ScummVM

Revision: 39369
          http://scummvm.svn.sourceforge.net/scummvm/?rev=39369&view=rev
Author:   thebluegr
Date:     2009-03-13 07:56:06 +0000 (Fri, 13 Mar 2009)

Log Message:
-----------
Added an explanatory comment for commit #39368

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-03-13 01:36:07 UTC (rev 39368)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-03-13 07:56:06 UTC (rev 39369)
@@ -414,6 +414,10 @@
 	if (len == 0) return 1;
 
 	while (*str) {
+		// We're ANDing the string with 0xFF to prevent a situation
+		// where MSVC could sometimes falsely parse the character as
+		// multibyte/Unicode, thereby overflowing an assertion inside
+		// isprint(). This occurs after the intro of LSL5, for example.
 		if (isprint((*str & 0xFF))) printable++;
 		str++;
 	}


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 39370
          http://scummvm.svn.sourceforge.net/scummvm/?rev=39370&view=rev
Author:   thebluegr
Date:     2009-03-13 10:36:57 +0000 (Fri, 13 Mar 2009)

Log Message:
-----------
Changed the parameter passed to isprint() to be unsigned, according to MSDN (thanks wjp)

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-03-13 07:56:06 UTC (rev 39369)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-03-13 10:36:57 UTC (rev 39370)
@@ -414,11 +414,12 @@
 	if (len == 0) return 1;
 
 	while (*str) {
-		// We're ANDing the string with 0xFF to prevent a situation
-		// where MSVC could sometimes falsely parse the character as
-		// multibyte/Unicode, thereby overflowing an assertion inside
-		// isprint(). This occurs after the intro of LSL5, for example.
-		if (isprint((*str & 0xFF))) printable++;
+		// The parameter passed to isprint() needs to be in the range
+		// 0 to 0xFF or EOF, according to MSDN, therefore we cast it
+		// to an unsigned char. Values outside this range (in this
+		// case, negative values) yield unpredictable results. Refer to:
+		// http://msdn.microsoft.com/en-us/library/ewx8s4kw.aspx
+		if (isprint((unsigned char)*str)) printable++;
 		str++;
 	}
 


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 39371
          http://scummvm.svn.sourceforge.net/scummvm/?rev=39371&view=rev
Author:   fingolfin
Date:     2009-03-13 16:31:38 +0000 (Fri, 13 Mar 2009)

Log Message:
-----------
SCI: cleanup

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-03-13 10:36:57 UTC (rev 39370)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-03-13 16:31:38 UTC (rev 39371)
@@ -407,7 +407,7 @@
 /* Simple heuristic to work around array handling peculiarity in SQ4:
 It uses StrAt() to read the individual elements, so we must determine
 whether a string is really a string or an array. */
-static int is_print_str(char *str) {
+static int is_print_str(const char *str) {
 	int printable = 0;
 	int len = strlen(str);
 
@@ -419,7 +419,8 @@
 		// to an unsigned char. Values outside this range (in this
 		// case, negative values) yield unpredictable results. Refer to:
 		// http://msdn.microsoft.com/en-us/library/ewx8s4kw.aspx
-		if (isprint((unsigned char)*str)) printable++;
+		if (isprint((byte)*str))
+			printable++;
 		str++;
 	}
 
@@ -428,20 +429,18 @@
 
 
 reg_t kStrAt(EngineState *s, int funct_nr, int argc, reg_t *argv) {
-	unsigned char *dest = (unsigned char *) kernel_dereference_bulk_pointer(s, argv[0], 0);
+	byte *dest = (byte *)kernel_dereference_bulk_pointer(s, argv[0], 0);
 	reg_t *dest2;
 
 	if (!dest) {
-		warning("Attempt to StrAt at invalid pointer "PREG"",
-		          PRINT_REG(argv[0]));
+		warning("Attempt to StrAt at invalid pointer "PREG"", PRINT_REG(argv[0]));
 		return NULL_REG;
 	}
 
 	if ((argc == 2) &&
 	        /* Our pathfinder already works around the issue we're trying to fix */
-	        (strcmp(s->seg_manager->getDescription(argv[0]),
-	                AVOIDPATH_DYNMEM_STRING) != 0)  &&
-	        ((strlen((const char*)dest) < 2) || (!is_print_str((char*)dest))))
+	        (strcmp(s->seg_manager->getDescription(argv[0]), AVOIDPATH_DYNMEM_STRING) != 0)  &&
+	        ((strlen((const char*)dest) < 2) || (!is_print_str((const char*)dest))))
 		/* SQ4 array handling detected */
 	{
 #ifndef SCUMM_BIG_ENDIAN
@@ -450,8 +449,9 @@
 		int odd = !(KP_UINT(argv[1]) & 1);
 #endif
 		dest2 = ((reg_t *) dest) + (KP_UINT(argv[1]) / 2);
-		dest = ((unsigned char *)(&dest2->offset)) + odd;
-	} else dest += KP_UINT(argv[1]);
+		dest = ((byte *)(&dest2->offset)) + odd;
+	} else
+		dest += KP_UINT(argv[1]);
 
 	s->r_acc = make_reg(0, *dest);
 


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 39673
          http://scummvm.svn.sourceforge.net/scummvm/?rev=39673&view=rev
Author:   joostp
Date:     2009-03-24 19:56:41 +0000 (Tue, 24 Mar 2009)

Log Message:
-----------
Fix smaller error introduced in r39671 refactoring (this check used to be "res == 0").
Thanks to syke for spotting it.

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-03-24 17:56:01 UTC (rev 39672)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-03-24 19:56:41 UTC (rev 39673)
@@ -256,7 +256,7 @@
 	bool res = vocab_tokenize_string(words, string, s->_parserWords, s->_parserSuffixes, &error);
 	s->parser_valid = 0; /* not valid */
 
-	if (res && !words.empty()) {
+	if (!res && !words.empty()) {
 
 		int syntax_fail = 0;
 


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 39683
          http://scummvm.svn.sourceforge.net/scummvm/?rev=39683&view=rev
Author:   joostp
Date:     2009-03-25 10:04:45 +0000 (Wed, 25 Mar 2009)

Log Message:
-----------
revert yesterday's change -- apparently this different behavior was intentional.

