This file contains the text of the FlexeLint/PC-lint error messages. It is to be used in conjunction with licensed FlexeLint/PC-lint software. Copyright (C) 1985-2008 Gimpel Software, All rights reserved Gimpel Software * 3207 Hogarth Lane * Collegeville, PA 19426 Phone (610) 584-4261 * FAX (610) 584-4266 * www.gimpel.com 19. MESSAGES 19.1 Syntax Errors 19.2 Internal Errors 19.3 Fatal Errors 19.4 Warning Messages 19.5 Informational Messages 19.6 Elective Notes 19.7 C++ Syntax Errors 19.8 Additional Internal Errors 19.9 C++ Warning Messages 19.10 C++ Informational Messages 19.11 C++ Elective Notes Most error messages have an associated error number. By looking up the number in the list below you can obtain additional information about the cause of the error. This information is also available from a self-extracting executable msg.exe, found on the PC-lint distribution diskette, which produces the machine-readable ASCII file msg.txt. For FlexeLint users, the file msg.txt can be found on the FlexeLint Supplementary diskette. Messages numbered 1000 and higher pertain generally to C++. This is summarized in the table below. After a possible 1000 is subtracted off, the remainder lies in the range 0-999. Remainders in the range 1-199 are syntax errors, 200-299 are PC-lint/FlexeLint internal errors and should never occur, 300-399 are fatal errors usually brought about by exceeding some limit, 400-699 are warning messages that indicate that something is likely to be wrong with the program being examined. Remainders in the range 700-899 designate informational messages. These may be errors but they also may represent legitimate programming practices depending upon personal programming style. Remainders in the range 900-999 are called "Elective Notes". They are not automatically output. You may examine the list to see if you wish to be alerted to any of them. C C++ Warning Level Syntax Errors 1 - 199 1001 - 1199 1 Internal Errors 200 - 299 1200 - 1299 0 Fatal Errors 300 - 399 0 Warnings 400 - 699 1400 - 1699 2 Informational 700 - 899 1700 - 1899 3 Elective Notes 900 - 999 1900 - 1999 4 Glossary A few of the terms used in the commentary below are: argument The actual argument of a function as opposed to a dummy (or formal) parameter of a function (see parameter). arithmetic Any of the integral types (see below) plus float, double, and long double. Boolean In general, the word Boolean refers to quantities that can be either true or false. An expression is said to be Boolean (perhaps it would be better to say 'definitely Boolean') if it is of the form: operand op operand where op is a relational (> >= < <=), an equality operator (== !=), logical And (&&) or logical Or (||). A context is said to require a Boolean if it is used in an if or while clause or if it is the 2nd expression of a for clause or if it is an argument to one of the operators: && or ||. An expression needn't be definitely Boolean to be acceptable in a context that requires a Boolean. Any integer or pointer is acceptable. declaration Gives properties about an object or function (as opposed to a definition). definition That which allocates space for an object or function (as opposed to a declaration) and which may also indicate properties about the object. There should be only one definition for an object but there may be many declarations. integral A type that has properties similar to integers. These include char, short, int, and long and the unsigned variations of any of these. scalar Any of the arithmetic types plus pointers. lvalue Is an expression that can be used on the Left hand side of an assignment operator (=). Some contexts require lvalues such as autoincrement (++) and autodecrement (--). macro An abbreviation defined by a #define statement. It may or may not have arguments. member Elements of a struct and of a union are called members. module That which is compiled by a compiler in a single independent compilation. It typically includes all the text of a .c (or a .cpp or .cxx, etc.) file plus any text within any #include file(s). parameter A formal parameter of a function as opposed to an actual argument (see argument). Message Parameters Some of the messages are parameterized with one or more of the following italicized names: Char Some character Context Specifies one of several contexts in which an assignment can be made. Can be one of: o assignment -- refers to an explicit assignment operator. o return -- refers to the implied assignment of a return statement. The type of the expression is converted implicitly to the type of the function. o initialization -- refers to the assignment implied by an initialization statement. o arg. no.... -- refers to the implied assignment of an argument in the presence of a prototype. The type of the expression is implicitly converted to the type within a prototype. o arg. 'this' -- refers to the implied argument of a member function call. FileName A filename. Messages containing this parameter can be suppressed with the -efile( ... ) option. Integer Some integer Invocation A function call with argument types. To suppress a message containing an Invocation you need the complete call, not just the function name. For example, the call f(1) could result in Error 1024 parameterized by 'f(int)'. To suppress this message you could use -esym(1024,f(int)). You could also use wild cards as in -esym(1024,f*). Kind A list of control structures. Location A line number followed optionally by a filename (if different from the current) and/or a module name if different from the current. Name A string, usually an identifier, that can be suppressed with a -esym but not with -elibsym. String A sequence of characters identified further in the message description. Symbol The name of a user identifier referring to a C or C++ object such as variable, function, structure, etc. Messages containing this parameter can be suppressed with the -esym( ... ) option. For C++, if Symbol is the name of a function the full function signature (including parameters) is given. Error suppression with -esym does not require (nor want) the parameters. Type A type or a top type base is provided. A top type base is one of pointer, function, array, struct, union, or enum. TypeDiff Specifies the way in which one type differs from another. Because of type qualification, function prototypes, and type compounding, it may not be obvious how two types differ. Also, see the -etd option to inhibit errors based on type differences. TypeDiff can be one or more of: o basic -- The two types differ in some fundamental way such as double versus int. o count -- Two function types differ in the number of arguments. o ellipsis -- Two function types differ in that one is prototyped using an ellipsis and the other is not prototyped. See Section 15.8 Plain Vanilla Functions. o incomplete -- At least one of the types is only partially specified such as an array without a dimension or a function without a prototype. o nominal -- The types are nominally different but are otherwise the same. For example, int versus long where these are the same size or double versus long double where these are the same size. The two types are either both integral or both float or are functions that return types or have arguments that differ nominally. If long is the same size as int then unsigned long will differ from int both as nominal and as signed/unsigned. If not the same size, then the difference is precision. o origin -- The types are not actually different but have different origins. For example a struct is defined in two separate modules rather than in one header file. If for some reason you want to do this then use the option -etd(origin). o precision -- Two arithmetic types differ in their precision such as int vs. long where these are different sizes. o promotion -- Two function types differ in that one is prototyped with a char, short or float type and the other is not prototyped. o ptrs to... -- Pointers point to different types, some TypeDiff code follows. o ptrs to incompatible types -- Pointers point to types which in turn differ in precision, count, size, ellipsis or promotion. o qualification -- Qualifiers such as const, volatile, etc. are inconsistent. o signed/unsigned -- The types differ in that one is a signed integral type and the other is unsigned of the same size, or they are both functions that return types that differ in this way, or they are both pointers to types that differ in this way. o size -- Two arrays differ in array dimension. o strong -- two types differ in that one is strong and the other is not the same strong type. o void/nonvoid -- The two types differ in that one is void and the other is not or, more frequently, they are both functions returning types that differ in this respect or pointers to types that differ in this respect. o int/enum-- One type is an enum and the other is an int. o Type = Type -- The two types in an assignment of some kind differ in some basic way and no more information is available. o Type vs. Type -- The two types differ in some basic way and no more information is available. 19.1 C Syntax Errors 1 Unclosed Comment (Location) -- End of file was reached with an open comment still unclosed. The Location of the open comment is shown. 2 Unclosed Quote -- An end of line was reached and a matching quote character (single or double) to an earlier quote character on the same line was not found. 3 #else without a #if -- A #else was encountered not in the scope of a #if, #ifdef or #ifndef. 4 Too many #if levels -- An internal limit was reached on the level of nesting of #if statements (including #ifdef and #ifndef). 5 Too many #endif's -- A #endif was encountered not in the scope of a #if or #ifdef or #ifndef. 6 Stack Overflow -- One of the built-in non-extendable stacks has been overextended. The possibilities are too many nested #if statements, #includes statements (including all recursive #include statements), static blocks (bounded by braces) or #define replacements. 7 Unable to open include file: FileName -- FileName is the name of the include file which could not be opened. 