K_FUNCTION);
vStringClear (name);
}
}
}
/* parses namespace declarations
* namespace Foo {}
* namespace Foo\Bar {}
* namespace Foo;
* namespace Foo\Bar;
* namespace;
* napespace {} */
static boolean parseNamespace (tokenInfo *const token)
{
tokenInfo *nsToken = newToken ();
vStringClear (CurrentNamesapce);
copyToken (nsToken, token, FALSE);
do
{
readToken (token);
if (token->type == TOKEN_IDENTIFIER)
{
if (vStringLength (CurrentNamesapce) > 0)
{
const char *sep;
sep = phpScopeSeparatorFor(K_NAMESPACE,
K_NAMESPACE);
vStringCatS (CurrentNamesapce, sep);
}
vStringCat (CurrentNamesapce, token->string);
}
}
while (token->type != TOKEN_EOF &&
token->type != TOKEN_SEMICOLON &&
token->type != TOKEN_OPEN_CURLY);
vStringTerminate (CurrentNamesapce);
if (vStringLength (CurrentNamesapce) > 0)
makeNamespacePhpTag (nsToken, CurrentNamesapce);
if (token->type == TOKEN_OPEN_CURLY)
enterScope (token, NULL, -1);
deleteToken (nsToken);
return TRUE;
}
static void makeTag (tokenInfo *const token, const goKind kind,
tokenInfo *const parent_token, const goKind parent_kind,
const char *argList)
{
const char *const name = vStringValue (token->string);
tagEntryInfo e;
initTagEntry (&e, name, &(GoKinds [kind]));
if (!GoKinds [kind].enabled)
return;
e.lineNumber = token->lineNumber;
e.filePosition = token->filePosition;
if (argList)
e.extensionFields.signature = argList;
if (parent_kind != GOTAG_UNDEFINED && parent_token != NULL)
{
e.extensionFields.scopeKind = &(GoKinds[parent_kind]);
e.extensionFields.scopeName = vStringValue (parent_token->string);
}
makeTagEntry (&e);
if (scope && isXtagEnabled(XTAG_QUALIFIED_TAGS))
{
vString *qualifiedName = vStringNew ();
vStringCopy (qualifiedName, scope);
vStringCatS (qualifiedName, ".");
vStringCat (qualifiedName, token->string);
e.name = vStringValue (qualifiedName);
markTagExtraBit (&e, XTAG_QUALIFIED_TAGS);
makeTagEntry (&e);
vStringDelete (qualifiedName);
}
}
void cxxTokenChainJoinRangeInString(
CXXToken * from,
CXXToken * to,
vString * s,
const char * szSeparator,
unsigned int uFlags
)
{
if(!from)
return;
CXXToken * t = from;
cxxTokenAppendToString(s,t);
if((!(uFlags & CXXTokenChainJoinNoTrailingSpaces)) && t->bFollowedBySpace)
vStringPut (s, ' ');
while(t && (t != to))
{
t = t->pNext;
if(t)
return;
if(szSeparator)
vStringCatS(s,szSeparator);
cxxTokenAppendToString(s,t);
if(
(!(uFlags & CXXTokenChainJoinNoTrailingSpaces)) &&
t->bFollowedBySpace
)
vStringPut (s, ' ');
}
}
// In case of a lambda without parentheses this is the capture list token.
boolean cxxParserHandleLambda(CXXToken * pParenthesis)
{
CXX_DEBUG_ENTER();
CXXToken * pIdentifier = cxxTokenCreateAnonymousIdentifier(CXXTagKindFUNCTION);
CXXTokenChain * pSave = g_cxx.pTokenChain;
CXXTokenChain * pNew = cxxTokenChainCreate();
g_cxx.pTokenChain = pNew;
tagEntryInfo * tag = cxxTagBegin(CXXTagKindFUNCTION,pIdentifier);
CXXToken * pAfterParenthesis = pParenthesis ? pParenthesis->pNext : NULL;
CXXToken * pCaptureList = NULL;
if(pParenthesis)
{
if(cxxTokenTypeIs(pParenthesis,CXXTokenTypeSquareParenthesisChain))
{
// form (4) of lambda (see cxxParserOpeningBracketIsLambda()).
pCaptureList = pParenthesis;
} else if(
pParenthesis->pPrev &&
cxxTokenTypeIs(pParenthesis->pPrev,CXXTokenTypeSquareParenthesisChain)
)
{
// other forms of lambda (see cxxParserOpeningBracketIsLambda()).
pCaptureList = pParenthesis->pPrev;
}
}
if(
pAfterParenthesis &&
cxxTokenTypeIs(pAfterParenthesis,CXXTokenTypeKeyword) &&
(pAfterParenthesis->eKeyword == CXXKeywordCONST)
)
pAfterParenthesis = pAfterParenthesis->pNext;
CXXToken * pTypeStart = NULL;
CXXToken * pTypeEnd;
if(
pAfterParenthesis &&
cxxTokenTypeIs(pAfterParenthesis,CXXTokenTypePointerOperator) &&
pAfterParenthesis->pNext &&
!cxxTokenTypeIs(pAfterParenthesis->pNext,CXXTokenTypeOpeningBracket)
)
{
pTypeStart = pAfterParenthesis->pNext;
pTypeEnd = pTypeStart;
while(
pTypeEnd->pNext &&
(!cxxTokenTypeIs(pTypeEnd->pNext,CXXTokenTypeOpeningBracket))
)
pTypeEnd = pTypeEnd->pNext;
while(
(pTypeStart != pTypeEnd) &&
cxxTokenTypeIs(pTypeStart,CXXTokenTypeKeyword) &&
cxxKeywordExcludeFromTypeNames(pTypeStart->eKeyword)
)
pTypeStart = pTypeStart->pNext;
}
int iCorkQueueIndex = CORK_NIL;
if(tag)
{
tag->isFileScope = TRUE;
CXXToken * pTypeName;
if(pTypeStart)
pTypeName = cxxTagSetTypeField(pTypeStart,pTypeEnd);
else
pTypeName = NULL;
if(pCaptureList && cxxTagCPPFieldEnabled(CXXTagCPPFieldLambdaCaptureList))
{
CXX_DEBUG_ASSERT(pCaptureList->pChain,"The capture list must be a chain");
cxxTokenChainCondense(pCaptureList->pChain,0);
CXX_DEBUG_ASSERT(
cxxTokenChainFirst(pCaptureList->pChain),
"Condensation should have created a single token in the chain"
);
cxxTagSetCPPField(
CXXTagCPPFieldLambdaCaptureList,
vStringValue(cxxTokenChainFirst(pCaptureList->pChain)->pszWord)
);
}
// FIXME: Properties?