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-03-25 08:06:17 UTC (rev 39682)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-03-25 10:04:45 UTC (rev 39683)
@@ -256,7 +256,7 @@
 	bool res = vocab_tokenize_string(words, string, s->_parserWords, s->_parserSuffixes, &error);
 	s->parser_valid = 0; /* not valid */
 
-	if (!res && !words.empty()) {
+	if (res && !words.empty()) {
 
 		int syntax_fail = 0;
 


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 39876
          http://scummvm.svn.sourceforge.net/scummvm/?rev=39876&view=rev
Author:   thebluegr
Date:     2009-04-06 17:13:07 +0000 (Mon, 06 Apr 2009)

Log Message:
-----------
Fixed an issue with the password screen in LSL5 and kStrAt(). The game password, saved in file memory.drv (which can either exist in the game directory, or is created in the saves folder by ScummVM) is now working correctly, and the game can be started. Note that it's NOT currently possible not to set a password (you'll get the password screen anyway the next time, and it won't work).

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-04-06 17:09:02 UTC (rev 39875)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-04-06 17:13:07 UTC (rev 39876)
@@ -420,12 +420,21 @@
 		return NULL_REG;
 	}
 
+	bool lsl5PasswordWorkaround = false;
+	// LSL5 stores the password at the beginning in memory.drv, using XOR encryption,
+	// which means that is_print_str() will fail. Therefore, do not use the heuristic to determine
+	// if we're handling a string or an array for LSL5's password screen (room 155)
+	// FIXME: implement function to get current room number
+	if (s->_gameName.equalsIgnoreCase("lsl5") && (KP_UINT(s->script_000->locals_block->locals[13]) == 155))
+		lsl5PasswordWorkaround = true;
+
+	const char* dst = (const char *)dest; // used just for code beautification purposes
+
 	if ((argc == 2) &&
 	        /* Our pathfinder already works around the issue we're trying to fix */
-	        (strcmp(s->seg_manager->getDescription(argv[0]), AVOIDPATH_DYNMEM_STRING) != 0)  &&
-	        ((strlen((const char*)dest) < 2) || (!is_print_str((const char*)dest))))
-		/* SQ4 array handling detected */
-	{
+	        (strcmp(s->seg_manager->getDescription(argv[0]), AVOIDPATH_DYNMEM_STRING) != 0) &&
+	        ((strlen(dst) < 2) || (!lsl5PasswordWorkaround && !is_print_str(dst)))) {
+		// SQ4 array handling detected
 #ifndef SCUMM_BIG_ENDIAN
 		int odd = KP_UINT(argv[1]) & 1;
 #else


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 40127
          http://scummvm.svn.sourceforge.net/scummvm/?rev=40127&view=rev
Author:   thebluegr
Date:     2009-04-24 20:09:52 +0000 (Fri, 24 Apr 2009)

Log Message:
-----------
Long messages in Eco Quest 1 are now shown correctly

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-04-24 19:00:55 UTC (rev 40126)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-04-24 20:09:52 UTC (rev 40127)
@@ -797,7 +797,7 @@
 
 	if (state.loadRes(module) && state.getSpecific(&tuple)) {
 		if (buffer)
-			state.getText(buffer, 100);
+			state.getText(buffer, 255);
 
 		return make_reg(0, *buffer);
 	} else {


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 40394
          http://scummvm.svn.sourceforge.net/scummvm/?rev=40394&view=rev
Author:   waltervn
Date:     2009-05-08 22:44:26 +0000 (Fri, 08 May 2009)

Log Message:
-----------
SCI: Fix return value of GetMessage.

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-05-08 21:25:56 UTC (rev 40393)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-05-08 22:44:26 UTC (rev 40394)
@@ -799,7 +799,7 @@
 		if (buffer)
 			state.getText(buffer, 255);
 
-		return make_reg(0, *buffer);
+		return argv[3];
 	} else {
 		return NULL_REG;
 	}


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 40556
          http://scummvm.svn.sourceforge.net/scummvm/?rev=40556&view=rev
Author:   waltervn
Date:     2009-05-14 11:27:54 +0000 (Thu, 14 May 2009)

Log Message:
-----------
SCI: Message: fix K_MESSAGE_SIZE

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-05-14 11:15:36 UTC (rev 40555)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-05-14 11:27:54 UTC (rev 40556)
@@ -817,7 +817,7 @@
 		MessageState tempState;
 
 		if (tempState.loadRes(s->resmgr, UKPV(1), false) && tempState.findTuple(tuple) && tempState.getMessage())
-			return make_reg(0, strlen(tempState.getText()));
+			return make_reg(0, strlen(tempState.getText()) + 1);
 		else
 			return NULL_REG;
 	}


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 44383
          http://scummvm.svn.sourceforge.net/scummvm/?rev=44383&view=rev
Author:   wjpalenstijn
Date:     2009-09-26 22:55:24 +0000 (Sat, 26 Sep 2009)

Log Message:
-----------
SCI: Fix regression in r44081 breaking walking in QfG2

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-09-26 22:50:35 UTC (rev 44382)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-09-26 22:55:24 UTC (rev 44383)
@@ -352,7 +352,7 @@
 
 
 reg_t kStrAt(EngineState *s, int, int argc, reg_t *argv) {
-	char *dest = s->segMan->derefString(argv[0]);
+	byte *dest = (byte*)s->segMan->derefString(argv[0]);
 	reg_t *dest2;
 
 	if (!dest) {
@@ -370,7 +370,8 @@
 	if ((argc == 2) &&
 	        /* Our pathfinder already works around the issue we're trying to fix */
 	        (strcmp(s->segMan->getDescription(argv[0]), AVOIDPATH_DYNMEM_STRING) != 0) &&
-	        ((strlen(dest) < 2) || (!lsl5PasswordWorkaround && !is_print_str(dest)))) {
+	        ((strlen((const char*)dest) < 2) ||
+	         (!lsl5PasswordWorkaround && !is_print_str((const char*)dest)))) {
 		// SQ4 array handling detected
 #ifndef SCUMM_BIG_ENDIAN
 		int odd = argv[1].toUint16() & 1;
@@ -378,7 +379,7 @@
 		int odd = !(argv[1].toUint16() & 1);
 #endif
 		dest2 = ((reg_t *) dest) + (argv[1].toUint16() / 2);
-		dest = ((char *)(&dest2->offset)) + odd;
+		dest = ((byte *)(&dest2->offset)) + odd;
 	} else
 		dest += argv[1].toUint16();
 


