Merge pull request #3690 from b4n/vstring/type

Improve vStringPut() usage safety and fix problematic callers.

dsl/optscript.c

@@ -4184,7 +4184,7 @@ op__put_str (OptVM *vm, EsObject *name,
 		for (size_t i = 0; i < d; i++)
 			vStringPut (vstr, ' ');
 		if (c != 0)
-			vStringPut (vstr, (char)c);
+			vStringPut (vstr, c);
 	}
 
 	ptrArrayDeleteLastInBatch (vm->ostack, 3);

main/field.c

@@ -863,13 +863,12 @@ extern bool  doesFieldHaveTabOrNewlineChar (fieldType type, const tagEntryInfo *
 static const char* renderCompactInputLine (vString *b,  const char *const line)
 {
 	bool lineStarted = false;
-	const char *p;
-	int c;
 
 	/*  Write everything up to, but not including, the newline.
 	 */
-	for (p = line, c = *p  ;  c != NEWLINE  &&  c != '\0'  ;  c = *++p)
+	for (const char *p = line; *p != NEWLINE && *p != '\0'; ++p)
 	{
+		int c = (unsigned char) *p;
 		if (lineStarted  || ! isspace (c))  /* ignore leading spaces */
 		{
 			lineStarted = true;

main/options.c

@@ -1286,7 +1286,7 @@ static void processExtraTagsOption (
 
 	longName = vStringNewOrClearWithAutoRelease (longName);
 
-	while ((c = *p++) != '\0')
+	while ((c = (unsigned char) *p++) != '\0')
 	{
 		switch (c)
 		{
@@ -1374,7 +1374,7 @@ static void processFieldsOption (
 	else if (*p != '+'  &&  *p != '-')
 		resetFieldsOption (LANG_IGNORE, false);
 
-	while ((c = *p++) != '\0') switch (c)
+	while ((c = (unsigned char) *p++) != '\0') switch (c)
 	{
 		case '+':
 			if (inLongName)
@@ -3084,7 +3084,7 @@ static bool processLangSpecificFieldsOption (const char *const option,
 		error (WARNING, "Wrong per language field specification: %s", p);
 
 	longName = vStringNewOrClearWithAutoRelease (longName);
-	while ((c = *p++) != '\0')
+	while ((c = (unsigned char) *p++) != '\0')
 	{
 		switch (c)
 		{
@@ -3185,7 +3185,7 @@ static bool processLangSpecificExtraOption (const char *const option,
 		error (WARNING, "Wrong per language extra specification: %s", p);
 
 	longName = vStringNewOrClearWithAutoRelease (longName);
-	while ((c = *p++) != '\0')
+	while ((c = (unsigned char) *p++) != '\0')
 	{
 		switch (c)
 		{

main/parse.c

@@ -751,7 +751,7 @@ static vString* determineInterpreter (const char* const cmd)
 		for ( ;  isspace ((int) *p)  ;  ++p)
 			;  /* no-op */
 		for ( ;  *p != '\0'  &&  ! isspace ((int) *p)  ;  ++p)
-			vStringPut (interpreter, (int) *p);
+			vStringPut (interpreter, *p);
 	} while (strcmp (vStringValue (interpreter), "env") == 0);
 	return interpreter;
 }
@@ -823,7 +823,7 @@ static vString* determineEmacsModeAtFirstLine (const char* const line)
 		for ( ;  isspace ((int) *p)  ;  ++p)
 			;  /* no-op */
 		for ( ;  *p != '\0'  &&  isLanguageNameChar ((int) *p)  ;  ++p)
-			vStringPut (mode, (int) *p);
+			vStringPut (mode, *p);
 
 		if ((strcmp(vStringValue (mode), "sh") == 0
 			 || strcmp(vStringValue (mode), "shell-script") == 0)
@@ -839,7 +839,7 @@ static vString* determineEmacsModeAtFirstLine (const char* const line)
 			goto out;
 
 		for ( ;  p < end &&  isLanguageNameChar ((int) *p)  ;  ++p)
-			vStringPut (mode, (int) *p);
+			vStringPut (mode, *p);
 