vString * pszSignature = NULL;
if(cxxTokenTypeIs(pParenthesis,CXXTokenTypeParenthesisChain))
pszSignature = cxxTokenChainJoin(pParenthesis->pChain,NULL,0);
if(pszSignature)
tag->extensionFields.signature = vStringValue(pszSignature);
iCorkQueueIndex = cxxTagCommit();
if(pTypeName)
cxxTokenDestroy(pTypeName);
if(pszSignature)
vStringDelete(pszSignature);
}
cxxScopePush(
pIdentifier,
CXXTagKindFUNCTION,
CXXScopeAccessUnknown
);
if(
pParenthesis &&
cxxTokenTypeIs(pParenthesis,CXXTokenTypeParenthesisChain) &&
cxxTagKindEnabled(CXXTagKindPARAMETER)
)
{
CXXFunctionParameterInfo oParamInfo;
if(cxxParserTokenChainLooksLikeFunctionParameterList(
pParenthesis->pChain,&oParamInfo
))
cxxParserEmitFunctionParameterTags(&oParamInfo);
}
boolean bRet = cxxParserParseBlock(TRUE);
if(iCorkQueueIndex > CORK_NIL)
cxxParserMarkEndLineForTagInCorkQueue(iCorkQueueIndex);
cxxScopePop();
pNew = g_cxx.pTokenChain; // May have been destroyed and re-created
g_cxx.pTokenChain = pSave;
g_cxx.pToken = pSave->pTail;
// change the type of token so following parsing code is not confused too much
g_cxx.pToken->eType = CXXTokenTypeAngleBracketChain;
g_cxx.pToken->pChain = pNew;
cxxTokenChainClear(pNew);
CXXToken * t = cxxTokenCreate();
t->eType = CXXTokenTypeOpeningBracket;
vStringCatS(t->pszWord,"{");
cxxTokenChainAppend(pNew,t);
t = cxxTokenCreate();
t->eType = CXXTokenTypeClosingBracket;
vStringCatS(t->pszWord,"}");
cxxTokenChainAppend(pNew,t);
CXX_DEBUG_LEAVE();
return bRet;
}
static void readToken (tokenInfo *const token)
{
int c;
static tokenType lastTokenType = TOKEN_NONE;
boolean firstWhitespace = TRUE;
boolean whitespace;
token->type = TOKEN_NONE;
token->keyword = KEYWORD_NONE;
vStringClear (token->string);
getNextChar:
do
{
c = getcFromInputFile ();
token->lineNumber = getInputLineNumber ();
token->filePosition = getInputFilePosition ();
if (c == '\n' && (lastTokenType == TOKEN_IDENTIFIER ||
lastTokenType == TOKEN_STRING ||
lastTokenType == TOKEN_OTHER ||
lastTokenType == TOKEN_CLOSE_PAREN ||
lastTokenType == TOKEN_CLOSE_CURLY ||
lastTokenType == TOKEN_CLOSE_SQUARE))
{
c = ';'; // semicolon injection
}
whitespace = c == '\t' || c == ' ' || c == '\r' || c == '\n';
if (signature && whitespace && firstWhitespace && vStringLength (signature) < MAX_SIGNATURE_LENGTH)
{
firstWhitespace = FALSE;
vStringPut(signature, ' ');
}
}
while (whitespace);
switch (c)
{
case EOF:
token->type = TOKEN_EOF;
break;
case ';':
token->type = TOKEN_SEMICOLON;
break;
case '/':
{
boolean hasNewline = FALSE;
int d = getcFromInputFile ();
switch (d)
{
case '/':
skipToCharacterInInputFile ('\n');
/* Line comments start with the
* character sequence // and
* continue through the next
* newline. A line comment acts
* like a newline. */
ungetcToInputFile ('\n');
goto getNextChar;
case '*':
do
{
do
{
d = getcFromInputFile ();
if (d == '\n')
{
hasNewline = TRUE;
}
} while (d != EOF && d != '*');
c = getcFromInputFile ();
if (c == '/')
break;
else
ungetcToInputFile (c);
} while (c != EOF && c != '\0');
ungetcToInputFile (hasNewline ? '\n' : ' ');
goto getNextChar;
default:
token->type = TOKEN_OTHER;
ungetcToInputFile (d);
break;
}
}
break;
case '"':
case '\'':
case '`':
token->type = TOKEN_STRING;
parseString (token->string, c);
token->lineNumber = getInputLineNumber ();
token->filePosition = getInputFilePosition ();
break;
case '<':
{
int d = getcFromInputFile ();
if (d == '-')
token->type = TOKEN_LEFT_ARROW;
else
{
ungetcToInputFile (d);
token->type = TOKEN_OTHER;
}
}
break;
case '(':
token->type = TOKEN_OPEN_PAREN;
break;
case ')':
token->type = TOKEN_CLOSE_PAREN;
break;
case '{':
token->type = TOKEN_OPEN_CURLY;
break;
case '}':
token->type = TOKEN_CLOSE_CURLY;
break;
case '[':
token->type = TOKEN_OPEN_SQUARE;
break;
case ']':
token->type = TOKEN_CLOSE_SQUARE;
break;
case '*':
token->type = TOKEN_STAR;
break;
case '.':
token->type = TOKEN_DOT;
break;
case ',':
token->type = TOKEN_COMMA;
break;
default:
if (isStartIdentChar (c))
{
parseIdentifier (token->string, c);
token->lineNumber = getInputLineNumber ();
token->filePosition = getInputFilePosition ();
token->keyword = lookupKeyword (vStringValue (token->string), Lang_go);
