/* Algorithm adapted from from GNU etags.
* Scheme tag functions
* look for (def... xyzzy
* look for (def... (xyzzy
* look for (def ... ((... (xyzzy ....
* look for (set! xyzzy
*/
static void readIdentifier (vString *const name, const unsigned char *cp)
{
const unsigned char *p;
vStringClear (name);
/* Go till you get to white space or a syntactic break */
for (p = cp; *p != '\0' && *p != '(' && *p != ')' && !isspace (*p); p++)
vStringPut (name, (int) *p);
}
static void readIdentifier (lexingState * st)
{
const unsigned char *p;
vStringClear (st->name);
/* first char is a simple letter */
if (isAlpha (*st->cp) || *st->cp == '_')
vStringPut (st->name, (int) *st->cp);
/* Go till you get identifier chars */
for (p = st->cp + 1; isIdent (*p); p++)
vStringPut (st->name, (int) *p);
st->cp = p;
vStringTerminate (st->name);
}
static void parseLet (const unsigned char *line, int infunction)
{
vString *name = vStringNew ();
const unsigned char *cp = line;
const unsigned char *np = line;
/* get the name */
if (isspace ((int) *cp))
{
while (*cp && isspace ((int) *cp))
++cp;
/*
* Ignore lets which set:
* & - local buffer vim settings
* @ - registers
* [ - Lists or Dictionaries
*/
if (!*cp || *cp == '&' || *cp == '@' || *cp == '[' )
goto cleanUp;
/*
* Ignore vim variables which are read only
* v: - Vim variables.
*/
np = cp;
++np;
if ((int) *cp == 'v' && (int) *np == ':' )
goto cleanUp;
/* Skip non-global vars in functions */
if (infunction && (int) *cp != 'g')
goto cleanUp;
/* deal with spaces, $, @ and & */
while (*cp && *cp != '$' && !isalnum ((int) *cp))
++cp;
if (!*cp)
goto cleanUp;
/* cp = skipPrefix (cp, &scope); */
do
{
if (!*cp)
break;
vStringPut (name, (int) *cp);
++cp;
} while (isalnum ((int) *cp) || *cp == '_' || *cp == '#' || *cp == ':' || *cp == '$');
vStringTerminate (name);
makeSimpleTag (name, VimKinds, K_VARIABLE);
vStringClear (name);
}
cleanUp:
vStringDelete (name);
}
static void makeJsTag (tokenInfo *const token, const jsKind kind, vString *const signature)
{
if (JsKinds [kind].enabled && ! token->ignoreTag )
{
const char *name = vStringValue (token->string);
vString *fullscope = vStringNewCopy (token->scope);
const char *p;
tagEntryInfo e;
if ((p = strrchr (name, '.')) != NULL)
{
if (vStringLength (fullscope) > 0)
vStringPut (fullscope, '.');
vStringNCatS (fullscope, name, p - name);
name = p + 1;
}
initTagEntry (&e, name);
e.lineNumber = token->lineNumber;
e.filePosition = token->filePosition;
e.kindName = JsKinds [kind].name;
e.kind = JsKinds [kind].letter;
if ( vStringLength(fullscope) > 0 )
{
jsKind parent_kind = JSTAG_CLASS;
/* if we're creating a function (and not a method),
* guess we're inside another function */
if (kind == JSTAG_FUNCTION)
parent_kind = JSTAG_FUNCTION;
e.extensionFields.scope[0] = JsKinds [parent_kind].name;
e.extensionFields.scope[1] = vStringValue (fullscope);
}
if (signature && vStringLength(signature))
{
size_t i;
/* sanitize signature by replacing all control characters with a
* space (because it's simple).