8 Unclosed #if (Location) -- A #if (or #ifdef or #ifndef) was encountered without a corresponding #endif. Location is the location of the #if. 9 Too many #else's in #if (Location) -- A given #if contained a #else which in turn was followed by either another #else or a #elif. The error message gives the line of the #if statement that started the conditional that contained the aberration. 10 Expecting 'String' -- String is the expected token. The expected token could not be found. This is commonly given when certain reserved words are not recognized. int __interrupt f(); will receive an Expecting ';' message at the f because it thinks you just declared __interrupt. The cure is to establish a new reserved word with +rw(__interrupt). Also, make sure you are using the correct compiler options file. 11 Excessive Size -- The filename specified on a #include line had a length that exceeded FILENAME_MAX characters. 12 Need < or " -- After a #include is detected and after macro substitution is performed, a file specification of the form or "filename" is expected. 13 Bad type -- A type adjective such as long, unsigned, etc. cannot be applied to the type which follows. 14 Symbol 'Symbol' previously defined (Location) -- The named object has been defined a second time. The location of the previous definition is provided. If this is a tentative definition (no initializer) then the message can be suppressed with the +fmd flag. 15 Symbol 'Symbol' redeclared (TypeDiff) (Location) -- The named symbol has been previously declared or defined in some other module (location given) with a type different from the type given by the declaration at the current location. The parameter TypeDiff provides further information on how the types differ. 16 Unrecognized name -- A # directive is not followed by a recognizable word. If this is not an error, use the +ppw option. 17 Unrecognized name -- A non-parameter is being declared where only parameters should be. 18 Symbol 'Symbol' redeclared (TypeDiff) conflicts with Location -- A symbol is being redeclared. The parameter TypeDiff provides further information on how the types differ. Location is the location of the previous definition. 19 Useless Declaration -- A type appeared by itself without an associated variable, and the type was not a struct and not a union and not an enum. A double semi-colon can cause this as in: int x;; 20 Illegal use of = -- A function declaration was followed by an = sign. 21 Expected { -- An initializer for an indefinite size array must begin with a left brace. 22 Illegal operator -- A unary operator was found following an operand and the operator is not a post operator. 23 Expected colon -- A ? operator was encountered but this was not followed by a : as was expected. 24 Expected an expression, found 'String' -- An operator was found at the start of an expression but it was not a unary operator. 25 Illegal constant -- Too many characters were encountered in a character constant (a constant bounded by ' marks). 26 Expected an expression, found 'String' -- An expression was not found where one was expected. The unexpected token is placed in the message. 27 Illegal character (0xhh) -- An illegal character was found in the source code. The hex code is provided in the message. A blank is assumed. If you are using strange characters in identifier names you will get this message for which you may use the -ident option. 28 Redefinition of symbol 'Symbol' Location -- The identifier preceding a colon was previously declared at the Location given as not being a label. 30 Expected a constant -- A constant was expected but not obtained. This could be following a case keyword, an array dimension, bit field length, enumeration value, #if expression, etc. 31 Redefinition of symbol 'Symbol' conflicts with Location -- A data object or function previously defined in this module is being redefined. 32 Field size (member 'Symbol') should not be zero -- The length of a field was given as non-positive, (0 or negative). 33 Illegal constant -- A constant was badly formed as when an octal constant contains one of the digits 8 or 9. 34 Non-constant initializer -- A non-constant initializer was found for a static data item. 35 Initializer has side-effects -- An initializer with side effects was found for a static data item. 36 Redefining the storage class of symbol 'Symbol' conflicts with Location -- An object's storage class is being changed. 37 Value of enumerator 'Symbol' inconsistent (conflicts with Location) -- An enumerator was inconsistently valued. 38 Offset of symbol 'Symbol' inconsistent (Location) -- A member of a class or struct appears in a different position (offset from the start of the structure) than an earlier declaration. This could be caused by array dimensions changing from one module to another. 39 Redefinition of symbol 'Symbol' conflicts with Location -- A struct or union is being redefined. 40 Undeclared identifier 'Name' -- Within an expression, an identifier was encountered that had not previously been declared and was not followed by a left parenthesis. Name is the name of the identifier. 41 Redefinition of symbol 'Symbol' -- A parameter of either a function or a macro is being repeated. 42 Expected a statement -- A statement was expected but a token was encountered that could not possibly begin a statement. 43 Vacuous type for variable 'Symbol' -- A vacuous type was found such as the void type in a context that expected substance. 44 Need a switch -- A case or default statement occurred outside a switch. 45 Bad use of register -- A variable is declared as a register but its type is inconsistent with it being a register (such as a function). 46 Field type should be int -- Bit fields in a structure should be typed unsigned or int. If your compiler allows other kinds of objects, such as char, then simply suppress this message. 47 Bad type -- Unary minus requires an arithmetic operand. 48 Bad type -- Unary * or the left hand side of the ptr (->) operator requires a pointer operand. 49 Expected a type -- Only types are allowed within prototypes. A prototype is a function declaration with a sequence of types within parentheses. The processor is at a state where it has detected at least one type within parentheses and so is expecting more types or a closing right parenthesis. 50 Attempted to take the address of a non-lvalue -- Unary & operator requires an lvalue (a value suitable for placement on the left hand side of an assignment operator). 51 Expected integral type -- Unary ~ expects an integral type (signed or unsigned char, short, int, or long). 52 Expected an lvalue -- autodecrement (--) and autoincrement (++) operators require an lvalue (a value suitable for placement on the left hand side of an assignment operator). Remember that casts do not normally produce lvalues. Thus ++(char *)p; is illegal according to the ANSI standard. This construct is allowed by some compilers and is allowed if you use the +fpc option (Pointer Casts are lvalues). 53 Expected a scalar -- Autodecrement (--) and autoincrement (++) operators may only be applied to scalars (arithmetics and pointers) or to objects for which these operators have been defined. 54 Division by 0 -- The constant 0 was used on the right hand side of the division operator (/) or the remainder operator (%). 55 Bad type -- The context requires a scalar, function, array, or struct (unless -fsa). 56 Bad type -- Add/subtract operator requires scalar types and pointers may not be added to pointers. 57 Bad type -- Bit operators ( &, | and ^ ) require integral arguments. 58 Bad type -- Bad arguments were given to a relational operator; these always require two scalars and pointers can't be compared with integers (unless constant 0). 59 Bad type -- The amount by which an item can be shifted must be integral. 60 Bad type -- The value to be shifted must be integral. 61 Bad type -- The context requires a Boolean. Booleans must be some form of arithmetic or pointer. 62 Incompatible types (TypeDiff) for operator ':' -- The 2nd and 3rd arguments to ? : must be compatible types. 63 Expected an lvalue -- Assignment expects its first operand to be an lvalue. Please note that a cast removes the lvaluedness of an expression. 64 Type mismatch (Context) (TypeDiff) -- There was a mismatch in types across an assignment (or implied assignment, see Context). TypeDiff specifies the type difference. 65 Expected a member name -- After a dot (.) or pointer (->) operator a member name should appear. 66 Bad type -- A void type was employed where it is not permitted. If a void type is placed in a prototype then it must be the only type within a prototype. (See error number 49.) 67 Can't cast from Type to Type -- Attempt to cast a non-scalar to an integral. 68 Can't cast from Type to Type -- Attempt to cast a non-arithmetic to a float. 69 Can't cast from Type to Type -- Bad conversion involving incompatible structures or a structure and some other object. 70 Can't cast from Type to Type -- Attempt to cast to a pointer from an unusual type (non-integral). 71 Can't cast from Type to Type -- Attempt to cast to a type that does not allow conversions. 72 Bad option 'String' -- Was not able to interpret an option. The option is given in String. 73 Bad left operand -- The cursor is positioned at or just beyond either an -> or a . operator. These operators expect an expression primary on their left. Please enclose any complex expression in this position within parentheses. 74 Address of Register -- An attempt was made to apply the address (&) operator to a variable whose storage class was given as register. 75 Too late to change sizes (option 'String') -- The size option was given after all or part of a module was processed. Make sure that any option to reset sizes of objects be done at the beginning of the first module processed or on the command line before any module is processed. 76 can't open file String -- String is the name of the file. The named file could not be opened for output. The file was destined to become a PC-lint/FlexeLint object module. 77 Address of bit-field cannot be taken -- The address of a bit-field cannot be taken. The rules of C only allow for taking the address of a whole byte (a whole char). 