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 44394
          http://scummvm.svn.sourceforge.net/scummvm/?rev=44394&view=rev
Author:   wjpalenstijn
Date:     2009-09-27 11:14:01 +0000 (Sun, 27 Sep 2009)

Log Message:
-----------
SCI: clean up kStrAt

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-09-27 10:36:45 UTC (rev 44393)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-09-27 11:14:01 UTC (rev 44394)
@@ -305,25 +305,35 @@
 		return NULL_REG;
 	}
 
-	byte* dest;
+	byte value;
+	byte newvalue = 0;
+	unsigned int offset = argv[1].toUint16();
+	if (argc > 2)
+		newvalue = argv[2].toSint16();
 
 	if (dest_r.isRaw) {
-		dest = (byte*)dest_r.raw + argv[1].toUint16();
+		value = dest_r.raw[offset];
+		if (argc > 2) /* Request to modify this char */
+			dest_r.raw[offset] = newvalue;
 	} else {
-#ifndef SCUMM_BIG_ENDIAN
-		int odd = argv[1].toUint16() & 1;
-#else
-		int odd = !(argv[1].toUint16() & 1);
-#endif
-		reg_t *tmp = dest_r.reg + (argv[1].toUint16() / 2);
-		dest = ((byte *)(&tmp->offset)) + odd;
+		reg_t &tmp = dest_r.reg[offset / 2];
+		if (!(offset & 1)) {
+			value = tmp.offset & 0x00ff;
+			if (argc > 2) { /* Request to modify this char */
+				tmp.offset &= 0xff00;
+				tmp.offset |= newvalue;
+			}
+		} else {
+			value = tmp.offset >> 8;
+			if (argc > 2)  { /* Request to modify this char */
+				tmp.offset &= 0x00ff;
+				tmp.offset |= newvalue << 8;
+			}
+		}
 	}
 
-	s->r_acc = make_reg(0, *dest);
+	s->r_acc = make_reg(0, value);
 
-	if (argc > 2)
-		*dest = argv[2].toSint16(); /* Request to modify this char */
-
 	return s->r_acc;
 }
 


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 44417
          http://scummvm.svn.sourceforge.net/scummvm/?rev=44417&view=rev
Author:   fingolfin
Date:     2009-09-27 19:24:18 +0000 (Sun, 27 Sep 2009)

Log Message:
-----------
Simplify kFormat slightly

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-09-27 19:23:44 UTC (rev 44416)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-09-27 19:24:18 UTC (rev 44417)
@@ -364,7 +364,7 @@
 ** the supplied parameters and writes it to the targ_address.
 */
 reg_t kFormat(EngineState *s, int, int argc, reg_t *argv) {
-	int *arguments;
+	uint16 *arguments;
 	reg_t dest = argv[0];
 	char targetbuf[512];
 	char *target = targetbuf;
@@ -392,9 +392,9 @@
 	debugC(2, kDebugLevelStrings, "Formatting \"%s\"\n", source);
 
 
-	arguments = (int*)malloc(sizeof(int) * argc);
+	arguments = (uint16 *)malloc(sizeof(uint16) * argc);
 #ifdef SATISFY_PURIFY
-	memset(arguments, 0, sizeof(int) * argc);
+	memset(arguments, 0, sizeof(uint16) * argc);
 #endif
 
 	for (i = startarg; i < argc; i++)
@@ -530,12 +530,11 @@
 				if (xfer == 'x')
 					format_string = "%x";
 
+				int val = arguments[paramindex];
 				if (!unsigned_var)
-					if (arguments[paramindex] & 0x8000)
-						/* sign extend */
-						arguments[paramindex] = (~0xffff) | arguments[paramindex];
+					val = (int16)arguments[paramindex];
 
-				target += sprintf(target, format_string, arguments[paramindex++]);
+				target += sprintf(target, format_string, val);
 				CHECK_OVERFLOW1(target, 0, NULL_REG);
 
 				unsigned_var = 0;


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 44484
          http://scummvm.svn.sourceforge.net/scummvm/?rev=44484&view=rev
Author:   thebluegr
Date:     2009-09-30 13:09:18 +0000 (Wed, 30 Sep 2009)

Log Message:
-----------
Fixed a crash when starting KQ1SCI

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-09-30 13:07:12 UTC (rev 44483)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-09-30 13:09:18 UTC (rev 44484)
@@ -447,7 +447,7 @@
 			switch (xfer) {
 			case 's': { /* Copy string */
 				reg_t reg = argv[startarg + paramindex];
-				Common::String tempsource = kernel_lookup_text(s, reg,
+				Common::String tempsource = (reg == NULL_REG) ? "" : kernel_lookup_text(s, reg,
 				                                  arguments[paramindex + 1]);
 				int slen = strlen(tempsource.c_str());
 				int extralen = str_leng - slen;


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 44490
          http://scummvm.svn.sourceforge.net/scummvm/?rev=44490&view=rev
Author:   thebluegr
Date:     2009-09-30 15:21:44 +0000 (Wed, 30 Sep 2009)

Log Message:
-----------
Fixed regression in r44417, found by waltervn. The score in KQ1SCI should be diplayed correctly again 

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-09-30 14:54:18 UTC (rev 44489)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-09-30 15:21:44 UTC (rev 44490)
@@ -529,6 +529,7 @@
 					val = (int16)arguments[paramindex];
 
 				target += sprintf(target, format_string, val);
+				paramindex++;
 				CHECK_OVERFLOW1(target, 0, NULL_REG);
 
 				unsigned_var = 0;


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 44597
          http://scummvm.svn.sourceforge.net/scummvm/?rev=44597&view=rev
Author:   m_kiewitz
Date:     2009-10-04 11:05:38 +0000 (Sun, 04 Oct 2009)