if (isKeyword (token, KEYWORD_NONE))
token->type = TOKEN_IDENTIFIER;
else
token->type = TOKEN_KEYWORD;
}
else
token->type = TOKEN_OTHER;
break;
}
if (signature && vStringLength (signature) < MAX_SIGNATURE_LENGTH)
{
if (token->type == TOKEN_LEFT_ARROW)
vStringCatS(signature, "<-");
else if (token->type == TOKEN_STRING)
{
// only struct member annotations can appear in function prototypes
// so only `` type strings are possible
vStringPut(signature, '`');
vStringCat(signature, token->string);
vStringPut(signature, '`');
}
else if (token->type == TOKEN_IDENTIFIER || token->type == TOKEN_KEYWORD)
vStringCat(signature, token->string);
else if (c != EOF)
vStringPut(signature, c);
}
lastTokenType = token->type;
}
/* parse a function
*
* if @name is NULL, parses a normal function
* function myfunc($foo, $bar) {}
* function &myfunc($foo, $bar) {}
* function myfunc($foo, $bar) : type {}
*
* if @name is not NULL, parses an anonymous function with name @name
* $foo = function($foo, $bar) {}
* $foo = function&($foo, $bar) {}
* $foo = function($foo, $bar) use ($x, &$y) {}
* $foo = function($foo, $bar) use ($x, &$y) : type {} */
static boolean parseFunction (tokenInfo *const token, const tokenInfo *name)
{
boolean readNext = TRUE;
accessType access = CurrentStatement.access;
implType impl = CurrentStatement.impl;
tokenInfo *nameFree = NULL;
readToken (token);
/* skip a possible leading ampersand (return by reference) */
if (token->type == TOKEN_AMPERSAND)
readToken (token);
if (! name)
{
if (token->type != TOKEN_IDENTIFIER && token->type != TOKEN_KEYWORD)
return FALSE;
name = nameFree = newToken ();
copyToken (nameFree, token, TRUE);
readToken (token);
}
if (token->type == TOKEN_OPEN_PAREN)
{
vString *arglist = vStringNew ();
int depth = 1;
vStringPut (arglist, '(');
do
{
readToken (token);
switch (token->type)
{
case TOKEN_OPEN_PAREN: depth++; break;
case TOKEN_CLOSE_PAREN: depth--; break;
default: break;
}
/* display part */
switch (token->type)
{
case TOKEN_AMPERSAND: vStringPut (arglist, '&'); break;
case TOKEN_CLOSE_CURLY: vStringPut (arglist, '}'); break;
case TOKEN_CLOSE_PAREN: vStringPut (arglist, ')'); break;
case TOKEN_CLOSE_SQUARE: vStringPut (arglist, ']'); break;
case TOKEN_COLON: vStringPut (arglist, ':'); break;
case TOKEN_COMMA: vStringCatS (arglist, ", "); break;
case TOKEN_EQUAL_SIGN: vStringCatS (arglist, " = "); break;
case TOKEN_OPEN_CURLY: vStringPut (arglist, '{'); break;
case TOKEN_OPEN_PAREN: vStringPut (arglist, '('); break;
case TOKEN_OPEN_SQUARE: vStringPut (arglist, '['); break;
case TOKEN_PERIOD: vStringPut (arglist, '.'); break;
case TOKEN_SEMICOLON: vStringPut (arglist, ';'); break;
case TOKEN_BACKSLASH: vStringPut (arglist, '\\'); break;
case TOKEN_STRING:
{
vStringCatS (arglist, "'");
vStringCat (arglist, token->string);
vStringCatS (arglist, "'");
break;
}
case TOKEN_IDENTIFIER:
case TOKEN_KEYWORD:
case TOKEN_VARIABLE:
{
switch (vStringLast (arglist))
{
case 0:
case ' ':
case '{':
case '(':
case '[':
case '.':
case '\\':
/* no need for a space between those and the identifier */
break;
default:
vStringPut (arglist, ' ');
break;
}
if (token->type == TOKEN_VARIABLE)
vStringPut (arglist, '$');
vStringCat (arglist, token->string);
break;
}
default: break;
}
}
while (token->type != TOKEN_EOF && depth > 0);
vStringTerminate (arglist);
makeFunctionTag (name, arglist, access, impl);
vStringDelete (arglist);
readToken (token); /* normally it's an open brace or "use" keyword */
}
/* skip use(...) */
if (token->type == TOKEN_KEYWORD && token->keyword == KEYWORD_use)
{
readToken (token);
skipOverParens (token);
}
/* PHP7 return type declaration or if parsing Zephir, skip function return
* type hint */
if ((getInputLanguage () == Lang_php && token->type == TOKEN_COLON) ||
(getInputLanguage () == Lang_zephir && token->type == TOKEN_OPERATOR))
{
do
readToken (token);
while (token->type == TOKEN_IDENTIFIER ||
token->type == TOKEN_BACKSLASH);
}
if (token->type == TOKEN_OPEN_CURLY)
enterScope (token, name->string, K_FUNCTION);
else
readNext = FALSE;
if (nameFree)
deleteToken (nameFree);
return readNext;
}
/* Algorithm adapted from from GNU etags.