* there should never be any junk in a valid signature, but who
* knows what the user wrote and CTags doesn't cope well with weird
* characters. */
for (i = 0; i < signature->length; i++)
{
unsigned char c = (unsigned char) signature->buffer[i];
if (c < 0x20 /* below space */ || c == 0x7F /* DEL */)
signature->buffer[i] = ' ';
}
e.extensionFields.arglist = vStringValue(signature);
}
makeTagEntry (&e);
vStringDelete (fullscope);
}
}
/* Match the name of a tag (function, variable, type, ...) starting at pos. */
static char const *extract_name (char const *pos, vString * name)
{
while (isspace (*pos))
pos++;
vStringClear (name);
for (; *pos && !isspace (*pos) && *pos != '(' && *pos != ',' && *pos != '='; pos++)
vStringPut (name, *pos);
return pos;
}
static void scanIdentifier (lexerState *lexer)
{
vStringClear(lexer->token_str);
do
{
vStringPut(lexer->token_str, (char) lexer->cur_c);
advanceChar(lexer);
} while(lexer->cur_c != EOF && isIdentifierContinue(lexer->cur_c));
}
/* parses a class or an interface:
* class Foo {}
* class Foo extends Bar {}
* class Foo extends Bar implements iFoo, iBar {}
* interface iFoo {}
* interface iBar extends iFoo {}
*
* if @name is not NULL, parses an anonymous class with name @name
* new class {}
* new class(1, 2) {}
* new class(1, 2) extends Foo implements iFoo, iBar {} */
static boolean parseClassOrIface (tokenInfo *const token, const phpKind kind,
const tokenInfo *name)
{
boolean readNext = TRUE;
implType impl = CurrentStatement.impl;
tokenInfo *nameFree = NULL;
vString *inheritance = NULL;
readToken (token);
if (name) /* anonymous class */
{
/* skip possible construction arguments */
skipOverParens (token);
}
else /* normal, named class */
{
if (token->type != TOKEN_IDENTIFIER)
return FALSE;
name = nameFree = newToken ();
copyToken (nameFree, token, TRUE);
readToken (token);
}
inheritance = vStringNew ();
/* read every identifiers, keywords and commas, and assume each
* identifier (not keyword) is an inheritance
* (like in "class Foo extends Bar implements iA, iB") */
while (token->type == TOKEN_IDENTIFIER ||
token->type == TOKEN_KEYWORD ||
token->type == TOKEN_COMMA)
{
if (token->type == TOKEN_IDENTIFIER)
{
if (vStringLength (inheritance) > 0)
vStringPut (inheritance, ',');
vStringCat (inheritance, token->string);
}
readToken (token);
}
makeClassOrIfaceTag (kind, name, inheritance, impl);
if (token->type == TOKEN_OPEN_CURLY)
enterScope (token, name->string, kind);
else
readNext = FALSE;
if (nameFree)
deleteToken (nameFree);
vStringDelete (inheritance);
return readNext;
}
static void parseSelector (vString *const string, const int firstChar)
{
int c = firstChar;
do
{
vStringPut (string, (char) c);
c = getcFromInputFile ();
} while (isSelectorChar (c));
ungetcToInputFile (c);
}
static void parseString (vString * const string, const int delimiter)
{
bool end = false;
while (!end)
{
int c = getcFromInputFile ();
if (c == EOF)
end = true;
else if (c == '\\')
{
c = getcFromInputFile (); /* This maybe a ' or ". */
vStringPut (string, c);
}
else if (c == delimiter)
end = true;
else
vStringPut (string, c);
}
}
static void parseTemplateString (vString *const string)
{
int c;
do
{
c = fileGetc ();
if (c == '`')
break;
vStringPut (string, c);
if (c == '\\')
{
c = fileGetc();
vStringPut(string, c);
}
else if (c == '$')
{
c = fileGetc ();
if (c != '{')
fileUngetc (c);
else
{
int depth = 1;
/* we need to use the real token machinery to handle strings,
* comments, regexes and whatnot */
tokenInfo *token = newToken ();
LastTokenType = TOKEN_UNDEFINED;
vStringPut(string, c);
do
{
readTokenFull (token, FALSE, string);
if (isType (token, TOKEN_OPEN_CURLY))
depth++;
else if (isType (token, TOKEN_CLOSE_CURLY))
depth--;
}
while (! isType (token, TOKEN_EOF) && depth > 0);
deleteToken (token);
}
}
}
while (c != EOF);
vStringTerminate (string);
}
static void add_tag(const char *token, haskellKind kind, vString *name)
{
int i;
for (i = 0; token[i] != '\0'; ++i)
vStringPut(name, token[i]);
vStringTerminate(name);
makeSimpleTag(name, HaskellKinds, kind);
vStringClear(name);
}
/* Reads an identifier, whose first character is given by "c", into "tag",
* together with the file location and corresponding line number.