 		for ( ;  isspace ((int) *p)  ;  ++p)
 			;  /* no-op */
@@ -891,7 +891,7 @@ static vString* determineEmacsModeAtEOF (MIO* const fp)
 			for ( ;  isspace ((int) *p)  ;  ++p)
 				;  /* no-op */
 			for ( ;  *p != '\0'  &&  isLanguageNameChar ((int) *p)  ;  ++p)
-				vStringPut (mode, (int) *p);
+				vStringPut (mode, *p);
 
 			is_shell_mode = ((strcasecmp (vStringValue (mode), "sh") == 0
 								 || strcasecmp (vStringValue (mode), "shell-script") == 0));
@@ -955,7 +955,7 @@ static vString* determineVimFileType (const char *const modeline)
 
 		p += strlen(filetype_prefix[i]);
 		for ( ;  *p != '\0'  &&  isalnum ((int) *p)  ;  ++p)
-			vStringPut (filetype, (int) *p);
+			vStringPut (filetype, *p);
 		break;
 	}
 	return filetype;
@@ -2542,7 +2542,7 @@ static void processLangKindDefinition (
 
 	longName = vStringNewOrClearWithAutoRelease (longName);
 
-	while ((c = *p++) != '\0')
+	while ((c = (unsigned char) *p++) != '\0')
 	{
 		switch (c)
 		{

main/vstring.c

@@ -244,11 +244,7 @@ extern void vStringCopyToLower (vString *const dest, const vString *const src)
 		vStringResize (dest, src->size);
 	d = dest->buffer;
 	for (i = 0  ;  i < length  ;  ++i)
-	{
-		int c = s [i];
-
-		d [i] = tolower (c);
-	}
+		d [i] = (char) tolower (s [i]);
 	d [i] = '\0';
 }
 
@@ -294,7 +290,7 @@ extern char    *vStringStrdup (const vString *const string)
 	return str;
 }
 
-static char valueToXDigit (int v)
+static int valueToXDigit (int v)
 {
 	Assert (v >= 0 && v <= 0xF);
 
@@ -308,7 +304,7 @@ extern void vStringCatSWithEscaping (vString* b, const char *s)
 {
 	for(; *s; s++)
 	{
-		int c = *s;
+		unsigned char c = (unsigned char) *s;
 
 		/* escape control characters (incl. \t) */
 		if ((c > 0x00 && c <= 0x1F) || c == 0x7F || c == '\\')

main/vstring.h

@@ -18,6 +18,7 @@
 
 #include <stdio.h>
 
+#include "debug.h"
 #include "inline.h"
 
 /*
@@ -101,18 +102,25 @@ CTAGS_INLINE void vStringPutNewlinAgainUnsafe (vString *const string)
 	string->buffer [string->length++] = '\n';
 }
 
-CTAGS_INLINE void vStringPut (vString *const string, const int c)
+CTAGS_INLINE void vStringPutImpl (vString *const string, const int c)
 {
+	/* verify the given character is an unsigned char value */
+	Assert (c >= 0 && c <= 0xff);
+
 	if (string->length + 1 == string->size)  /*  check for buffer overflow */
 		vStringResize (string, string->size * 2);
 
-	string->buffer [string->length] = c;
+	string->buffer [string->length] = (char) c;
 	if (c != '\0')
 		string->buffer [++string->length] = '\0';
 }
 
-CTAGS_INLINE bool vStringPutWithLimit (vString *const string, const int c,
-									   unsigned int maxlen)
+#define vStringPut(s, c) (sizeof(c) == sizeof(char) \
+						  ? vStringPutImpl((s), (unsigned char) (c)) \
+						  : vStringPutImpl((s), (c)))
+
+CTAGS_INLINE bool vStringPutWithLimitImpl (vString *const string, const int c,
+										   unsigned int maxlen)
 {
 	if (vStringLength (string) < maxlen || maxlen == 0)
 	{
@@ -122,4 +130,9 @@ CTAGS_INLINE bool vStringPutWithLimit (vString *const string, const int c,
 	return false;
 }
 