78 Symbol 'Symbol' typedef'ed at Location used in expression -- The named symbol was defined in a typedef statement and is therefore considered a type. It was subsequently found in a context where an expression was expected. 79 Bad type for % operator -- The % operator should be used with some form of integer. 80 this use of ellipsis is not strictly ANSI -- The ellipsis should be used in a prototype only after a sequence of types not after a sequence of identifiers. Some compilers support this extension. If you want to use this feature suppress this message. 81 struct/union not permitted in equality comparison -- Two struct's or union's are being compared with one of == or !=. This is not permitted by the ANSI standard. If your compiler supports this, suppress this message. 82 return ; illegal with void function -- The ANSI standard does not allow an expression form of the return statement with a void function. If you are trying to cast to void as in return (void)f(); and your compiler allows it, suppress this message. 83 Incompatible pointer types with subtraction -- Two pointers being subtracted have indirect types which differ. You can get PC-lint/FlexeLint to ignore slight differences in the pointers by employing one or more of the -ep... options 84 sizeof object is zero or object is undefined -- A sizeof returned a 0 value. This could happen if the object were undefined or incompletely defined. Make sure a complete definition of the object is in scope when you use sizeof. 85 Array 'Symbol' has dimension 0 -- An array (named Symbol) was declared without a dimension in a context that required a non-zero dimension. 86 Structure 'Symbol' has no data elements -- A structure was declared (in a C module) that had no data members. Though legal in C++ this is not legal C. 87 Expression too complicated for #ifdef or #ifndef -- By the rules of C there should be only a single identifier following a #ifdef or a #ifndef. You may also supply a validly constructed C (or C++) comment. 88 Symbol 'Symbol' is an array of empty elements -- An array was declared (in a C module) whose elements were each of 0 length. Though legal in C++ this is not permitted C. 89 Argument or option too long ('String') -- The length of an option (shown in String) exceeds an internal limit. Please try to decompose the option into something smaller. At this writing the limit is 610 characters. 90 Option 'String' is only appropriate within a lint comment -- The indicated option is not appropriate at the command or the .lnt level. For example if -unreachable is given on the command line you will get this message. 91 Line exceeds Integer characters (use +linebuf) -- A line read from one of the input files is longer than anticipated. By default the line buffer size is 600 characters. Each time you use the +linebuf option you can double this size. The size can be doubled ad infinitum. 92 Negative array dimension or bit field length (Integer) -- A negative array dimension or bit field length is not permitted. 93 New-line is not permitted within string arguments to macros -- A macro invocation contains a string that is split across more than one line. For example: A( "Hello World" ); will trigger this message. Some compilers accept this construct and you can suppress this message with -e93 if this is your current practice. But it is more portable to place the string constant on one line. Thus A( "Hello World" ); would be better. 95 Expected a macro parameter but instead found 'Name' -- The # operator (or the non-standard extension to the # operator spelled #@) was found within a macro definition but was not immediately followed by a parameter of the macro as is required by the standards. Name identifies the token immediately to the right of the operator. 96 Unmatched left brace for String on Location -- The purpose of this message is to report the location of a left curly brace that is unmatched by a right curly brace. Such an unmatched left curly can be far removed from the point at which the unbalance was detected (often the end of the compilation unit). Providing the location of the left curly can be extremely helpful in determining the source of the imbalance. 98 Recovery Error (String) -- A recovery error is issued when an inconsistent state was found while attempting to recover from a syntactic error. The String provided in the message serves as a clue to this inconsistent state. Since the presumptive cause of the error is an earlier error, priority should be placed on resolving the original error. This "Recovery Error" is meant only to provide additional information on the state of the parser. 101 Expected an identifier -- While processing a function declarator, a parameter specifier was encountered that was not an identifier, whereas a prior parameter was specified as an identifier. This is mixing old-style function declarations with the new-style and is not permitted. For example void f(n,int m) will elicit this message. 102 Illegal parameter specification -- Within a function declarator, a parameter must be specified as either an identifier or as a type followed by a declarator. 103 Unexpected declaration -- After a prototype, only a comma, semi-colon, right parenthesis or a left brace may occur. This error could occur if you have omitted a terminating character after a declaration or if you are mixing old-style parameter declarations with new-style prototypes. 104 Conflicting types -- Two consecutive conflicting types were found such as int followed by double. Remove one of the types! 105 Conflicting modifiers -- Two consecutive conflicting modifiers were found such as far followed by near. Remove one of the modifiers! 106 Illegal constant -- A string constant was found within a preprocessor expression as in #if ABC == "abc" Such expressions should be integral expressions. 107 Label 'Symbol' (Location) not defined -- The Symbol at the given Location appeared in a goto but there was no corresponding label. 108 Invalid context -- A continue or break statement was encountered without an appropriate surrounding context such as a for, while, or do loop or, for the break statement only, a surrounding switch statement. 109 The combination 'short long' is not standard, 'long' is assumed -- Some compilers support the non-standard sequence short long. This message reports, as an error, that this sequence is being used. If you are required to use the construct then simply suppress this message. As the message indicates, that type will be presumed to be long. 110 Attempt to assign to void -- An attempt was made to assign a value to an object designated (possibly through a pointer) as void. 111 Assignment to const object -- An object declared as const was assigned a value. This could arise via indirection. For example, if p is a pointer to a const int then assigning to *p will raise this error. 113 Inconsistent enum declaration -- The sequence of members within an enum (or their values) is inconsistent with that of another enum (usually in some other module) having the same name. 114 Inconsistent structure declaration for tag 'Symbol' -- The sequence of members within a structure (or union) is inconsistent with another structure (usually in some other module) having the same name. 115 Struct/union not defined -- A reference to a structure or a union was made that required a definition and there is no definition in scope. For example, a reference to p->a where p is a pointer to a struct that had not yet been defined in the current module. 116 Inappropriate storage class -- A storage class other than register was given in a section of code that is dedicated to declaring parameters. The section is that part of a function preceding the first left brace. 117 Inappropriate storage class -- A storage class was provided outside any function that indicated either auto or register. Such storage classes are appropriate only within functions. 118 Too few arguments for prototype -- The number of arguments provided for a function was less than the number indicated by a prototype in scope. 119 Too many arguments for prototype -- The number of arguments provided for a function was greater than the number indicated by a prototype in scope. 120 Initialization without braces of dataless type 'Symbol' -- There was an attempt to initialize a nested object (e.g., an array element) without braces. Additionally, that object type possesses no data members. class A { public: void f(); }; class B { public: A a; int k; } ; A a[4] = { {}, {}, {}, {} }; // OK B b = { , 34 }; // Error 120 121 Attempting to initialize an object of undefined type 'Symbol' -- The initialization of an object was attempted where that object type has no visible definition. For example: class Undefined u = { 5 }; 122 Digit (Char) too large for radix -- The indicated character was found in a constant beginning with zero. For example, 08 is accepted by some compilers to represent 8 but it should be 010 or plain 8. 123 Macro 'Symbol' defined with arguments at Location this is just a warning -- The name of a macro defined with arguments was subsequently used without a following '('. This is legal but may be an oversight. It is not uncommon to suppress this message (with -e123), because some compilers allow, for example, the macro max() to coexist with a variable max. 124 Pointer to void not allowed -- A pointer to void was used in a context that does not permit void. This includes subtraction, addition and the relationals (> >= < <=). 125 Too many storage class specifiers -- More than one storage class specifier (static, extern, typedef, register or auto) was found. Only one is permitted. 126 Inconsistent structure definition 'Symbol' -- The named structure (or union or enum) was inconsistently defined across modules. The inconsistency was recognized while processing a lint object module. Line number information was not available with this message. Alter the structures so that the member information is consistent. 127 Illegal constant -- An empty character constant ('') was found. 128 Pointer to function not allowed -- A pointer to a function was found in an arithmetic context such as subtraction, addition, or one of the relationals (> >= < <=). 