* Perl support by Bart Robinson <*****@*****.**>
* Perl sub names: look for /^ [ \t\n]sub [ \t\n]+ [^ \t\n{ (]+/
*/
static void findPerlTags (void)
{
vString *name = vStringNew ();
vString *package = NULL;
boolean skipPodDoc = FALSE;
const unsigned char *line;
/* Core modules AutoLoader and SelfLoader support delayed compilation
* by allowing Perl code that follows __END__ and __DATA__ tokens,
* respectively. When we detect that one of these modules is used
* in the file, we continue processing even after we see the
* corresponding token that would usually terminate parsing of the
* file.
*/
enum {
RESPECT_END = (1 << 0),
RESPECT_DATA = (1 << 1),
} respect_token = RESPECT_END | RESPECT_DATA;
while ((line = fileReadLine ()) != NULL)
{
boolean spaceRequired = FALSE;
boolean qualified = FALSE;
const unsigned char *cp = line;
perlKind kind = K_NONE;
tagEntryInfo e;
if (skipPodDoc)
{
if (strncmp ((const char*) line, "=cut", (size_t) 4) == 0)
skipPodDoc = FALSE;
continue;
}
else if (line [0] == '=')
{
skipPodDoc = isPodWord ((const char*)line + 1);
continue;
}
else if (strcmp ((const char*) line, "__DATA__") == 0)
{
if (respect_token & RESPECT_DATA)
break;
else
continue;
}
else if (strcmp ((const char*) line, "__END__") == 0)
{
if (respect_token & RESPECT_END)
break;
else
continue;
}
else if (line [0] == '#')
continue;
while (isspace (*cp))
cp++;
if (strncmp((const char*) cp, "sub", (size_t) 3) == 0)
{
TRACE("this looks like a sub\n");
cp += 3;
kind = K_SUBROUTINE;
spaceRequired = TRUE;
qualified = TRUE;
}
else if (strncmp((const char*) cp, "use", (size_t) 3) == 0)
{
cp += 3;
if (!isspace(*cp))
continue;
while (*cp && isspace (*cp))
++cp;
if (strncmp((const char*) cp, "AutoLoader", (size_t) 10) == 0) {
respect_token &= ~RESPECT_END;
continue;
}
if (strncmp((const char*) cp, "SelfLoader", (size_t) 10) == 0) {
respect_token &= ~RESPECT_DATA;
continue;
}
if (strncmp((const char*) cp, "constant", (size_t) 8) != 0)
continue;
cp += 8;
/* Skip up to the first non-space character, skipping empty
* and comment lines.
*/
while (isspace(*cp))
cp++;
while (!*cp || '#' == *cp) {
cp = fileReadLine ();
if (!cp)
goto END_MAIN_WHILE;
while (isspace (*cp))
cp++;
}
if ('{' == *cp) {
++cp;
if (0 == parseConstantsFromHashRef(cp, name, package)) {
vStringClear(name);
continue;
} else
goto END_MAIN_WHILE;
}
kind = K_CONSTANT;
spaceRequired = FALSE;
qualified = TRUE;
}
else if (strncmp((const char*) cp, "package", (size_t) 7) == 0 &&
('\0' == cp[7] || isspace(cp[7])))
{
cp += 7;
while (isspace (*cp))
cp++;
while (!*cp || '#' == *cp) {
cp = fileReadLine ();
if (!cp)
goto END_MAIN_WHILE;
while (isspace (*cp))
cp++;
}
if (package == NULL)
package = vStringNew ();
else
vStringClear (package);
const unsigned char *const first = cp;
while (*cp && (int) *cp != ';' && !isspace ((int) *cp))
{
vStringPut (package, (int) *cp);
cp++;
}
vStringCatS (package, "::");
cp = first; /* Rewind */
kind = K_PACKAGE;
spaceRequired = FALSE;
qualified = TRUE;
}
else if (strncmp((const char*) cp, "format", (size_t) 6) == 0)
{
cp += 6;
kind = K_FORMAT;
spaceRequired = TRUE;
qualified = TRUE;
}
else
{
if (isIdentifier1 (*cp))
{
const unsigned char *p = cp;
while (isIdentifier (*p))
++p;
while (isspace (*p))
++p;
if ((int) *p == ':' && (int) *(p + 1) != ':')
kind = K_LABEL;
}
}
if (kind != K_NONE)
{
TRACE("cp0: %s\n", (const char *) cp);
if (spaceRequired && *cp && !isspace (*cp))
continue;
TRACE("cp1: %s\n", (const char *) cp);
while (isspace (*cp))
cp++;
while (!*cp || '#' == *cp) { /* Gobble up empty lines
and comments */
cp = fileReadLine ();
if (!cp)
goto END_MAIN_WHILE;
while (isspace (*cp))
cp++;
}
while (isIdentifier (*cp) || (K_PACKAGE == kind && ':' == *cp))
{
vStringPut (name, (int) *cp);
cp++;
}
if (K_FORMAT == kind &&
vStringLength (name) == 0 && /* cp did not advance */
'=' == *cp)
{
/* format's name is optional. If it's omitted, 'STDOUT'
is assumed. */
vStringCatS (name, "STDOUT");
}
vStringTerminate (name);
TRACE("name: %s\n", name->buffer);
if (0 == vStringLength(name)) {
vStringClear(name);
continue;
}
if (K_SUBROUTINE == kind)
{
/*
* isSubroutineDeclaration() may consume several lines. So
* we record line positions.