+#define vStringPutWithLimit(s, c, l) \
+	(sizeof(c) == sizeof(char) \
+	 ? vStringPutWithLimitImpl((s), (unsigned char) (c), (l)) \
+	 : vStringPutWithLimitImpl((s), (c), (l)))
+
 #endif  /* CTAGS_MAIN_VSTRING_H */

old-docs/EXTENDING.html

@@ -239,7 +239,7 @@ <h2>Examples</h2>
                 ++cp;
             while (isalnum ((int) *cp)  ||  *cp == '_')
             {
-                vStringPut (name, (int) *cp);
+                vStringPut (name, *cp);
                 ++cp;
             }
             makeSimpleTag (name, SwineKinds, K_DEFINE);

parsers/abaqus.c

@@ -70,7 +70,7 @@ static void createTag(AbaqusKind kind, const char *buf)
 
 	do
 	{
-		vStringPut(name, (int) *buf);
+		vStringPut(name, *buf);
 		++buf;
 	} while ((*buf != '\0') && (*buf != ','));
 	makeSimpleTag(name, kind);

parsers/asp.c

@@ -127,7 +127,7 @@ static void findAspTags (void)
 						    ++cp;
 					    while (isalnum ((int) *cp)  ||  *cp == '_')
 					    {
-						    vStringPut (name, (int) *cp);
+						    vStringPut (name, *cp);
 						    ++cp;
 					    }
 					    makeSimpleTag (name, K_FUNCTION);
@@ -140,7 +140,7 @@ static void findAspTags (void)
 						    ++cp;
 					    while (isalnum ((int) *cp)  ||  *cp == '_')
 					    {
-						    vStringPut (name, (int) *cp);
+						    vStringPut (name, *cp);
 						    ++cp;
 					    }
 					    makeSimpleTag (name, K_SUB);
@@ -149,7 +149,7 @@ static void findAspTags (void)
 					else {
 					    while (isalnum ((int) *cp)  ||  *cp == '_')
 					    {
-						    vStringPut (name, (int) *cp);
+						    vStringPut (name, *cp);
 						    ++cp;
 					    }
 					    makeSimpleTag (name, K_DIM);
@@ -171,7 +171,7 @@ static void findAspTags (void)
 						    ++cp;
 					    while (isalnum ((int) *cp)  ||  *cp == '_')
 					    {
-						    vStringPut (name, (int) *cp);
+						    vStringPut (name, *cp);
 						    ++cp;
 					    }
 					    makeSimpleTag (name, K_FUNCTION);
@@ -184,7 +184,7 @@ static void findAspTags (void)
 						    ++cp;
 					    while (isalnum ((int) *cp)  ||  *cp == '_')
 					    {
-						    vStringPut (name, (int) *cp);
+						    vStringPut (name, *cp);
 						    ++cp;
 					    }
 					    makeSimpleTag (name, K_SUB);
@@ -193,7 +193,7 @@ static void findAspTags (void)
 					else {
 					    while (isalnum ((int) *cp)  ||  *cp == '_')
 					    {
-						    vStringPut (name, (int) *cp);
+						    vStringPut (name, *cp);
 						    ++cp;
 					    }
 					    makeSimpleTag (name, K_DIM);
@@ -213,7 +213,7 @@ static void findAspTags (void)
 						++cp;
 					while (isalnum ((int) *cp)  ||  *cp == '_')
 					{
-						vStringPut (name, (int) *cp);
+						vStringPut (name, *cp);
 						++cp;
 					}
 					makeSimpleTag (name, K_FUNCTION);
@@ -231,7 +231,7 @@ static void findAspTags (void)
 						++cp;
 					while (isalnum ((int) *cp)  ||  *cp == '_')
 					{
-						vStringPut (name, (int) *cp);
+						vStringPut (name, *cp);
 						++cp;
 					}
 					makeSimpleTag (name, K_SUB);
@@ -249,7 +249,7 @@ static void findAspTags (void)
 						++cp;
 					while (isalnum ((int) *cp)  ||  *cp == '_')
 					{
-						vStringPut (name, (int) *cp);
+						vStringPut (name, *cp);
 						++cp;
 					}
 					makeSimpleTag (name, K_DIM);
@@ -267,7 +267,7 @@ static void findAspTags (void)
 						++cp;
 					while (isalnum ((int) *cp)  ||  *cp == '_')
 					{
-						vStringPut (name, (int) *cp);
+						vStringPut (name, *cp);
 						++cp;
 					}
 					makeSimpleTag (name, K_CLASS);
@@ -285,7 +285,7 @@ static void findAspTags (void)
 						++cp;
 					while (isalnum ((int) *cp)  ||  *cp == '_')
 					{
-						vStringPut (name, (int) *cp);
+						vStringPut (name, *cp);
 						++cp;
 					}
 					makeSimpleTag (name, K_CONST);