129 declaration expected, identifier 'Symbol' ignored -- In a context in which a declaration was expected an identifier was found. Moreover, the identifier was not followed by '(' or a '[' 130 Expected integral type -- The expression in a switch statement must be some variation of an int (possibly long or unsigned) or an enum. 131 syntax error in call of macro 'Symbol' at location Location -- This message is issued when a macro with arguments (function-like macro) is invoked and an incorrect number of arguments is provided. Location is the location of the start of the macro call. This can be useful because an errant macro call can extend over many lines. 132 Expected function definition -- A function declaration with identifiers between parentheses is the start of an old-style function definition (K&R style). This is normally followed by optional declarations and a left brace to signal the start of the function body. Either replace the identifier(s) with type(s) or complete the function with a function body. 133 Too many initializers for aggregate 'Symbol' -- In a brace-enclosed initializer, there are more items than there are elements of the aggregate. 134 Missing initializer -- An initializer was expected but only a comma was present. 135 comma assumed in initializer -- A comma was missing between two initializers. For example: int a[2][2] = { { 1, 2 } { 3, 4 } }; is missing a comma after the first right brace (}). 136 Illegal macro name -- The ANSI standard restricts the use of certain names as macros. defined is on the restricted list. 137 constant 'Symbol' used twice within switch -- The indicated constant was used twice as a case within a switch statement. Currently only enumerated types are checked for repeated occurrence. 138 Can't add parent 'Symbol' to strong type String; creates loop -- An attempt was made to add a strong type parent to a typedef type. The attempt is either explicit (with the -strong option) or implicit with the use of a typedef to a known strong type. This attempt would have caused a loop in the strong parent relationship. Such loops are simply not tolerated. 139 Can't take sizeof function -- There is an attempt to take the sizeof a function. 140 Type appears after modifier -- Microsoft modifiers such as far, _near, __huge, _pascal, etc. etc. modify the declarator to its immediate right. It therefore should not appear before the type. For example, you should write int pascal f(void); rather than pascal int f(void);. Note that const and volatile differ from the Microsoft modifiers. They may appear before or after the type. After reporting the error an attempt is made to process the modifiers as the programmer probably intended. 141 The following option has too many elements: 'String' -- The indicated option (given by 'String') is too big. It most likely consists of an itemized list that has too many items. You should decompose the large option into two or more smaller options that in sum are equivalent to the one large option. 143 Erroneous option: String -- An option contained information that was inconsistent with itself or with an earlier option. The String provided in the message explains more fully what the problem is. 144 Non-existent return value for symbol 'Symbol', compare with Location -- An attempt was made to use a non-existent return value of the named function (identified by Symbol). It was previously decided that the function did not return a value or was declared with void. 145 Type expected before operator, void assumed -- In a context in which a type is expected no type is found. Rather, an operator '*' or '&' was encountered. The keyword void was assumed to have preceded this operator. 146 Assuming a binary constant -- A constant of the form 0b... was encountered. This was taken to be a binary constant. For example, 0b100 represents the value 4. If your compiler supports binary constants you may suppress this message. 147 sizeof takes just one argument -- An expression of the form sizeof(a,b) was detected. A second argument is non standard and has been used by some compilers to denote an option to the sizeof operator. If your compiler has a use for the second argument then suppress this message. 148 member 'Symbol' previously declared at Location -- The indicated member was previously declared within the same structure or union. Although a redeclaration of a function may appear benign it is just not permitted by the rules of the language. One of the declarations should be removed. 149 C++ construct 'String' found in C code -- An illegal construct was found in C code. It looked as though it might be suitable for C++. The quoted string identifies the construct further. 150 Token 'String' unexpected String -- An unexpected token was encountered. The action taken, if any, is identified by the second message parameter. 151 Token 'Name' inconsistent with abstract type -- In a context in which an abstract type is allowed such as within a cast or after a sizeof, and after starting to parse the abstract type, an identifier was found. For example: x = (int y) z; 152 Lob base file 'file name' missing -- The indicated file has been specified as the base of lob production via the option -lobbase(). On output, this message is given if the lob base is missing. The situation is correctable by simply producing the missing lob output. This will not be a problem given the appropriate dependencies in the make file. On input, the most likely cause of this message is an out-of-date base file. A hash code within the lob file being read, did not match a similar code already embedded within the base. The input lob file should be considered in error and should be regenerated. 153 Could not create temporary file -- This message is produced when generating a lob output file based upon some lob base file. When the lob file is produced, it is first written to a temporary. The temporary is generated by the C library function tmpnam(). 154 Could not evaluate type 'String', int assumed -- String in the message is the second argument to either a printf_code option or a scanf_code option. When used, it was to be evaluated as a type. Unfortunately the type could not be identified. 155 Ignoring { }'ed sequence within an expression, 0 assumed -- Some compilers support what looks like a compound statement as a C/C++ expression. For example to define the absolute value of an integer which guarantees that it will be read only once you may use: #define abs(a) { int b = a; b >= 0 ? b : -b; } The last expression in the list is the result. To syntactically support the construct without running amuck we recognize the sequence and issue this message. If you want to use the facility just suppress the message. 156 Braced initializer for scalar type 'Name' -- An example of an initializer that will draw this complaint is as follows. int s[] = { { 1 } }; After the compiler has seen the first curly it is expecting to see a number (or other numeric expression). Compilers that strictly adhere to the ISO C and C++ Standards will flag this as ill-formed code. Note that it is legal (but somewhat arcane) to employ a left curly at the top-level when initializing an object of scalar type. For example, the following is well-formed: int i = { 0 }; // OK; initialize scalar i with 0. char *t = { "bar" }; // OK; initialize scalar t with a pointer to // a statically allocated array. Also note: as the example above implies, this message can apply to pointers to arrays of char; it does not apply to arrays. 157 No data may follow an incomplete array -- An incomplete array is allowed within a struct of a C99 or C++ program but no data is allowed to appear after this array. For example: struct A { int x; int a[]; int b; }; This diagnostic is issued when the 'b' is seen. 158 Assignment to variable 'Symbol' (Location) increases capability -- An assignment has been made to a variable that increases capability. A typical capability increase is to remove const protection as in the following example: int *p; const int *q; p = q; // Error 158 If a capability increase is seen in situations other than an assignment or if the variable is not available, Warning 605 is issued. Please see the description of that message for further information concerning capability increase. See also Informational messages 1776 and 1778 in Section 13.8 C++ Informational Messages. 159 enum following a type is non-standard -- Normally two different types are not permitted within the same type specification; this will ordinarily result in Error 104. However, some compilers support 'sized' enumerations wherein a scalar type can precede the enum keyword. E.g. char enum color { red, green, blue }; When the second type is an enum we do not issue a 104 but emit Error 159 instead. By suppressing this message (with -e159) such constructs will be supported. 160 The sequence '( {' is non standard and is taken to introduce a GNU statement expression -- Lint encountered the sequence '( {' in a context where an expression (possibly a sub-expression) is expected. int n = ({ // Error 160 here int y = foo (); int z; if (y > 0) z = y; else z = - y; z; }) // Now n has the last value of z. The primary intention of this message is to alert the user to the non-standard nature of this construct. The typical response is to suppress the message and go on. But a few caveats are in order. Programmers who intend to work only with C code with the GNU extensions may safely disable this diagnostic but C++ users should think twice. This is partly for the reasons given in GCC's documentation (see the section entitled "Statements and Declarations in Expressions") and partly because the meaning of '( {' will change in G++ when its maintainers implement Initializer Lists (a new core language feature that is expected to appear in the 2010 version of the ISO C++ Standard). 161 Repeated use of parameter 'Symbol' in parameter list -- The name of a function parameter was repeated. For example: void f( int n, int m, int n ) {} will cause this message to be issued. Names of parameters for a given function must all be different. 