Log Message:
-----------
SCI: kReadNumber crash fix for lsl1

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-10-04 10:48:49 UTC (rev 44596)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-10-04 11:05:38 UTC (rev 44597)
@@ -335,7 +335,7 @@
 	Common::String source_str = s->segMan->getString(argv[0]);
 	const char *source = source_str.c_str();
 
-	while (isspace(*source))
+	while (isspace((unsigned char)*source))
 		source++; /* Skip whitespace */
 
 	if (*source == '$') /* SCI uses this for hex numbers */


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 44738
          http://scummvm.svn.sourceforge.net/scummvm/?rev=44738&view=rev
Author:   thebluegr
Date:     2009-10-07 16:47:06 +0000 (Wed, 07 Oct 2009)

Log Message:
-----------
Fixed a potential buffer overflow in kFormat, which occurs in the "About" screen in KQ4

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-10-07 16:26:16 UTC (rev 44737)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-10-07 16:47:06 UTC (rev 44738)
@@ -359,7 +359,8 @@
 reg_t kFormat(EngineState *s, int argc, reg_t *argv) {
 	uint16 *arguments;
 	reg_t dest = argv[0];
-	char targetbuf[512];
+	int maxsize = 4096; /* Arbitrary... */
+	char targetbuf[4096];
 	char *target = targetbuf;
 	reg_t position = argv[1]; /* source */
 	int index = argv[2].toUint16();
@@ -371,9 +372,7 @@
 	int startarg;
 	int str_leng = 0; /* Used for stuff like "%13s" */
 	int unsigned_var = 0;
-	int maxsize = 4096; /* Arbitrary... */
 
-
 	if (position.segment)
 		startarg = 2;
 	else


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 45113
          http://scummvm.svn.sourceforge.net/scummvm/?rev=45113&view=rev
Author:   waltervn
Date:     2009-10-15 11:57:36 +0000 (Thu, 15 Oct 2009)

Log Message:
-----------
SCI: Add support for reg_t based segments to Message()

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-10-15 11:47:05 UTC (rev 45112)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-10-15 11:57:36 UTC (rev 45113)
@@ -670,20 +670,35 @@
 
 		s->_msgState->lastQuery(module, msg);
 
-		byte *buffer = s->_segMan->derefBulkPtr(argv[1], 10);
+		bool ok = false;
 
-		if (buffer) {
-			assert(s->_segMan->dereference(argv[1]).isRaw);
+		if (s->_segMan->dereference(argv[1]).isRaw) {
+			byte *buffer = s->_segMan->derefBulkPtr(argv[1], 10);
 
-			WRITE_LE_UINT16(buffer, module);
-			WRITE_LE_UINT16(buffer + 2, msg.noun);
-			WRITE_LE_UINT16(buffer + 4, msg.verb);
-			WRITE_LE_UINT16(buffer + 6, msg.cond);
-			WRITE_LE_UINT16(buffer + 8, msg.seq);
+			if (buffer) {
+				ok = true;
+				WRITE_LE_UINT16(buffer, module);
+				WRITE_LE_UINT16(buffer + 2, msg.noun);
+				WRITE_LE_UINT16(buffer + 4, msg.verb);
+				WRITE_LE_UINT16(buffer + 6, msg.cond);
+				WRITE_LE_UINT16(buffer + 8, msg.seq);
+			}
 		} else {
-			warning("Message: buffer %04x:%04x invalid or too small to hold the tuple", PRINT_REG(argv[1]));
+			reg_t *buffer = s->_segMan->derefRegPtr(argv[1], 5);
+
+			if (buffer) {
+				ok = true;
+				buffer[0] = make_reg(0, module);
+				buffer[1] = make_reg(0, msg.noun);
+				buffer[2] = make_reg(0, msg.verb);
+				buffer[3] = make_reg(0, msg.cond);
+				buffer[4] = make_reg(0, msg.seq);
+			}
 		}
 
+		if (!ok)
+			warning("Message: buffer %04x:%04x invalid or too small to hold the tuple", PRINT_REG(argv[1]));
+
 		return NULL_REG;
 	}
 	default:


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 45203
          http://scummvm.svn.sourceforge.net/scummvm/?rev=45203&view=rev
Author:   m_kiewitz
Date:     2009-10-17 21:22:27 +0000 (Sat, 17 Oct 2009)

Log Message:
-----------
SCI: kStrAt added fixme comment

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-10-17 21:11:56 UTC (rev 45202)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-10-17 21:22:27 UTC (rev 45203)
@@ -305,6 +305,8 @@
 		newvalue = argv[2].toSint16();
 
 	if (dest_r.isRaw) {
+		// FIXME: in kq5 this here gets called with offset = 0xFFFF, we should implement maxSize check in here
+		//  i dont know the exact behaviour, so i dont know how to do this correctly
 		value = dest_r.raw[offset];
 		if (argc > 2) /* Request to modify this char */
 			dest_r.raw[offset] = newvalue;


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 45213
          http://scummvm.svn.sourceforge.net/scummvm/?rev=45213&view=rev
Author:   thebluegr
Date:     2009-10-18 10:39:10 +0000 (Sun, 18 Oct 2009)

Log Message:
-----------
Added a sanity check in kStrAt()

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-10-18 10:32:08 UTC (rev 45212)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-10-18 10:39:10 UTC (rev 45213)
@@ -292,6 +292,11 @@
 
 
 reg_t kStrAt(EngineState *s, int argc, reg_t *argv) {
+	if (argv[0] == SIGNAL_REG) {
+		warning("Attempt to perform kStrAt() on a signal reg");
+		return NULL_REG;
+	}
+
 	SegmentRef dest_r = s->_segMan->dereference(argv[0]);
 	if (!dest_r.raw) {
 		warning("Attempt to StrAt at invalid pointer %04x:%04x", PRINT_REG(argv[0]));


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 45530
          http://scummvm.svn.sourceforge.net/scummvm/?rev=45530&view=rev
Author:   m_kiewitz
Date:     2009-10-30 10:06:35 +0000 (Fri, 30 Oct 2009)

Log Message:
-----------
SCI: kStrAt / dont do call, when offset exceeds maxSize. fixes crash in kq5