parsers/autoit.c

@@ -145,7 +145,7 @@ static int parseFunc (const unsigned char *p, NestingLevels *nls)
 	skipSpaces (&p);
 	while (isIdentChar ((int) *p))
 	{
-		vStringPut (name, (int) *p);
+		vStringPut (name, *p);
 		++p;
 	}
 	skipSpaces (&p);
@@ -154,7 +154,7 @@ static int parseFunc (const unsigned char *p, NestingLevels *nls)
 		vString *signature = vStringNew ();
 
 		do
-			vStringPut (signature, (int) *p);
+			vStringPut (signature, *p);
 		while (*p != ')' && *p++);
 
 		k = makeAutoItTag (nls, name, K_FUNCTION, signature);
@@ -218,7 +218,7 @@ static void findAutoItTags (void)
 					++p;
 					while (*p != '\0' && *p != '>' && *p != '"')
 					{
-						vStringPut (name, (int) *p);
+						vStringPut (name, *p);
 						++p;
 					}
 					if (vStringLength(name) > 0)
@@ -284,7 +284,7 @@ static void findAutoItTags (void)
 					p++;
 					while (isIdentChar ((int) *p))
 					{
-						vStringPut (name, (int) *p);
+						vStringPut (name, *p);
 						++p;
 					}
 					if (vStringLength(name) > 0)

parsers/awk.c

@@ -53,7 +53,7 @@ static void findAwkTags (void)
 				++cp;
 			while (isalnum ((int) *cp)  ||  *cp == '_')
 			{
-				vStringPut (name, (int) *cp);
+				vStringPut (name, *cp);
 				++cp;
 			}
 			while (isspace ((int) *cp))

parsers/c-based.c

@@ -1486,7 +1486,22 @@ static void skipToMatch (const char *const pair)
 	while (matchLevel > 0  &&  (c = skipToNonWhite ()) != EOF)
 	{
 		if (CollectingSignature)
-			vStringPut (Signature, c);
+		{
+			if (c <= 0xff)
+				vStringPut (Signature, c);
+			else
+			{
+				switch (c)
+				{
+					case CHAR_SYMBOL:
+					case STRING_SYMBOL:
+						vStringCat (Signature, cppGetLastCharOrStringContents ());
+						break;
+					default:
+						AssertNotReached();
+				}
+			}
+		}
 		if (c == begin)
 		{
 			++matchLevel;

parsers/css.c

@@ -58,7 +58,7 @@ static void parseSelector (vString *const string, const int firstChar)
 	int c = firstChar;
 	do
 	{
-		vStringPut (string, (char) c);
+		vStringPut (string, c);
 		c = getcFromInputFile ();
 	} while (isSelectorChar (c));
 	ungetcToInputFile (c);

parsers/erlang.c

@@ -66,7 +66,7 @@ static const unsigned char *parseIdentifier (
 	vStringClear (identifier);
 	while (isIdentifierCharacter ((int) *cp))
 	{
-		vStringPut (identifier, (int) *cp);
+		vStringPut (identifier, *cp);
 		++cp;
 	}
 	return cp;