*/
static void readIdentifier (int c, vString *const name)
{
vStringClear (name);
do
{
vStringPut (name, c);
c = getcAndCollect ();
} while (c != EOF && cppIsident (c));
ungetcAndCollect (c);
}
static void pushScope (tokenInfo *const token,
const tokenInfo *const parent,
const jsonKind parentKind)
{
if (vStringLength (token->scope) > 0)
vStringPut (token->scope, '.');
vStringCat (token->scope, parent->string);
vStringTerminate (token->scope);
token->scopeKind = parentKind;
}
static void readAndEmitTag (const unsigned char** cp, rubyKind expected_kind)
{
if (isspace (**cp))
{
vString *name = vStringNew ();
vString *chunk = vStringNew ();
rubyKind actual_kind;
unsigned int i = 0;
/* parse the identifier, allowing scoping like "class Foo::Bar::Baz" */
while (1)
{
actual_kind = parseIdentifier (cp, chunk, expected_kind);
if (i++ > 0)
vStringPut (name, '.');
vStringCat (name, chunk);
vStringClear (chunk);
if (actual_kind != K_UNDEFINED && (*cp)[0] == ':' && (*cp)[1] == ':')
*cp += 2;
else
break;
}
vStringDelete (chunk);
vStringTerminate (name);
if (actual_kind == K_UNDEFINED || vStringLength (name) == 0)
{
/*
* What kind of tags should we create for code like this?
*
* %w(self.clfloor clfloor).each do |name|
* module_eval <<-"end;"
* def #{name}(x, y=1)
* q, r = x.divmod(y)
* q = q.to_i
* return q, r
* end
* end;
* end
*
* Or this?
*
* class << HTTP
*
* For now, we don't create any.
*/
}
else
{
emitRubyTag (name, actual_kind);
}
vStringDelete (name);
}
}
/*
* Emits a tag for the given 'name' of kind 'kind' at the current nesting.
*/
static void emitRubyTag (vString* name, rubyKind kind)
{
tagEntryInfo tag;
vString* scope;
tagEntryInfo *parent;
rubyKind parent_kind = K_UNDEFINED;
NestingLevel *lvl;
const char *unqualified_name;
const char *qualified_name;
int r;
if (!RubyKinds[kind].enabled) {
return;
}
vStringTerminate (name);
scope = nestingLevelsToScope (nesting);
lvl = nestingLevelsGetCurrent (nesting);
parent = getEntryOfNestingLevel (lvl);
if (parent)
parent_kind = parent->kind - RubyKinds;
qualified_name = vStringValue (name);
unqualified_name = strrchr (qualified_name, SCOPE_SEPARATOR);
if (unqualified_name && unqualified_name[1])
{
if (unqualified_name > qualified_name)
{
if (vStringLength (scope) > 0)
vStringPut (scope, SCOPE_SEPARATOR);
vStringNCatS (scope, qualified_name,
unqualified_name - qualified_name);
/* assume module parent type for a lack of a better option */
parent_kind = K_MODULE;
}
unqualified_name++;
}
else
unqualified_name = qualified_name;
initTagEntry (&tag, unqualified_name, &(RubyKinds [kind]));
if (vStringLength (scope) > 0) {
Assert (0 <= parent_kind &&
(size_t) parent_kind < (ARRAY_SIZE (RubyKinds)));
tag.extensionFields.scopeKind = &(RubyKinds [parent_kind]);
tag.extensionFields.scopeName = vStringValue (scope);
}
r = makeTagEntry (&tag);
nestingLevelsPush (nesting, r);
vStringClear (name);
vStringDelete (scope);
}
/**
* Reads from 'start' to 'end' and eliminates all spaces
* inbetween.