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-10-30 09:51:00 UTC (rev 45529)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-10-30 10:06:35 UTC (rev 45530)
@@ -309,8 +309,11 @@
 		newvalue = argv[2].toSint16();
 
 	if (dest_r.isRaw) {
-		// FIXME: in kq5 this here gets called with offset = 0xFFFF, we should implement maxSize check in here
-		//  i dont know the exact behaviour, so i dont know how to do this correctly
+		// in kq5 this here gets called with offset 0xFFFF
+		if ((int)offset > dest_r.maxSize) {
+			warning("kStrAt offset %X exceeds maxSize", offset);
+			return s->r_acc;
+		}
 		value = dest_r.raw[offset];
 		if (argc > 2) /* Request to modify this char */
 			dest_r.raw[offset] = newvalue;


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 45537
          http://scummvm.svn.sourceforge.net/scummvm/?rev=45537&view=rev
Author:   m_kiewitz
Date:     2009-10-30 14:22:22 +0000 (Fri, 30 Oct 2009)

Log Message:
-----------
SCI: kStrAt offset checking fixed

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-10-30 12:33:05 UTC (rev 45536)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-10-30 14:22:22 UTC (rev 45537)
@@ -308,12 +308,14 @@
 	if (argc > 2)
 		newvalue = argv[2].toSint16();
 
+	// in kq5 this here gets called with offset 0xFFFF
+	//  (in the desert wheng getting the staff)
+	if ((int)offset >= dest_r.maxSize) {
+		warning("kStrAt offset %X exceeds maxSize", offset);
+		return s->r_acc;
+	}
+
 	if (dest_r.isRaw) {
-		// in kq5 this here gets called with offset 0xFFFF
-		if ((int)offset > dest_r.maxSize) {
-			warning("kStrAt offset %X exceeds maxSize", offset);
-			return s->r_acc;
-		}
 		value = dest_r.raw[offset];
 		if (argc > 2) /* Request to modify this char */
 			dest_r.raw[offset] = newvalue;


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 45652
          http://scummvm.svn.sourceforge.net/scummvm/?rev=45652&view=rev
Author:   thebluegr
Date:     2009-11-04 09:31:35 +0000 (Wed, 04 Nov 2009)

Log Message:
-----------
Cleanup

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2009-11-04 08:38:32 UTC (rev 45651)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2009-11-04 09:31:35 UTC (rev 45652)
@@ -32,12 +32,6 @@
 
 namespace Sci {
 
-#define CHECK_OVERFLOW1(pt, size, rv) \
-	if (((pt) - (str_base)) + (size) > maxsize) { \
-		error("String expansion exceeded heap boundaries"); \
-		return rv;\
-	}
-
 /* Returns the string the script intended to address */
 Common::String kernel_lookup_text(EngineState *s, reg_t address, int index) {
 	char *seeker;
@@ -361,6 +355,12 @@
 #define ALIGN_LEFT -1
 #define ALIGN_CENTRE 2
 
+#define CHECK_OVERFLOW1(pt, size, rv) \
+	if (((pt) - (targetbuf)) + (size) > maxsize) { \
+		error("String expansion exceeded heap boundaries"); \
+		return rv;\
+	}
+
 /*  Format(targ_address, textresnr, index_inside_res, ...)
 ** or
 **  Format(targ_address, heap_text_addr, ...)
@@ -375,7 +375,6 @@
 	char *target = targetbuf;
 	reg_t position = argv[1]; /* source */
 	int index = argv[2].toUint16();
-	char *str_base = target;
 	int mode = 0;
 	int paramindex = 0; /* Next parameter to evaluate */
 	char xfer;
@@ -584,6 +583,7 @@
 	return dest; /* Return target addr */
 }
 
+#undef CHECK_OVERFLOW1
 
 reg_t kStrLen(EngineState *s, int argc, reg_t *argv) {
 	return make_reg(0, s->_segMan->strlen(argv[0]));


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 47496
          http://scummvm.svn.sourceforge.net/scummvm/?rev=47496&view=rev
Author:   thebluegr
Date:     2010-01-24 11:00:21 +0000 (Sun, 24 Jan 2010)

Log Message:
-----------
Replaced CHECK_OVERFLOW1 with asserts

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-01-24 10:06:27 UTC (rev 47495)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-01-24 11:00:21 UTC (rev 47496)
@@ -179,12 +179,6 @@
 #define ALIGN_LEFT -1
 #define ALIGN_CENTRE 2
 
-#define CHECK_OVERFLOW1(pt, size, rv) \
-	if (((pt) - (targetbuf)) + (size) > maxsize) { \
-		error("String expansion exceeded heap boundaries"); \
-		return rv;\
-	}
-
 /*  Format(targ_address, textresnr, index_inside_res, ...)
 ** or
 **  Format(targ_address, heap_text_addr, ...)
@@ -229,7 +223,7 @@
 	while ((xfer = *source++)) {
 		if (xfer == '%') {
 			if (mode == 1) {
-				CHECK_OVERFLOW1(target, 2, NULL_REG);
+				assert((target - targetbuf) + 2 <= maxsize);
 				*target++ = '%'; /* Literal % by using "%%" */
 				mode = 0;
 			} else {
@@ -274,7 +268,7 @@
 			} else
 				str_leng = 0;
 
-			CHECK_OVERFLOW1(target, str_leng + 1, NULL_REG);
+			assert((target - targetbuf) + str_leng + 1 <= maxsize);
 
 			switch (xfer) {
 			case 's': { /* Copy string */
@@ -283,7 +277,7 @@
 				                                  arguments[paramindex + 1]);
 				int slen = strlen(tempsource.c_str());
 				int extralen = str_leng - slen;
-				CHECK_OVERFLOW1(target, extralen, NULL_REG);
+				assert((target - targetbuf) + extralen <= maxsize);
 				if (extralen < 0)
 					extralen = 0;
 