19.2 Internal Errors 2XX Internal Error -- Some inconsistency or contradiction was discovered in the PC-lint/FlexeLint system. This may or may not be the result of a user error. This inconsistency should be brought to the attention of Gimpel Software. 19.3 Fatal Errors Errors in this category are normally fatal and suppressing the error is normally impossible. However, those errors marked with an asterisk(*) can be suppressed and processing will be continued. For example -e306 will allow reprocessing of modules. 301 Stack overflow -- There was a stack overflow while processing declarations. Approximately 50 nested declarators were found. For example, if a '/' followed by 50 consecutive '*'s were to introduce a box-like comment and if the '/' were omitted, then this message would be produced. 302 Exceeded Available Memory -- Main memory has been exhausted. 303 String too long (try +macros) -- A single #define definition or macro invocation exceeded an internal limit (of 4096 characters). As the diagnostic indicates the problem can be corrected with an option. 304 Corrupt object file, code Integer, symbol=String -- A PC-lint/FlexeLint object file is apparently corrupted. Please delete the object module and recreate it using the -oo option. The special code identifier number as well as a list of symbol names are optionally suffixed to the message as an aid in diagnosing the problem by technical support. 305 Unable to open module 'file name' -- file name is the name of the file. The named module could not be opened for reading. Perhaps you misspelled the name. 306 Previously encountered module 'FileName' -- FileName is the name of the module. The named module was previously encountered. This is probably a user blunder. 307 Can't open indirect file 'FileName' -- FileName is the name of the indirect file. The named indirect file (ending in .lnt) could not be opened for reading. 308 Can't write to standard out -- stdout was found to equal NULL. This is most unusual. 309 #error ... -- The #error directive was encountered. The ellipsis reflects the original line. Normally processing is terminated at this point. If you set the fce (continue on #error) flag, processing will continue. 310 Declaration too long: 'String...' -- A single declaration was found to be too long for an internal buffer (about 2000 characters). This occurred when attempting to write out the declaration using the -o... option. The first 30 characters of the declaration is given in String. Typically this is caused by a very long struct whose substructures, if any, are untagged. First identify the declaration that is causing the difficulty. If a struct or union, assign a tag to any unnamed substructures or subunion. A typedef can also be used to reduce the size of such a declaration. 312 Lint Object Module has obsolete or foreign version id: Integer -- A lint object module was produced with a prior or different version of PC-lint/FlexeLint. Delete the.lob file and recreate it using your new version of PC-lint/FlexeLint. 313 Too many files -- The number of files that PC-lint/FlexeLint can process has exceeded an internal limit. The FlexeLint user may recompile his system to increase this limit. Look for symbol FSETLEN in custom.h. Currently, the number of files is limited to 4096. 314 Previously used .lnt file: FileName -- The indirect file named was previously encountered. If this was not an accident, you may suppress this message. 315 Exceeded message limit (see -limit) -- The maximum number of messages was exceeded. Normally there is no limit unless one is imposed by the -limit(n) option. 316 Error while writing to file "file name" -- The given file could not be opened for output. 317 File encoding, String, not currently supported -- Lint detected a byte order mark at the beginning of a file which indicated the file is encoded in the given format. As of this writing, the only formats supported to any extent are ASCII and UTF-8 (for which Lint presumes ASCII encoding). 321 Declaration stack overflow -- An overflow occurred in the stack used to contain array, pointer, function or reference modifiers when processing a declarator. 322 Unable to open include file FileName -- FileName is the name of the include file which could not be opened. Directory search is controlled by options: -i +fdi and the INCLUDE environment variable. This is a suppressible fatal message. If option -e322 is used, Error message 7 will kick in. A diagnostic will be issued but processing will continue. 323 Token String too long -- In attempting to save a token for later reuse, a fixed size buffer was exceeded (governed by the size M_TOKEN). 324 Too many symbols Integer -- Too many symbols were encountered. An internal limit was reached. 325 Cannot re-open file 'file name' -- In the case of a large number of nested includes, files in the outer fringe need to be closed before new ones are opened. These outer files then need to be re-opened. An error occurred when attempting to re-open such a file. 326 String 'String ...' too long, exceeds Integer characters -- A string (first 40 characters provided in the message) exceeds some internal limit (provided in the message). There is no antidote to this condition in the form of an option. FlexeLint customers may recompile with a redefinition of either M_STRING (maximum string) or M_NAME (maximum name). To override the definition in custom.h we suggest recompiling with an appropriate -dvar=value option assuming your compiler supports the option. 328 Bypass header 'Name' follows a different header sequence than in module 'String' which includes File1 where the current module includes File2 -- This message is issued when a header is #include'd that had previously been designated as bypass and it has been determined that this header follows a different header include sequence than in some other module. The name of the other module is given by the second parameter of this message. In order not to bury the programmer under a ton of header names, we have made an effort to determine the precise point where the two modules went their separate ways. The first include file difference occurred when that other module included the header identified by File1, whereas the current module was attempting to include the header identified by File2. Each Filei is a pair of parameters of the form 'String' (Location) where the location is the point of the #include. For example: Module x.cpp: #include "alpha.h" #include "delta.h" #include "beta.h" #include "gamma.h" Module y.cpp: #include "alpha.h" #include "beta.h" #include "gamma.h" When the include of "beta.h" occurs in module y.cpp (and if beta.h has been designated as bypass), there will be a Fatal Error 328 that the header sequence of module 'x.cpp' differs from the current module in that the former module included 'delta.h' at a point where the current module included 'beta.h'. It was necessary to make this message a fatal error since attempting to bypass headers that do not follow a consistent header sequence is an act of folly. It is possible to continue on after the 328 in hopes of picking up more inconsistencies in other modules. This can be done using the +fce (Continue-on-Error) flag. 19.4 C Warning Messages 401 symbol 'Symbol' not previously declared static at Location -- The indicated Symbol declared static was previously declared without the static storage class. This is technically a violation of the ANSI standard. Some compilers will accept this situation without complaint and regard the Symbol as static. 402 static function 'Symbol' (Location) not defined -- The named Symbol was declared as a static function in the current module and was referenced but was not defined (in the module). 403 static symbol 'Symbol' has unusual type modifier -- Some type modifiers such as _export are inconsistent with the static storage class. 404 struct not completed within file 'FileName' -- A struct (or union or enum) definition was started within a header file but was not completed within the same header file. 405 #if not closed off within file 'FileName' -- An #if construct was begun within a header file (name given) but was not completed within that header file. Was this intentional? 406 Comment not closed off within file 'FileName' -- A comment was begun within a header file (name given) but was not completed within that header file. Was this intentional? 407 Inconsistent use of tag 'Symbol' conflicts with Location -- A tag specified as a union, struct or enum was respecified as being one of the other two in the same module. For example: struct tag *p; union tag *q; will elicit this message. 408 Type mismatch with switch expression -- The expression within a case does not agree exactly with the type within the switch expression. For example, an enumerated type is matched against an int. 409 Expecting a pointer or array -- An expression of the form i[...] was encountered where i is an integral expression. This could be legitimate depending on the subscript operand. For example, if i is an int and a is an array then i[a] is legitimate but unusual. If this is your coding style, suppress this message. 410 size_t not what was expected from fzl and/or fzu, using 'Type' -- This warning is issued if you had previously attempted to set the type of sizeof by use of the options +fzl, -fzl, or -fzu, and a later size_t declaration contradicts the setting. This usually means you are attempting to lint programs for another system using header files for your own system. If this is the case we suggest you create a directory housing header files for that foreign system, alter size_t within that directory, and lint using that directory. 411 ptrdiff_t not what was expected from fdl option, using 'Type' -- This warning is issued if you had previously attempted to set the type of pointer differences by use of the fdl option and a later ptrdiff_t declaration contradicts the setting. See suggestion in Error Message 410. 412 Ambiguous format specifier '%X' -- The format specifier %X when used with one of the scanf family, is ambiguous. With Microsoft C it means %lx whereas in ANSI C it has the meaning of %x. This ambiguous format specification has no place in any serious C program and should be replaced by one of the above. 