*/
static char *readBetweenDelimitersWhileTrimmingSpaces(char start, char end) {
static vString *wordBuffer = 0;
int z;
if (!wordBuffer)
wordBuffer = vStringNew();
else
vStringClear(wordBuffer);
z = skipToNonWhite();
if (z != start) return NULL;
while ((z = cppGetc()) != EOF && z != end) {
if (isspace(z)) continue;
vStringPut(wordBuffer, z);
}
if (z == EOF) return NULL;
vStringPut(wordBuffer, z);
vStringPut(wordBuffer, 0);
return vStringValue(wordBuffer);
}
static void parseFunction (const unsigned char *line)
{
vString *name = vStringNew ();
/* boolean inFunction = FALSE; */
int scope;
const unsigned char *cp = line + 1;
if ((int) *++cp == 'n' && (int) *++cp == 'c' &&
(int) *++cp == 't' && (int) *++cp == 'i' &&
(int) *++cp == 'o' && (int) *++cp == 'n')
++cp;
if ((int) *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 */
{
do
{
vStringPut (name, (int) *cp);
++cp;
} while (isalnum ((int) *cp) || *cp == '_' || *cp == '.' || *cp == '#');
vStringTerminate (name);
makeSimpleTag (name, VimKinds, K_FUNCTION);
vStringClear (name);
}
}
}
/* TODO - update struct to indicate inside function */
while ((line = readVimLine ()) != NULL)
{
/*
* Vim7 added the for/endfo[r] construct, so we must first
* check for an "endfo", before a "endf"
*/
if ( (!strncmp ((const char*) line, "endfo", (size_t) 5) == 0) &&
(strncmp ((const char*) line, "endf", (size_t) 4) == 0) )
break;
/* TODO - call parseVimLine */
}
vStringDelete (name);
}
static void parseIdentifier (vString *const string, const int firstChar)
{
int c = firstChar;
do
{
vStringPut (string, (char) c);
c = getcFromInputFile ();
} while (isIdentChar (c));
ungetcToInputFile (c);
vStringTerminate (string);
}
/*
* Copies the characters forming an identifier from *cp into
* name, leaving *cp pointing to the character after the identifier.
*/
static rubyKind parseIdentifier (
const unsigned char** cp, vString* name, rubyKind kind)
{
/* Method names are slightly different to class and variable names.
* A method name may optionally end with a question mark, exclamation
* point or equals sign. These are all part of the name.
* A method name may also contain a period if it's a singleton method.
*/
const char* also_ok = (kind == K_METHOD) ? "_.?!=" : "_";
skipWhitespace (cp);
/* Check for an anonymous (singleton) class such as "class << HTTP". */
if (kind == K_CLASS && **cp == '<' && *(*cp + 1) == '<')
{
return K_UNDEFINED;
}
/* Check for operators such as "def []=(key, val)". */
if (kind == K_METHOD || kind == K_SINGLETON)
{
if (parseRubyOperator (name, cp))
{
return kind;
}
}
/* Copy the identifier into 'name'. */
while (**cp != 0 && (isalnum (**cp) || charIsIn (**cp, also_ok)))
{
char last_char = **cp;
vStringPut (name, last_char);
++*cp;
if (kind == K_METHOD)
{
/* Recognize singleton methods. */
if (last_char == '.')
{
vStringTerminate (name);
vStringClear (name);
return parseIdentifier (cp, name, K_SINGLETON);
}
/* Recognize characters which mark the end of a method name. */
if (charIsIn (last_char, "?!="))
{
break;
}
}
}
return kind;
}
static const unsigned char *parseIdentifier (
const unsigned char *cp, vString *const identifier)
{
vStringClear (identifier);
while (isIdentifierCharacter ((int) *cp))
{
vStringPut (identifier, (int) *cp);
++cp;
}
return cp;
}
static void parseString (vString * const string, const int delimiter)
{
boolean end = FALSE;
while (!end)
{
int c = fileGetc ();
if (c == EOF)
end = TRUE;
else if (c == '\\')
{
c = fileGetc (); /* This maybe a ' or ". */
vStringPut (string, c);
}
else if (c == delimiter)
end = TRUE;
else
vStringPut (string, c);
}
vStringTerminate (string);
}
static void parseIdentifier (vString *const string, const int firstChar)
{
int c = firstChar;
do
{
vStringPut (string, c);
c = getcFromInputFile ();
} while (isIdentChar (c));
vStringTerminate (string);
ungetcToInputFile (c); /* always unget, LF might add a semicolon */
}
static void parseSelector (vString *const string, const int firstChar)
{
int c = firstChar;
do
{
vStringPut (string, (char) c);
c = fileGetc ();
} while (isSelectorChar (c));
fileUngetc (c);
vStringTerminate (string);
}
static void parseString (vString *const string, const int delimiter)
{
boolean end = FALSE;
while (! end)
{
int c = fileGetc ();
if (c == EOF)
end = TRUE;
else if (c == '\\')
{
/* Eat the escape sequence (\", \', etc). We properly handle
* <LineContinuation> by eating a whole \<CR><LF> not to see <LF>
* as an unescaped character, which is invalid and handled below.