@@ -336,7 +330,7 @@
 			break;
 
 			case 'c': { /* insert character */
-				CHECK_OVERFLOW1(target, 2, NULL_REG);
+				assert((target - targetbuf) + 2 <= maxsize);
 				if (align >= 0)
 					while (str_leng-- > 1)
 						*target++ = ' '; /* Format into the text */
@@ -362,7 +356,7 @@
 
 				target += sprintf(target, format_string, val);
 				paramindex++;
-				CHECK_OVERFLOW1(target, 0, NULL_REG);
+				assert((target - targetbuf) <= maxsize);
 
 				unsigned_var = 0;
 
@@ -407,8 +401,6 @@
 	return dest; /* Return target addr */
 }
 
-#undef CHECK_OVERFLOW1
-
 reg_t kStrLen(EngineState *s, int argc, reg_t *argv) {
 	return make_reg(0, s->_segMan->strlen(argv[0]));
 }


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 47498
          http://scummvm.svn.sourceforge.net/scummvm/?rev=47498&view=rev
Author:   thebluegr
Date:     2010-01-24 11:57:10 +0000 (Sun, 24 Jan 2010)

Log Message:
-----------
Always memset the newly malloc'ed argument memory to 0 inside kFormat

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-01-24 11:54:55 UTC (rev 47497)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-01-24 11:57:10 UTC (rev 47498)
@@ -213,9 +213,7 @@
 
 
 	arguments = (uint16 *)malloc(sizeof(uint16) * argc);
-#ifdef SATISFY_PURIFY
 	memset(arguments, 0, sizeof(uint16) * argc);
-#endif
 
 	for (i = startarg; i < argc; i++)
 		arguments[i-startarg] = argv[i].toUint16(); /* Parameters are copied to prevent overwriting */


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 47500
          http://scummvm.svn.sourceforge.net/scummvm/?rev=47500&view=rev
Author:   waltervn
Date:     2010-01-24 15:21:51 +0000 (Sun, 24 Jan 2010)

Log Message:
-----------
SCI: Fix left aligning and centering in kFormat.

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-01-24 12:19:09 UTC (rev 47499)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-01-24 15:21:51 UTC (rev 47500)
@@ -241,16 +241,11 @@
 
 				if (xfer == '0')
 					fillchar = '0';
-				else
+				else if (xfer == '=')
+					align = ALIGN_CENTRE;
+				else if (isdigit(xfer) || (xfer == '-'))
+					source--; // Go to start of length argument
 
-					if (xfer == '=') {
-						align = ALIGN_CENTRE;
-						source++;
-					} else
-
-						if (isdigit(xfer))
-							source--; /* Stepped over length argument */
-
 				str_leng = strtol(source, &destp, 10);
 
 				if (destp > source)


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 48066
          http://scummvm.svn.sourceforge.net/scummvm/?rev=48066&view=rev
Author:   mthreepwood
Date:     2010-02-15 17:10:40 +0000 (Mon, 15 Feb 2010)

Log Message:
-----------
SCI32 reorders the REF* subops in kMessage; fixes GK1 conversations.

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-02-15 10:01:26 UTC (rev 48065)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-02-15 17:10:40 UTC (rev 48066)
@@ -475,10 +475,16 @@
 #ifdef ENABLE_SCI32
 	if (getSciVersion() >= SCI_VERSION_2) {
 		// In complete weirdness, SCI32 bumps up subops 3-8 to 4-9 and stubs off subop 3.
+		// In addition, SCI32 reorders the REF* subops.
 		if (func == 3)
-			warning("SCI32 kMessage(3)");
-		else if (func > 3)
+			error("SCI32 kMessage(3)");
+		else if (func > 3) {
 			func--;
+			if (func == K_MESSAGE_REFCOND)
+				func = K_MESSAGE_REFNOUN;
+			else if (func == K_MESSAGE_REFNOUN || func == K_MESSAGE_REFVERB)
+				func--;
+		}
 	}
 #endif
 


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 49250
          http://scummvm.svn.sourceforge.net/scummvm/?rev=49250&view=rev
Author:   m_kiewitz
Date:     2010-05-26 20:48:08 +0000 (Wed, 26 May 2010)

Log Message:
-----------
SCI: fixing kReadNumber to behave like in sierra sci (non standard atoi implementation) - fixes big door not unlocking in sq4

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-05-26 19:54:50 UTC (rev 49249)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-05-26 20:48:08 UTC (rev 49250)
@@ -138,10 +138,28 @@
 	while (isspace((unsigned char)*source))
 		source++; /* Skip whitespace */
 
-	if (*source == '$') /* SCI uses this for hex numbers */
-		return make_reg(0, (int16)strtol(source + 1, NULL, 16)); /* Hex */
-	else
-		return make_reg(0, (int16)strtol(source, NULL, 10)); /* Force decimal */
+	int16 result = 0;
+
+	if (*source == '$') {
+		// hexadecimal input
+		result = (int16)strtol(source + 1, NULL, 16);
+	} else {
+		// decimal input, we can not use strtol/atoi in here, because sierra used atoi BUT it was a non standard compliant
+		//  atoi, that didnt do clipping. In SQ4 we get the door code in here and that's even larger than uint32!
+		if (*source == '-') {
+			result = -1;
+			source++;
+		}
+		while (*source) {
+			result *= 10;
+			if ((*source < '0') || (*source > '9'))
+				error("Invalid character in kReadNumber input");
+			result += *source - 0x30;
+			source++;
+		}
+	}
+
+	return make_reg(0, result);
 }
 
 


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 49344
          http://scummvm.svn.sourceforge.net/scummvm/?rev=49344&view=rev
Author:   m_kiewitz
Date:     2010-05-31 09:21:15 +0000 (Mon, 31 May 2010)

Log Message:
-----------
SCI: when getting invalid chars in kReadNumber don't error() out, but create a warning - we get invalid chars in the intro of lsl5, cause currently unknown

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-05-31 08:11:49 UTC (rev 49343)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-05-31 09:21:15 UTC (rev 49344)
@@ -151,9 +151,13 @@
 			source++;
 		}
 		while (*source) {
+			if ((*source < '0') || (*source > '9')) {
+				// TODO: this happens in lsl5 right in the intro -> we get '1' '3' 0xCD 0xCD 0xCD 0xCD 0xCD
+				//       find out why this happens and fix it
+				warning("Invalid character in kReadNumber input");
+				break;
+			}
 			result *= 10;
-			if ((*source < '0') || (*source > '9'))
-				error("Invalid character in kReadNumber input");
 			result += *source - 0x30;
 			source++;
 		}