413 Likely use of null pointer 'Symbol' in [left/right] argument to operator 'String' Reference -- From information gleaned from earlier statements, it appears certain that a null pointer (a pointer whose value is 0) has been used in a context where null pointers are inappropriate. These include: Unary *, pointer increment (++) or decrement(--), addition of pointer to numeric, and subtraction of two pointers. In the case of binary operators, one of the words 'left' or 'right' is used to designate which operand is null. Symbol identifies the pointer variable that may be null. 414 Possible division by 0 -- The second argument to either the division operator (/) or the modulus operator (%) may be zero. Information is taken from earlier statements including assignments, initialization and tests. 415 access of out-of-bounds pointer ('Integer' beyond end of data) by operator 'String' -- An out-of-bounds pointer was accessed. String designates the operator. The parameter 'Integer' gives some idea how far out of bounds the pointer may be. It is measured in units given by the size of the pointed to object. The value is relative to the last item of good data and therefore should always be greater than zero. For example: int a[10]; a[10] = 0; results in an overflow message containing the phrase '1 beyond end of data'. 416 creation of out-of-bounds pointer ('Integer' beyond end of data) by operator 'String' -- An out-of-bounds pointer was created. See message 415 for a description of the parameters Integer and String. For example: int a[10]; ... f( a + 11 ); Here, an illicit pointer value is created and is flagged as such by PC-lint/FlexeLint. Note that the pointer a+10 is not considered by PC-lint/FlexeLint to be the creation of an out-of-bounds pointer. This is because ANSI C explicitly allows pointing just beyond an array. Access through a+10, however, as in *(a+10) or the more familiar a[10], would be considered erroneous but in that case message 415 would be issued. 417 integral constant 'String' has precision Number which is longer than long long int -- The longest possible integer is by default 8 bytes (see the +fll flag and then the -sll# option). An integral constant was found to be even larger than such a quantity. For example: 0xFFFF0000FFFF0000F. String is the token in error. 418 Passing null pointer to function 'Symbol', Context Reference -- A NULL pointer is being passed to a function identified by Symbol. The argument in question is given by Context. The function is either a library function designed not to receive a NULL pointer or a user function dubbed so via the option -function. 419 Apparent data overrun for function 'Symbol', argument Integer exceeds argument Integer -- This message is for data transfer functions such as memcpy, strcpy, fgets, etc. when the size indicated by the first cited argument (or arguments) exceeds the size of the buffer area cited by the second. The message may also be issued for user functions via the -function option. 420 Apparent access beyond array for function 'Symbol', argument Integer exceeds Integer Reference -- This message is issued for several library functions (such as fwrite, memcmp, etc.) wherein there is an apparent attempt to access more data than exist. For example, if the length of data specified in the fwrite call exceeds the size of the data specified. The function is specified by Symbol and the arguments are identified by argument number. 421 Caution -- function 'Symbol' is considered dangerous -- This message is issued (by default) for the built-in function gets. This function is considered dangerous because there is no mechanism to ensure that the buffer provided as first argument will not overflow. A well known computer virus (technically a worm) was created based on this defect. Through the -function option, the user may designate other functions as dangerous. 422 Passing to function 'Symbol' a negative value (Integer), Context Reference -- An integral value that appears to be negative is being passed to a function that is expecting only positive values for a particular argument. The message contains the name of the function (Symbol), the questionable value (Integer) and the argument number (Context). The function may be a standard library function designed to accept only positive values such as malloc or memcpy (third argument), or may have been identified by the user as such through the -function or -sem options. The negative integral value may in fact be unsigned. Thus: void *malloc( unsigned ); void f() { int n = -1; int *p; p = malloc(n); // Warning 422 p = malloc( (unsigned) n ); // Warning 422 } will result in the warnings indicated. Note that casting the expression does not inhibit the warning. There is a slight difference in behavior on 32-bit systems versus 16-bit systems. If long is the same size as int (as in 32-bit systems) the warning is issued based upon the sign bit. If long is larger than an int (as is true on typical 16-bit systems) the warning is issued if the value was a converted negative as in the examples above. It is not issued if an unsigned int has the high-order bit set. This is because it is not unreasonable to malloc more that 32,176 bytes in a 16-bit system. 423 Creation of memory leak in assignment to variable 'Symbol' -- An assignment was made to a pointer variable (designated by Symbol) which appeared to already be holding the address of an allocated object which had not been freed. The allocation of memory which is not freed is considered a memory leak. 424 Inappropriate deallocation (Name1) for 'Name2' data. -- This message indicates that a deallocation (free(), delete, or delete[]) as specified by String1 is inappropriate for the data being freed. [12, Item 5] The kind of data (specified by String2) is one or more of: malloc, new, new[], static, auto, member, modified or constant. These have the meanings as described below: malloc data is data obtained from a call to malloc, calloc or realloc. new and new[] data is data derived from calls to new. static data is either static data within a function or external data. auto data is non-static data in a function. member data is a component of a structure (and hence can't be independently freed). modified data is the result of applying pointer arithmetic to some other pointer. E.g. p = malloc(100); free( p+1 ); // warning p+1 is considered modified. constant data is the result of casting a constant to a pointer. E.g. int *p = (int *) Ox80002; free(p); // warning 425 'Message' in processing semantic 'String' at token 'String' -- This warning is issued when a syntax error is encountered while processing a Semantic option (-sem). The 'Message' depends upon the error. The first 'String' represents the portion of the semantic being processed. The second 'String' denotes the token being scanned when the error is first noticed. 426 Call to function 'Symbol' violates semantic 'String' -- This Warning message is issued when a user semantic (as defined by -sem) is violated. 'String' is the subportion of the semantic that was violated. For example: //lint -sem( f, 1n > 10 && 2n > 10 ) void f( int, int ); ... f( 2, 20 ); results in the message: Call to function 'f(int, int)' violates semantic '(1n>10)' 427 // comment terminates in \ -- A one-line comment terminates in the back-slash escape sequence. This means that the next line will be absorbed in the comment (by a standards-conforming compiler -- not all compilers do the absorption, so beware). It is much safer to end the line with something other than a back-slash. Simply tacking on a period will do. If you really intend the next line to be a comment, the line should be started with its own double slash (//). 428 negative subscript (Integer) in operator 'String' -- A negative integer was added to an array or to a pointer to an allocated area (allocated by malloc, operator new, etc.) This message is not given for pointers whose origin is unknown since a negative subscript is, in general, legal. The addition could have occurred as part of a subscript operation or as part of a pointer arithmetic operation. The operator is denoted by String. The value of the integer is given by Integer. 429 Custodial pointer 'Symbol' (Location) has not been freed or returned -- A pointer of auto storage class was allocated storage which was neither freed nor returned to the caller. This represents a "memory leak". A pointer is considered custodial if it uniquely points to the storage area. It is not considered custodial if it has been copied. Thus: int *p = new int[20]; // p is a custodial pointer int *q = p; // p is no longer custodial p = new int[20]; // p again becomes custodial q = p + 0; // p remains custodial Here p does not lose its custodial property by merely participating in an arithmetic operation. A pointer can lose its custodial property by passing the pointer to a function. If the parameter of the function is typed pointer to const or if the function is a library function, that assumption is not made. For example p = malloc(10); strcpy (p, "hello"); Then p still has custody of storage allocated. It is possible to indicate via semantic options that a function will take custody of a pointer. 430 Character '@', taken to specify variable location, is not standard C/C++ -- Many compilers for embedded systems have a declaration syntax that specifies a location in place of an initial value for a variable. For example: int x @0x2000; specifies that variable x is actually location 0x2000. This message is a reminder that this syntax is non-standard (although quite common). If you are using this syntax on purpose, suppress this message. 431 Missing identifier for template parameter number Integer -- A template object parameter (as opposed to a type parameter) was not provided with an identifier. Was this an oversight? 