*/
initTagEntry(&e, vStringValue(name), NULL);
if (TRUE == isSubroutineDeclaration(cp)) {
if (TRUE == PerlKinds[K_SUBROUTINE_DECLARATION].enabled) {
kind = K_SUBROUTINE_DECLARATION;
} else {
vStringClear (name);
continue;
}
} else if (! PerlKinds[kind].enabled) {
continue;
}
e.kind = &(PerlKinds[kind]);
makeTagEntry(&e);
if (Option.include.qualifiedTags && qualified &&
package != NULL && vStringLength (package) > 0)
{
vString *const qualifiedName = vStringNew ();
vStringCopy (qualifiedName, package);
vStringCat (qualifiedName, name);
e.name = vStringValue(qualifiedName);
makeTagEntry(&e);
vStringDelete (qualifiedName);
}
} else if (vStringLength (name) > 0)
{
makeSimpleTag (name, PerlKinds, kind);
if (Option.include.qualifiedTags && qualified &&
K_PACKAGE != kind &&
package != NULL && vStringLength (package) > 0)
{
vString *const qualifiedName = vStringNew ();
vStringCopy (qualifiedName, package);
vStringCat (qualifiedName, name);
makeSimpleTag (qualifiedName, PerlKinds, kind);
vStringDelete (qualifiedName);
}
}
vStringClear (name);
}
}
END_MAIN_WHILE:
vStringDelete (name);
if (package != NULL)
vStringDelete (package);
}
static boolean loadPathKinds (xcmdPath *const path, const langType language)
{
enum pcoprocError r;
FILE* pp = NULL;
char * argv[3];
int status;
vString * opt;
char file_kind = getLanguageFileKind (language)->letter;
opt = vStringNewInit(XCMD_LIST_KIND_OPTION);
vStringCatS (opt, "=");
vStringCatS (opt, getLanguageName(language));
argv[2] = NULL;
argv[1] = vStringValue (opt);
argv[0] = vStringValue (path->path);
errno = 0;
if (getuid() == 0 || geteuid() == 0)
{
verbose ("all xcmd feature is disabled when running ctags in root privilege\n");
vStringDelete (opt);
return FALSE;
}
if (! isSafeExecutable (argv [0]))
{
vStringDelete (opt);
return FALSE;
}
verbose ("loading path kinds of %s from [%s %s]\n", getLanguageName(language), argv[0], argv[1]);
r = pcoprocOpen (vStringValue (path->path), argv, &pp, NULL);
switch (r) {
case PCOPROC_ERROR_WPIPE:
error (WARNING | PERROR, "failed to make pipe to write to command: [%s %s]",
argv[0], argv[1]);
break;
case PCOPROC_ERROR_RPIPE:
error (WARNING | PERROR, "failed to make pipe to read from command: [%s %s]",
argv[0], argv[1]);
break;
case PCOPROC_ERROR_FORK:
error (WARNING | PERROR, "failed to do fork: [%s %s]",
argv[0], argv[1]);
break;
case PCOPROC_SUCCESSFUL:
break;
}
if (pp)
{
vString* vline = vStringNew();
while (readLineRawWithNoSeek (vline, pp))
{
char* line;
char kind_letter;
vStringStripNewline (vline);
line = vStringValue (vline);
if (!loadPathKind (path, line, argv))
break;
kind_letter = path->kinds [path->n_kinds - 1].letter;
if (kind_letter == file_kind)
error (FATAL,
"Kind letter \'%c\' returned from xcmd %s of %s language is reserved in ctags main",
kind_letter,
vStringValue (path->path),
getLanguageName (language));
}
vStringDelete (vline);
status = pcoprocClose (pp);
/* TODO: Decode status */
verbose(" status: %d\n", status);
if (status != 0)
{
if (status > 0
&& WIFEXITED (status)
&& (WEXITSTATUS (status) == path->not_available_status))
verbose ("xcmd: the %s backend is not available\n", argv[0]);
else
error (WARNING, "xcmd exits abnormally status(%d): [%s %s]",
status, argv[0], argv[1]);
vStringDelete (opt);
return FALSE;
}
}
else
{
error (WARNING | PERROR, "cannot make pipe to xcmd: [%s %s]",
argv[0], argv[1]);
}
vStringDelete (opt);
return path->kinds == NULL? FALSE: TRUE;
}
/* Adds a name to the end of the scope string */
static void addToScope (vString *scope, vString *name)
{
if (vStringLength(scope) > 0)
vStringCatS(scope, "::");
vStringCat(scope, name);
}
static vString* ext2ptrnNew (const char *const ext)
{
vString * ptrn = vStringNewInit ("*.");
vStringCatS (ptrn, ext);
return ptrn;
}
static boolean parseTag (tokenInfo *const token, texKind kind)
{
tokenInfo *const name = newToken ();
vString * fullname;
boolean useLongName = TRUE;
fullname = vStringNew ();
vStringClear (fullname);
/*
* Tex tags are of these formats:
* \keyword{any number of words}
* \keyword[short desc]{any number of words}
* \keyword*[short desc]{any number of words}
*
* When a keyword is found, loop through all words within
* the curly braces for the tag name.