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 49345
          http://scummvm.svn.sourceforge.net/scummvm/?rev=49345&view=rev
Author:   m_kiewitz
Date:     2010-05-31 09:31:05 +0000 (Mon, 31 May 2010)

Log Message:
-----------
SCI: skip spaces in kReadNumber - this happens in lsl3 intro

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-05-31 09:21:15 UTC (rev 49344)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-05-31 09:31:05 UTC (rev 49345)
@@ -151,6 +151,10 @@
 			source++;
 		}
 		while (*source) {
+			if (*source == ' ') {
+				source++; // skip spaces - happens in lsl3 intro
+				continue;
+			}
 			if ((*source < '0') || (*source > '9')) {
 				// TODO: this happens in lsl5 right in the intro -> we get '1' '3' 0xCD 0xCD 0xCD 0xCD 0xCD
 				//       find out why this happens and fix it


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 49348
          http://scummvm.svn.sourceforge.net/scummvm/?rev=49348&view=rev
Author:   m_kiewitz
Date:     2010-05-31 12:11:42 +0000 (Mon, 31 May 2010)

Log Message:
-----------
SCI: added TODO at kReadNumber space code

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-05-31 12:10:30 UTC (rev 49347)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-05-31 12:11:42 UTC (rev 49348)
@@ -153,6 +153,8 @@
 		while (*source) {
 			if (*source == ' ') {
 				source++; // skip spaces - happens in lsl3 intro
+				// TODO: find the cause for those spaces. ssci breaks when encountering spaces, but we even get a leading
+				//  space in lsl3 (" -97 ") which would actually mean we need to return 0
 				continue;
 			}
 			if ((*source < '0') || (*source > '9')) {


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 49355
          http://scummvm.svn.sourceforge.net/scummvm/?rev=49355&view=rev
Author:   m_kiewitz
Date:     2010-05-31 13:52:07 +0000 (Mon, 31 May 2010)

Log Message:
-----------
SCI: break on anything, warn on anything but spaces in kReadNumber

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-05-31 13:09:24 UTC (rev 49354)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-05-31 13:52:07 UTC (rev 49355)
@@ -151,16 +151,14 @@
 			source++;
 		}
 		while (*source) {
-			if (*source == ' ') {
-				source++; // skip spaces - happens in lsl3 intro
-				// TODO: find the cause for those spaces. ssci breaks when encountering spaces, but we even get a leading
-				//  space in lsl3 (" -97 ") which would actually mean we need to return 0
-				continue;
-			}
 			if ((*source < '0') || (*source > '9')) {
-				// TODO: this happens in lsl5 right in the intro -> we get '1' '3' 0xCD 0xCD 0xCD 0xCD 0xCD
-				//       find out why this happens and fix it
-				warning("Invalid character in kReadNumber input");
+				// Sierras atoi stopped processing at anything different than number
+				//  Sometimes the input has a trailing space, that's fine (example: lsl3)
+				if (*source != ' ') {
+					// TODO: this happens in lsl5 right in the intro -> we get '1' '3' 0xCD 0xCD 0xCD 0xCD 0xCD
+					//       find out why this happens and fix it
+					warning("Invalid character in kReadNumber input");
+				}
 				break;
 			}
 			result *= 10;


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 50090
          http://scummvm.svn.sourceforge.net/scummvm/?rev=50090&view=rev
Author:   m_kiewitz
Date:     2010-06-20 18:22:49 +0000 (Sun, 20 Jun 2010)

Log Message:
-----------
SCI: dont display string set by kSetQuitStr

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-06-20 18:20:05 UTC (rev 50089)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-06-20 18:22:49 UTC (rev 50090)
@@ -565,7 +565,7 @@
 
 reg_t kSetQuitStr(EngineState *s, int argc, reg_t *argv) {
 	Common::String quitStr = s->_segMan->getString(argv[0]);
-	debug("Setting quit string to '%s'", quitStr.c_str());
+	//debug("Setting quit string to '%s'", quitStr.c_str());
 	return s->r_acc;
 }
 


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 50234
          http://scummvm.svn.sourceforge.net/scummvm/?rev=50234&view=rev
Author:   wjpalenstijn
Date:     2010-06-24 21:09:38 +0000 (Thu, 24 Jun 2010)

Log Message:
-----------
SCI: Make kStrAt clear segment when writing characters

This is necessary since the uninitialized value detection from r50211,
and is analogous to seg_manager.cpp's setChar.
(Triggered in LSL3 age verification.)

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-06-24 20:52:35 UTC (rev 50233)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-06-24 21:09:38 UTC (rev 50234)
@@ -115,12 +115,14 @@
 			if (argc > 2) { /* Request to modify this char */
 				tmp.offset &= 0xff00;
 				tmp.offset |= newvalue;
+				tmp.segment = 0;
 			}
 		} else {
 			value = tmp.offset >> 8;
 			if (argc > 2)  { /* Request to modify this char */
 				tmp.offset &= 0x00ff;
 				tmp.offset |= newvalue << 8;
+				tmp.segment = 0;
 			}
 		}
 	}


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 51376
          http://scummvm.svn.sourceforge.net/scummvm/?rev=51376&view=rev
Author:   mthreepwood
Date:     2010-07-27 15:45:21 +0000 (Tue, 27 Jul 2010)

Log Message:
-----------
SCI: Fix bug #3035058 - ECOQUEST demo: Missing subtitles

The demo uses a special version of kMessage to get its messages.