432 Suspicious argument to malloc -- The following pattern was detected: malloc( strlen(e+1) ) where e is some expression. This is suspicious because it closely resembles the commonly used pattern: malloc( strlen(e)+1 ) If you really intended to use the first pattern then an equivalent expression that will not raise this error is: malloc( strlen(e)-1 ) 433 Allocated area not large enough for pointer -- An allocation was assigned to a pointer whose reach extends beyond the area that was allocated. This would usually happen only with library allocation routines such as malloc and calloc. For example: int *p = malloc(1); This message is also provided for user-declared allocation functions. For example, if a user's own allocation function is provided with the following semantic: -sem(ouralloc,@P==malloc(1n)) We would report the same message. Please note that it is necessary to designate that the returned area is freshly allocated (ala malloc). This message is always given in conjunction with the more general Informational Message 826. 434 White space ignored between back-slash and new-line -- According to the C and C++ standards, any back-slash followed immediately by a new-line results in the deletion of both characters. For example: #define A \ 34 defines A to be 34. If a blank or tab intervenes between the back-slash and the new-line then according to a strict interpretation of the standard you have defined A to be a back-slash. But this blank is invisible to the naked eye and hence could lead to confusion. Worse, some compilers silently ignore the white-space and the program becomes non-portable. You should never deliberately place a blank at the end of a line and any such blanks should be removed. If you really need to define a macro with a terminal back-slash you can use a comment as in: #define A \ /* commentary */ 435 integral constant 'String' has precision Integer, use +fll to enable long long" -- An integer constant was found that had a precision that was too large for a long but would fit within a long long. Yet the +fll flag that enables the long long type was not set. Check the sizes that you specified for long (-sl#) and for long long (-sll#) and make sure they are correct. Turn on +fll if your compiler supports long long. Otherwise use smaller constants. 436 Apparent preprocessor directive in invocation of macro 'Symbol' -- A function like macro was invoked whose arguments extended for multiple lines which included preprocessor statements. This is almost certainly an error brought about by a missing right parenthesis. By the rules of Standard C the preprocessing directive is absorbed into the macro argument but then will not subsequently get executed. For this reason some compilers treat the apparent preprocessor directive as a directive. This is logical but not portable. It is therefore best to avoid this construct. 437 Passing struct 'Symbol' to ellipsis -- A struct is being passed to a function at a parameter position identified by an ellipsis. For example: void g() { struct A { int a; } x; void f( int, ... ); f( 1, x ); ... } This is sufficiently unusual that it is worth pointing out on the likely hood that this is unintended. The situation becomes more severe in the case of a Non-POD struct [10]. In this case the behavior is considered undefined. 438 Last value assigned to variable 'Symbol' not used -- A value had been assigned to a variable that was not subsequently used. The message is issued either at a return statement or at the end of a block when the variable goes out of scope. For example, consider the following function: void f( int n ) { int x = 0, y = 1; if( n > 0 ) { int z; z = x + y; if( n > z ) { x = 3; return; } z = 12; } } Here we can report that x was assigned a value that had not been used by the time the return statement had been encountered. We also report that the most recently assigned value to z is unused at the point that z goes out of scope. See message 838 in Section 13.4 C Informational Messages and flags -fiw and -fiz in Sections 2.3.6 Initialization-is-considered-a-Write flag (-fiw) and 2.3.7 Initialization-by-Zero-is-considered-a-Write flag (-fiz) . This message is suppressed if the variable's address is assigned to a pointer (or, equivalently, the variable is used to directly initialize a reference to non-const). 440 for clause irregularity: variable 'Symbol' tested in 2nd expression does not match 'Symbol' modified in 3rd -- A for clause has a suspicious structure. The loop variable, as determined by an examination of the 3rd for clause expression, does not match the variable that is tested in the 2nd for clause expression. For example: for( i = 0; i < 10; j++ ) ... would draw this complaint since the 'i' of the 2nd expression does not match the 'j' of the third expression. 441 for clause irregularity: loop variable 'Symbol' not found in 2nd for expression -- The loop variable is determined by an examination of the 3rd for clause expression. A loop variable was found (and its name is given in the message) but it did not appear as one of the accessed symbols of the condition expression (the 2nd for expression). For example: for( p = a; *p; j++ ) ... would draw this complaint since the 2nd expression does not contain the 'j' of the third expression. 442 for clause irregularity: testing direction inconsistent with increment direction -- A for clause was encountered that appeared to have a parity problem. For example: for( i = 0; i < 10; i-- ) ... Here the test for i less than 10 seems inconsistent with the 3rd expression of the for clause which decreases the value of i. This same message would be given if i were being increased by the 3rd expression and was being tested for being greater than some value in the 2nd expression. 443 for clause irregularity: variable 'Symbol' initialized in 1st expression does not match 'Symbol' modified in 3rd -- A for clause has a suspicious structure. The loop variable, as determined by an examination of the 3rd for clause expression, does not match the variable that is initialized in the 1st expression. For example: for( ii = 0; i < 10; i++ ) ... would draw this complaint since the 'ii' of the 1st expression does not match the 'i' of the third expression. 444 for clause irregularity: pointer 'Symbol' incremented in 3rd expression is tested for NULL in 2nd expression -- The following kind of situation has been detected: for( ... ; p == NULL; p++ ) ... A loop variable being incremented or decremented would not normally be checked to see if it is NULL. This is more likely a programmer error. 445 reuse of for loop variable 'Symbol' at 'Location' could cause chaos -- A for loop nested within another for loop employed the same loop variable. For example: for( i = 0; i < 100; i++ ) { ... for( i = 0; i < n; i++ ) { ... } } 446 for loop index variable 'Symbol' modified in body of the for loop -- A for loop with an identifiable loop index variable was programmed in such a way that the loop body also modifies the index variable. For example: for( i = 0; i < 100; i++ ) { a[i++] = 0; } In general it is better to restrict modifications to for loop index variables to the for clause if at all possible. If this is not possible, you can prefix the for loop with an appropriate lint comment such as: /*lint -e{446} i is modified in the body of the for loop */ 447 Extraneous whitespace ignored in include directive for file 'FileName'; opening file 'FileName' -- A named file was found to contain either leading or trailing whitespace in the #include directive. While legal, the ISO Standards allow compilers to define how files are specified or the header is identified, including the appearance of whitespace characters immediately after the < or opening " or before the > or closing ". Since filenames tend not to contain leading or trailing whitespace, Lint ignores the (apparently) extraneous characters and processes the directive as though the characters were never given. The use of a -efile option on either String for this message will cause Lint to process #include's with whitespace intact. 448 Likely access of pointer pointing Integer bytes past nul character by operator 'String' -- Accessing past the terminating nul character is often an indication of a programmer error. For example: char buf[20]; strcpy( buf, "a" ); char c = buf[4]; // legal but suspect. Although buf has 20 characters, after the strcpy, there would be only two that the programmer would normally be interested in. 449 Pointer variable 'Symbol' previously deallocated -- A pointer variable (designated in the message) was freed or deleted in an earlier statement. 451 Header file 'FileName' repeatedly included but does not have a standard include guard -- The file named in the message has already been included in the current module. Moreover it has been determined that this header does not have a standard include guard. A standard include guard has the form #ifndef Name #define Name ... #endif with nothing but comments before and after this sequence and nothing but comments between the #ifndef and the #define Name. This warning may also be accompanied by a 537 (repeated include header). Message 537 is often suppressed because if you are working with include guards, it is not a helpful message. However, the message 451 should be left on in order to check the consistency of the include guards themselves. See also Elective Note 967 in Section 13.5 C Elective Notes. 452 typedef Symbol 'Symbol' redeclared (TypeDiff) conflicts with Location", -- A typedef symbol is being declared to be a different type. This can be legal, especially with multiple modules, but is not good programming practice. It inteferes with program legibility. 453 Function 'Symbol', previously designated pure, String 'Name' -- A semantic option designated that the named function, Symbol, is pure (lacking non-local side-effects; see the pure semantic in Section 4. Semantics ). However, an impurity was detected. Such impurities include calling a function through a function pointer, accessing a volatile variable, modifying a static variable or calling a function whose purity PC-lint/FlexeLint cannot verify. String describes which of these reasons apply and Name shows the related variable or function, as appropriate. Despite the inconsistency reported, the function will continue to be regarded as pure. 454 A thread mutex has been locked but not unlocked -- A return point in a function has been reached such that a mutex lock that had been previously set has not been unlocked. E.g., //lint -sem( lock, thread_lock ) void f( int x ) { lock(); if( x < 0 ) return; // Warning 454 ... 