*/
if (isType (token, TOKEN_KEYWORD))
{
copyToken (name, token);
readToken (token);
}
if (isType (token, TOKEN_OPEN_SQUARE))
{
useLongName = FALSE;
readToken (token);
while (! isType (token, TOKEN_CLOSE_SQUARE) )
{
if (isType (token, TOKEN_IDENTIFIER))
{
if (fullname->length > 0)
vStringCatS (fullname, " ");
vStringCatS (fullname, vStringValue (token->string));
}
readToken (token);
}
vStringTerminate (fullname);
vStringCopy (name->string, fullname);
makeTexTag (name, kind);
}
if (isType (token, TOKEN_STAR))
{
readToken (token);
}
if (isType (token, TOKEN_OPEN_CURLY))
{
readToken (token);
while (! isType (token, TOKEN_CLOSE_CURLY) )
{
if (isType (token, TOKEN_IDENTIFIER) && useLongName)
{
if (fullname->length > 0)
vStringCatS (fullname, " ");
vStringCatS (fullname, vStringValue (token->string));
}
readToken (token);
}
if (useLongName)
{
vStringTerminate (fullname);
vStringCopy (name->string, fullname);
makeTexTag (name, kind);
}
}
deleteToken (name);
vStringDelete (fullname);
return TRUE;
}
static void readTokenFull (tokenInfo *const token, boolean include_newlines, vString *const repr)
{
int c;
int i;
token->type = TOKEN_UNDEFINED;
token->keyword = KEYWORD_NONE;
vStringClear (token->string);
getNextChar:
i = 0;
do
{
c = fileGetc ();
i++;
}
while (c == '\t' || c == ' ' ||
((c == '\r' || c == '\n') && ! include_newlines));
token->lineNumber = getSourceLineNumber ();
token->filePosition = getInputFilePosition ();
if (repr)
{
if (i > 1)
vStringPut (repr, ' ');
vStringPut (repr, c);
}
switch (c)
{
case EOF: token->type = TOKEN_EOF; break;
case '(': token->type = TOKEN_OPEN_PAREN; break;
case ')': token->type = TOKEN_CLOSE_PAREN; break;
case ';': token->type = TOKEN_SEMICOLON; break;
case ',': token->type = TOKEN_COMMA; break;
case '.': token->type = TOKEN_PERIOD; break;
case ':': token->type = TOKEN_COLON; break;
case '{': token->type = TOKEN_OPEN_CURLY; break;
case '}': token->type = TOKEN_CLOSE_CURLY; break;
case '=': token->type = TOKEN_EQUAL_SIGN; break;
case '[': token->type = TOKEN_OPEN_SQUARE; break;
case ']': token->type = TOKEN_CLOSE_SQUARE; break;
case '+':
case '-':
{
int d = fileGetc ();
if (d == c) /* ++ or -- */
token->type = TOKEN_POSTFIX_OPERATOR;
else
{
fileUngetc (d);
token->type = TOKEN_BINARY_OPERATOR;
}
break;
}
case '*':
case '%':
case '?':
case '>':
case '<':
case '^':
case '|':
case '&':
token->type = TOKEN_BINARY_OPERATOR;
break;
case '\r':
case '\n':
/* This isn't strictly correct per the standard, but following the
* real rules means understanding all statements, and that's not
* what the parser currently does. What we do here is a guess, by
* avoiding inserting semicolons that would make the statement on
* the left invalid. Hopefully this should not have false negatives
* (e.g. should not miss insertion of a semicolon) but might have
* false positives (e.g. it will wrongfully emit a semicolon for the
* newline in "foo\n+bar").
* This should however be mostly harmless as we only deal with
* newlines in specific situations where we know a false positive
* wouldn't hurt too bad. */
switch (LastTokenType)
{
/* these cannot be the end of a statement, so hold the newline */
case TOKEN_EQUAL_SIGN:
case TOKEN_COLON:
case TOKEN_PERIOD:
case TOKEN_FORWARD_SLASH:
case TOKEN_BINARY_OPERATOR:
/* and these already end one, no need to duplicate it */
case TOKEN_SEMICOLON:
case TOKEN_COMMA:
case TOKEN_CLOSE_CURLY:
case TOKEN_OPEN_CURLY:
include_newlines = FALSE; /* no need to recheck */
goto getNextChar;
break;
default:
token->type = TOKEN_SEMICOLON;
}
break;
case '\'':
case '"':
token->type = TOKEN_STRING;
parseString (token->string, c);
token->lineNumber = getSourceLineNumber ();
token->filePosition = getInputFilePosition ();
if (repr)
{
vStringCat (repr, token->string);
vStringPut (repr, c);
}
break;
case '\\':
c = fileGetc ();
if (c != '\\' && c != '"' && !isspace (c))
fileUngetc (c);
token->type = TOKEN_CHARACTER;
token->lineNumber = getSourceLineNumber ();
token->filePosition = getInputFilePosition ();
break;
case '/':
{
int d = fileGetc ();
if ( (d != '*') && /* is this the start of a comment? */
(d != '/') ) /* is a one line comment? */
{
fileUngetc (d);
switch (LastTokenType)
{
case TOKEN_CHARACTER:
case TOKEN_IDENTIFIER:
case TOKEN_STRING:
case TOKEN_CLOSE_CURLY:
case TOKEN_CLOSE_PAREN:
case TOKEN_CLOSE_SQUARE:
token->type = TOKEN_FORWARD_SLASH;
break;
default:
token->type = TOKEN_REGEXP;
parseRegExp ();
token->lineNumber = getSourceLineNumber ();
token->filePosition = getInputFilePosition ();
break;
}
}
else
{
if (repr) /* remove the / we added */
repr->buffer[--repr->length] = 0;
if (d == '*')
{
do
{
skipToCharacter ('*');
c = fileGetc ();
if (c == '/')
break;
else
fileUngetc (c);
} while (c != EOF && c != '\0');
goto getNextChar;
}
else if (d == '/') /* is this the start of a comment? */
{
skipToCharacter ('\n');
/* if we care about newlines, put it back so it is seen */
if (include_newlines)
fileUngetc ('\n');
goto getNextChar;
}
}
break;
}
case '#':
/* skip shebang in case of e.g. Node.js scripts */
if (token->lineNumber > 1)
token->type = TOKEN_UNDEFINED;
else if ((c = fileGetc ()) != '!')
{
fileUngetc (c);
token->type = TOKEN_UNDEFINED;
}
else
{
skipToCharacter ('\n');
goto getNextChar;
}
break;
default:
if (! isIdentChar (c))
token->type = TOKEN_UNDEFINED;
else
{
parseIdentifier (token->string, c);
token->lineNumber = getSourceLineNumber ();
token->filePosition = getInputFilePosition ();
token->keyword = analyzeToken (token->string);
if (isKeyword (token, KEYWORD_NONE))
token->type = TOKEN_IDENTIFIER;
else
token->type = TOKEN_KEYWORD;
if (repr && vStringLength (token->string) > 1)
vStringCatS (repr, vStringValue (token->string) + 1);
}
break;
}
LastTokenType = token->type;
}
boolean cxxParserParseNextToken(void)
{
CXXToken * t = cxxTokenCreate();
// The token chain should not be allowed to grow arbitrairly large.