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-07-27 15:29:58 UTC (rev 51375)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-07-27 15:45:21 UTC (rev 51376)
@@ -456,7 +456,8 @@
 	K_MESSAGE_REFNOUN,
 	K_MESSAGE_PUSH,
 	K_MESSAGE_POP,
-	K_MESSAGE_LASTMESSAGE
+	K_MESSAGE_LASTMESSAGE,
+	K_MESSAGE_ECOQUEST1_DEMO = 99
 };
 
 reg_t kGetMessage(EngineState *s, int argc, reg_t *argv) {
@@ -558,6 +559,18 @@
 
 		return NULL_REG;
 	}
+	case K_MESSAGE_ECOQUEST1_DEMO:
+		// The EcoQuest 1 demo uses a special version of kMessage. It's one of the
+		// earliest SCI 1.1 games. The noun is always 99 in this case, so we can
+		// treat it as a subop. If any other games require this syntax, we can change
+		// this to work with those games too.
+
+		if (g_sci->getGameId() != GID_ECOQUEST || !g_sci->isDemo())
+			error("kMessage called with EcoQuest 1 demo syntax in a different game");
+
+		tuple.noun = argv[0].toUint16();
+		tuple.verb = argv[2].toUint16();
+		return make_reg(0, s->_msgState->getMessage(argv[1].toUint16(), tuple, argv[3]));
 	default:
 		warning("Message: subfunction %i invoked (not implemented)", func);
 	}


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 52654
          http://scummvm.svn.sourceforge.net/scummvm/?rev=52654&view=rev
Author:   lskovlun
Date:     2010-09-09 12:44:34 +0000 (Thu, 09 Sep 2010)

Log Message:
-----------
SCI: Support push and pop subfunctions in kMessage()
Glossary seems to work now.

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-09-09 12:01:04 UTC (rev 52653)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-09-09 12:44:34 UTC (rev 52654)
@@ -486,10 +486,12 @@
 	}
 #endif
 
-	if ((func != K_MESSAGE_NEXT) && (argc < 2)) {
-		warning("Message: not enough arguments passed to subfunction %d", func);
-		return NULL_REG;
-	}
+//	TODO: Perhaps fix this check, currently doesn't work with PUSH and POP subfunctions
+//	Pepper uses them to to handle the glossary
+//	if ((func != K_MESSAGE_NEXT) && (argc < 2)) {
+//		warning("Message: not enough arguments passed to subfunction %d", func);
+//		return NULL_REG;
+//	}
 
 	MessageTuple tuple;
 
@@ -558,6 +560,12 @@
 
 		return NULL_REG;
 	}
+	case K_MESSAGE_PUSH:
+		s->_msgState->pushCursorStack();
+		break;
+	case K_MESSAGE_POP:
+		s->_msgState->popCursorStack();
+		break;
 	default:
 		warning("Message: subfunction %i invoked (not implemented)", func);
 	}


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 52676
          http://scummvm.svn.sourceforge.net/scummvm/?rev=52676&view=rev
Author:   lskovlun
Date:     2010-09-11 13:30:42 +0000 (Sat, 11 Sep 2010)

Log Message:
-----------
SCI: In kFormat, handle arguments pointing to text resource 0 correctly.
Fixes Colonel's Bequest inventory formatting.

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-09-11 10:03:31 UTC (rev 52675)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-09-11 13:30:42 UTC (rev 52676)
@@ -275,7 +275,7 @@
 					reg = readSelector(s->_segMan, reg, SELECTOR(data));
 #endif
 
-				Common::String tempsource = (reg == NULL_REG) ? "" : g_sci->getKernel()->lookupText(reg,
+				Common::String tempsource = g_sci->getKernel()->lookupText(reg,
 				                                  arguments[paramindex + 1]);
 				int slen = strlen(tempsource.c_str());
 				int extralen = str_leng - slen;


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 54374
          http://scummvm.svn.sourceforge.net/scummvm/?rev=54374&view=rev
Author:   lskovlun
Date:     2010-11-19 12:52:17 +0000 (Fri, 19 Nov 2010)

Log Message:
-----------
Some subfunctions changed positions in SCI3 kString. WIP.
And with this, LSL7 starts up. Yeah, baby!

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-11-19 12:49:46 UTC (rev 54373)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-11-19 12:52:17 UTC (rev 54374)
@@ -616,6 +616,15 @@
 }
 
 reg_t kString(EngineState *s, int argc, reg_t *argv) {
+
+	// TODO: Find out how exactly subfunctions work in SCI3
+	if (getSciVersion() == SCI_VERSION_3 &&
+	    argv[0].toUint16() == 8)
+		argv[0].offset = 10;
+	if (getSciVersion() == SCI_VERSION_3 &&
+	    argv[0].toUint16() == 11)
+		argv[0].offset = 13;
+
 	switch (argv[0].toUint16()) {
 	case 0: { // New
 		reg_t stringHandle;
@@ -780,9 +789,13 @@
 		Common::String string = s->_segMan->getString(argv[1]);
 		return make_reg(0, (uint16)atoi(string.c_str()));
 	}
-	case 14: // Unknown (SCI3)
-		warning("Unknown kString subop %d", argv[0].toUint16());
+	case 14: { // lower (SCI3)
+		Common::String string = s->_segMan->getString(argv[1]);
+		
+		string.toLowercase();
+		s->_segMan->strcpy(argv[1], string.c_str());
 		return NULL_REG;
+	}
 	default:
 		error("Unknown kString subop %d", argv[0].toUint16());
 	}


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.

Revision: 54403
          http://scummvm.svn.sourceforge.net/scummvm/?rev=54403&view=rev
Author:   thebluegr
Date:     2010-11-21 02:07:55 +0000 (Sun, 21 Nov 2010)

Log Message:
-----------
SCI: Removed the correct subops from kString in late SCI2.1/SCI3 games (thanks to waltervn)

Modified Paths:
--------------
    scummvm/trunk/engines/sci/engine/kstring.cpp

Modified: scummvm/trunk/engines/sci/engine/kstring.cpp
===================================================================
--- scummvm/trunk/engines/sci/engine/kstring.cpp	2010-11-21 02:06:20 UTC (rev 54402)
+++ scummvm/trunk/engines/sci/engine/kstring.cpp	2010-11-21 02:07:55 UTC (rev 54403)
@@ -620,7 +620,7 @@
 	uint16 op = argv[0].toUint16();
 
 	if (g_sci->_features->detectSci2StringFunctionType() == kSci2StringFunctionNew) {
-		if (op >= 7)	// Cpy, Cmp have been removed
+		if (op >= 8)	// Dup, GetData have been removed
 			op += 2;
 	}
 


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.