* Also, handle the fact that <LineContinuation> produces an empty
* sequence.
* See ECMA-262 7.8.4 */
c = fileGetc();
if (c != '\r' && c != '\n')
vStringPut(string, c);
else if (c == '\r')
{
c = fileGetc();
if (c != '\n')
fileUngetc (c);
}
}
else if (c == delimiter)
end = TRUE;
else if (c == '\r' || c == '\n')
{
/* those are invalid when not escaped */
end = TRUE;
/* we don't want to eat the newline itself to let the automatic
* semicolon insertion code kick in */
fileUngetc (c);
}
else
vStringPut (string, c);
}
vStringTerminate (string);
}
/* While ANSI only permits lines of the form:
* # line n "filename"
* Earlier compilers generated lines of the form
* # n filename
* GNU C will output lines of the form:
* # n "filename"
* So we need to be fairly flexible in what we accept.
*/
static vString *readFileName (char *s)
{
vString *const fileName = vStringNew ();
bool quoteDelimited = false;
skipWhite (&s);
if (*s == '"')
{
s++; /* skip double-quote */
quoteDelimited = true;
}
while (*s != '\0' && *s != '\n' &&
(quoteDelimited ? (*s != '"') : (*s != ' ' && *s != '\t')))
{
vStringPut (fileName, *s);
s++;
}
vStringPut (fileName, '\0');
return fileName;
}
/* Read a VHDL identifier beginning with "firstChar" and place it into "name".
*/
static void parseIdentifier (vString * const string, const int firstChar)
{
int c = firstChar;
Assert (isIdentChar1 (c));
do
{
vStringPut (string, c);
c = getcFromInputFile ();
} while (isIdentChar (c));
if (!isspace (c))
ungetcToInputFile (c); /* unget non-identifier character */
}
/* Read a C identifier beginning with "firstChar" and places it into
* "name".
*/
static void parseIdentifier (vString *const string, const int firstChar)
{
int c = firstChar;
Assert (isIdentChar (c));
do
{
vStringPut (string, c);
c = fileGetc ();
} while (isIdentChar (c));
vStringTerminate (string);
fileUngetc (c); /* unget non-identifier character */
}
static void readIdentifier (const int first, vString *const id)
{
int c = first;
vStringClear (id);
while (isIdentifier (c))
{
vStringPut (id, c);
c = nextChar ();
}
fileUngetc (c);
vStringTerminate (id);
}
CXXToken * cxxTokenChainExtractRange(
CXXToken * from,
CXXToken * to,
unsigned int uFlags
)
{
if(!from)
return NULL;
CXXToken * pToken = from;
CXXToken * pRet = cxxTokenCreate();
pRet->iLineNumber = pToken->iLineNumber;
pRet->oFilePosition = pToken->oFilePosition;
pRet->eType = pToken->eType;
cxxTokenAppendToString(pRet->pszWord,pToken);
if(
(!(uFlags & CXXTokenChainExtractRangeNoTrailingSpaces)) &&
pToken->bFollowedBySpace
)
vStringPut (pRet->pszWord, ' ');
pRet->bFollowedBySpace = pToken->bFollowedBySpace;
while(pToken != to)
{
pToken = pToken->pNext;
if(!pToken)
return pRet;
cxxTokenAppendToString(pRet->pszWord,pToken);
if(
(!(uFlags & CXXTokenChainExtractRangeNoTrailingSpaces)) &&
pToken->bFollowedBySpace
)
vStringPut (pRet->pszWord, ' ');
pRet->bFollowedBySpace = pToken->bFollowedBySpace;
}
return pRet;
}
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;
}
}