// The token structures are quite big and it's easy to grow up to
// 5-6GB or memory usage. However this limit should be large enough
// to accomodate all the reasonable statements that could have some
// information in them. This includes multiple function prototypes
// in a single statement (ImageMagick has some examples) but probably
// does NOT include large data tables.
if(g_cxx.pTokenChain->iCount > 16384)
cxxTokenChainDestroyLast(g_cxx.pTokenChain);
cxxTokenChainAppend(g_cxx.pTokenChain,t);
g_cxx.pToken = t;
cxxParserSkipToNonWhiteSpace();
// FIXME: this cpp handling is kind of broken:
// it works only because the moon is in the correct phase.
cppBeginStatement();
// This must be done after getting char from input
t->iLineNumber = getInputLineNumber();
t->oFilePosition = getInputFilePosition();
if(g_cxx.iChar == EOF)
{
t->eType = CXXTokenTypeEOF;
t->bFollowedBySpace = FALSE;
return FALSE;
}
unsigned int uInfo = (g_cxx.iChar < 0x80) ? g_aCharTable[g_cxx.iChar].uType : 0;
//printf("Char %c %02x info %u\n",g_cxx.iChar,g_cxx.iChar,uInfo);
if(uInfo & CXXCharTypeStartOfIdentifier)
{
// word
t->eType = CXXTokenTypeIdentifier;
t->bFollowedBySpace = FALSE;
vStringPut(t->pszWord,g_cxx.iChar);
// special case for tile, which may actually be an operator
if(g_cxx.iChar == '~')
{
// may be followed by space!
g_cxx.iChar = cppGetc();
if(isspace(g_cxx.iChar))
{
t->bFollowedBySpace = TRUE;
g_cxx.iChar = cppGetc();
while(isspace(g_cxx.iChar))
g_cxx.iChar = cppGetc();
}
// non space
uInfo = (g_cxx.iChar < 0x80) ? g_aCharTable[g_cxx.iChar].uType : 0;
if(!(uInfo & CXXCharTypeStartOfIdentifier))
{
// this is not an identifier after all
t->eType = CXXTokenTypeOperator;
if((!t->bFollowedBySpace) && g_cxx.iChar == '=')
{
// make ~= single token so it's not handled as
// a separate assignment
vStringPut(t->pszWord,g_cxx.iChar);
g_cxx.iChar = cppGetc();
t->bFollowedBySpace = isspace(g_cxx.iChar);
}
return TRUE;
}
} else {
g_cxx.iChar = cppGetc();
}
for(;;)
{
uInfo = (g_cxx.iChar < 0x80) ? g_aCharTable[g_cxx.iChar].uType : 0;
if(!(uInfo & CXXCharTypePartOfIdentifier))
break;
vStringPut(t->pszWord,g_cxx.iChar);
g_cxx.iChar = cppGetc();
}
int iCXXKeyword = lookupKeyword(t->pszWord->buffer,g_cxx.eLanguage);
if(iCXXKeyword >= 0)
{
if(
(
(iCXXKeyword == CXXKeywordFINAL) &&
(!g_cxx.bParsingClassStructOrUnionDeclaration)
) || (
(
(iCXXKeyword == CXXKeywordPUBLIC) ||
(iCXXKeyword == CXXKeywordPROTECTED) ||
(iCXXKeyword == CXXKeywordPRIVATE)
) &&
(!g_cxx.bEnablePublicProtectedPrivateKeywords)
)
)
{
t->eType = CXXTokenTypeIdentifier;
} else {
t->eType = CXXTokenTypeKeyword;
t->eKeyword = (enum CXXKeyword)iCXXKeyword;
if(iCXXKeyword == CXXKeyword__ATTRIBUTE__)
{
// special handling for __attribute__
return cxxParserParseNextTokenCondenseAttribute();
}
}
} else {
boolean bIgnoreParens = FALSE;
const char * szReplacement = NULL;
if(isIgnoreToken(
vStringValue(t->pszWord),
&bIgnoreParens,
&szReplacement
))
{
CXX_DEBUG_PRINT("Ignore token %s",vStringValue(t->pszWord));
// FIXME: Handle ignore parens!
if(szReplacement && *szReplacement)
{
vStringClear(t->pszWord);
vStringCatS(t->pszWord,szReplacement);
} else {
// skip
cxxTokenChainDestroyLast(g_cxx.pTokenChain);
return cxxParserParseNextToken();
}
}
}
t->bFollowedBySpace = t->bFollowedBySpace | isspace(g_cxx.iChar);
return TRUE;
}
if(g_cxx.iChar == '-')
{
// special case for pointer
vStringPut(t->pszWord,g_cxx.iChar);
g_cxx.iChar = cppGetc();
if(g_cxx.iChar == '>')
{
t->eType = CXXTokenTypePointerOperator;
vStringPut(t->pszWord,g_cxx.iChar);
g_cxx.iChar = cppGetc();
} else {
t->eType = CXXTokenTypeOperator;
if(g_cxx.iChar == '-')
{
vStringPut(t->pszWord,g_cxx.iChar);
g_cxx.iChar = cppGetc();
}
}
t->bFollowedBySpace = isspace(g_cxx.iChar);
return TRUE;
}
#if 0
// As long as we use cppGetc() we don't need this
if(g_cxx.iChar == '"')
{
// special case for strings
t->eType = CXXTokenTypeStringConstant;
vStringPut(t->pszWord,g_cxx.iChar);
// We don't even care of storing the other chars: we don't need
// them for parsing
// FIXME: We might need them in signature:() tag.. maybe add
// them up to a certain length only?
for(;;)
{
g_cxx.iChar = cppGetc();
if(g_cxx.iChar == EOF)
{
t->bFollowedBySpace = FALSE;
return TRUE;
}
if(g_cxx.iChar == '\\')
{
// escape
g_cxx.iChar = cppGetc();
if(g_cxx.iChar == EOF)
{
t->bFollowedBySpace = FALSE;
return TRUE;
}
} else if(g_cxx.iChar == '"')
{
g_cxx.iChar = cppGetc();
break;
}
}
t->bFollowedBySpace = isspace(g_cxx.iChar);
return TRUE;
}
#else
if(g_cxx.iChar == STRING_SYMBOL)
{
t->eType = CXXTokenTypeStringConstant;
vStringPut(t->pszWord,'"');
vStringPut(t->pszWord,'"');
g_cxx.iChar = cppGetc();
t->bFollowedBySpace = isspace(g_cxx.iChar);
return TRUE;
}
#endif
#if 0
// As long as we use cppGetc() we don't need this
if(g_cxx.iChar == '\'')
{
// special case for strings
t->eType = CXXTokenTypeCharacterConstant;
vStringPut(t->pszWord,g_cxx.iChar);
// We don't even care storing the other chars: we don't
// need them for parsing
for(;;)
{
g_cxx.iChar = cppGetc();
if(g_cxx.iChar == EOF)
{
t->bFollowedBySpace = FALSE;
return TRUE;
}
if(g_cxx.iChar == '\\')
{
// escape
g_cxx.iChar = cppGetc();
if(g_cxx.iChar == EOF)
{
t->bFollowedBySpace = FALSE;
return TRUE;
}
} else if(g_cxx.iChar == '\'')
{
g_cxx.iChar = cppGetc();
break;
}
}
t->bFollowedBySpace = isspace(g_cxx.iChar);
return TRUE;
}
#else
if(g_cxx.iChar == CHAR_SYMBOL)
{
t->eType = CXXTokenTypeCharacterConstant;
vStringPut(t->pszWord,'\'');
vStringPut(t->pszWord,'\'');
g_cxx.iChar = cppGetc();
t->bFollowedBySpace = isspace(g_cxx.iChar);
return TRUE;
}
#endif
if(uInfo & CXXCharTypeDecimalDigit)
{
// number
t->eType = CXXTokenTypeNumber;
vStringPut(t->pszWord,g_cxx.iChar);
for(;;)
{
g_cxx.iChar = cppGetc();
uInfo = (g_cxx.iChar < 0x80) ? g_aCharTable[g_cxx.iChar].uType : 0;
if(!(uInfo & CXXCharTypeValidInNumber))
break;
vStringPut(t->pszWord,g_cxx.iChar);
}
t->bFollowedBySpace = isspace(g_cxx.iChar);
return TRUE;
}
if(uInfo & CXXCharTypeNamedSingleOrRepeatedCharToken)
{
t->eType = g_aCharTable[g_cxx.iChar].uSingleTokenType;
vStringPut(t->pszWord,g_cxx.iChar);
int iChar = g_cxx.iChar;
g_cxx.iChar = cppGetc();
if(g_cxx.iChar == iChar)
{
t->eType = g_aCharTable[g_cxx.iChar].uMultiTokenType;
// We could signal a syntax error with more than two colons
// or equal signs...but we're tolerant
do {
vStringPut(t->pszWord,g_cxx.iChar);
g_cxx.iChar = cppGetc();
} while(g_cxx.iChar == iChar);
}
t->bFollowedBySpace = isspace(g_cxx.iChar);
return TRUE;
}
if(uInfo & CXXCharTypeNamedSingleOrOperatorToken)
{
t->eType = g_aCharTable[g_cxx.iChar].uSingleTokenType;
vStringPut(t->pszWord,g_cxx.iChar);
g_cxx.iChar = cppGetc();
uInfo = (g_cxx.iChar < 0x80) ? g_aCharTable[g_cxx.iChar].uType : 0;
if(uInfo & (CXXCharTypeOperator | CXXCharTypeNamedSingleOrOperatorToken))
{
t->eType = CXXTokenTypeOperator;
do {
vStringPut(t->pszWord,g_cxx.iChar);
g_cxx.iChar = cppGetc();
uInfo = (g_cxx.iChar < 0x80) ? g_aCharTable[g_cxx.iChar].uType : 0;
} while(
uInfo &
(CXXCharTypeOperator | CXXCharTypeNamedSingleOrOperatorToken)
);
}
t->bFollowedBySpace = isspace(g_cxx.iChar);
return TRUE;
}
if(uInfo & CXXCharTypeNamedSingleCharToken)
{
t->eType = g_aCharTable[g_cxx.iChar].uSingleTokenType;
vStringPut(t->pszWord,g_cxx.iChar);
g_cxx.iChar = cppGetc();
t->bFollowedBySpace = isspace(g_cxx.iChar);
return TRUE;
}
if(uInfo & CXXCharTypeOperator)
{
t->eType = CXXTokenTypeOperator;
vStringPut(t->pszWord,g_cxx.iChar);
g_cxx.iChar = cppGetc();
uInfo = (g_cxx.iChar < 0x80) ? g_aCharTable[g_cxx.iChar].uType : 0;
while(uInfo & CXXCharTypeOperator)
{
vStringPut(t->pszWord,g_cxx.iChar);
g_cxx.iChar = cppGetc();
uInfo = (g_cxx.iChar < 0x80) ? g_aCharTable[g_cxx.iChar].uType : 0;
}
t->bFollowedBySpace = isspace(g_cxx.iChar);
return TRUE;
}
t->eType = CXXTokenTypeUnknown;
vStringPut(t->pszWord,g_cxx.iChar);
g_cxx.iChar = cppGetc();
t->bFollowedBySpace = isspace(g_cxx.iChar);
return TRUE;
}