455 A thread mutex that had not been locked is being unlocked -- A call to an unlock() function was made that was not preceded by a balancing lock(). It is assumed that every mutex lock() function must be balanced by exactly one unlock() function, no more, no less. For example: //lint -sem( lock, thread_lock ) //lint -sem( unlock, thread_unlock ) void f( bool x ) { lock(); /* something */; unlock(); /* something else */ unlock(); // Warning 455 } 456 Two execution paths are being combined with different mutex lock states -- It is the purpose of this message to make absolutely certain that every lock has a corresponding unlock in the same unbroken sequence of code in the same function. Execution paths can be combined at the end of an if statement, switch statement, or the begining of while, for and do statements, a label (target of goto), etc. In all these cases we check to make sure that the mutex lock states are the same. For example: //lint -sem( lock, thread_lock ) void f( bool x ) { if( x ) lock(); // Warning 456 issued here ... It could be argued that if an unlock() call would appear under control of the very same bool x in the example above then all would be well. And if this is your coding style you are free to turn this message off. But errors in mutex locking have such horrible programming consequences as to suggest especially strong measures to assure code correctness. We recommend, for example: //lint -sem( lock, thread_lock ) //lint -sem( unlock, thread_unlock ) void f( bool x ) { if( x ) { lock(); /* something */; unlock(); } else { /* something */ } } If the 'something' that is being executed is sufficently complex, then it can be made into a function. 457 Function 'Symbol1' of thread 'Symbol2' has an unprotected write access to variable 'Symbol3' which is used by function 'Symbol4' of thread 'Symbol5' -- A variable (Symbol3) was modified by function (Symbol1) of thread (Symbol2) outside of any recognized mutex lock. It was also accessed by a function (Symbol3) of a second thread (Symbol5). The latter access may or may not have been protected. If unprotected, a second message will be issued with the roles of Symbol1 and Symbol4 interchanged. 458 Function 'Symbol1' of thread 'Symbol2' has an unprotected read access to variable 'Symbol3' which is modified by function 'Symbol4' of thread 'Symbol5' -- A variable identified in the message was accessed (non-modifiing) by a function (Symbol1) of thread (Symbol2) outside of any recognized mutex lock. It was also modified by a function (Symbol4) in a second thread (Symbol5). The modification may or may not have been protected. If unprotected, Warning 457 will also be issued. 459 Function 'Symbol' whose address was taken has an unprotected access to variable 'Symbol' -- This message is activated only when it appears that the program has more than one thread. See Section 8. Multi-thread Suppport to determine what those conditions might be. If a function's address is taken, we presume that we are unable to determine statically all the locations from which the function may be called and so we presume that any and all threads can call this function and so the function needs to have protected access to every static variable that it might touch. There are several remedies to such a message. If multiple threads can indeed access this function, then place a mutex lock in the function. If there already is a mutex lock and we don't recognize it, then set the thread_protected semantic for the function. If only one thread really accesses this function or if the access is guaranteed to be benign, then, after making sure this condition is commented in the code, use the same thread_protected semantic for the function. 460 Thread 'Symbol' has unprotected call to thread unsafe function 'Symbol' which is also called by thread 'Symbol' -- The second symbol in the message represents a function that was designated as being thread_unsafe through the -sem option. It was being called in an unprotected region of a thread whose root function is the first symbol in the message. Another thread is also accessing this function and this thread is identified by the third parameter of the message. Calls to thread unsafe functions need to be protected by mutex locks if they are to be employed by more than one thread. 461 Thread 'Symbol' has unprotected call to function 'Symbol' of group 'Name' while thread 'Symbol' calls function 'Symbol' of the same group -- This message is similar to Warning 460 in that a thread (identified in the message as the first Symbol) is making a call on a function (the second Symbol) which had been deduced (through options) as being thread unsafe. Like message 460 there is another thread that is also doing some calling. In this case the other thread is not calling the same function as the first but one which has been placed within the same group (identified by the third parameter) as the first function. See Section 8. Multi-thread Support to obtain further information on thread unsafe function groups and options to determine them. 462 Thread 'Symbol' calling function 'Symbol' is inconsistent with the 'String' semantic -- The first Symbol in the message identifies a thread. The second Symbol identifies a function called directly or indirectly by the thread. The String argument specifies a semantic that had been attributed to the functon. It should have one of the following forms: thread_not thread_not( list ) thread_only( list ) If the second form is given, it means that the thread appears on the list. If the 3rd form is given it means that the thread was not on the list. 464 Buffer argument will be copied into itself -- This is issued when we encounter a function argument expression used in such a way that there will be an attempt to copy its contents onto itself. E.g. sprintf( s, "%s", s ); 501 Expected signed type -- The unary minus operator was applied to an unsigned type. The resulting value is a positive unsigned quantity and may not be what was intended. 502 Expected unsigned type -- Unary ~ being a bit operator would more logically be applied to unsigned quantities rather than signed quantities. 503 Boolean argument to relational -- Normally a relational would not have a Boolean as argument. An example of this is a < b < c which is technically legal but does not produce the same result as the mathematical expression which it resembles. 504 Unusual shift operation (String) -- Either the quantity being shifted or the amount by which a quantity is to be shifted was derived in an unusual way such as with a bit-wise logical operator, a negation, or with an unparenthesized expression. If the shift value is a compound expression that is not parenthesized, parenthesize it. 505 Redundant left argument to comma -- The left argument to the comma operator had no side effects in its top-most operator and hence is redundant. 506 Constant value Boolean -- A Boolean, i.e., a quantity found in a context that requires a Boolean such as an argument to && or || or an if() or while() clause or ! was found to be a constant and hence will evaluate the same way each time. 507 Size incompatibility -- A cast was made to an integral quantity from a pointer and according to other information given or implied it would not fit. For example a cast to an unsigned int was specified and information provided by the options indicate that a pointer is are larger than an int. 508 extern used with definition -- A function definition was accompanied with an extern storage class. extern is normally used with declarations rather than with definitions. At best the extern is redundant. At worst you may trip up a compiler. 509 extern used with definition -- A data object was defined with a storage class of extern. This is technically legal in ANSI and you may want to suppress this message. However, it can easily trip up a compiler and so the practice is not recommended at this time. 511 Size incompatibility -- A cast was made from an integral type to a pointer and the size of the quantity was too large to fit into the pointer. For example if a long is cast to a pointer and if options indicate that a long ise larger than a pointer, this warning would be reported. 512 Symbol 'Symbol' previously used as static (Location) -- The Symbol name given is a function name that was declared as static in some other module (the location of that declaration is provided). The use of a name as static in one module and external in another module is legal but suspect. 514 Unusual use of a Boolean -- An argument to an arithmetic operator (+ - / * %) or a bit-wise logical operator (| & ^) was a Boolean. This can often happen by accident as in: if( flags & 4 == 0 ) where the ==, having higher precedence than &, is done first (to the puzzlement of the programmer). 515 Symbol 'Symbol' has arg. count conflict (Integer vs. Integer) with Location -- An inconsistency was found in the number of actual arguments provided in a function call and either the number of formal parameters in its definition or the number of actual arguments in some other function call. See the +fva option to selectively suppress this message. 516 Symbol 'Symbol' has arg. type conflict (no. Integer -- TypeDiff) with Location -- An inconsistency was found in the type of an actual argument in a function call with either the type of the corresponding formal parameter in the function definition or the type of an actual argument in another call to the same function or with the type specified for the argument in the function's prototype. The call is not made in the presence of a prototype. See options -ean, -eau, -eas and -eai for selective suppression of some kinds of type differences. If the conflict involves types char or short then you may want to consider using the +fxc or +fxs option. 517 defined not K&R -- The defined function (not a K&R construct) was employed and the K&R preprocessor flag (+fkp) was set. Either do not set the flag or do not use defined. 518 Expected '(' -- sizeof type is not strict C. sizeof(type) or sizeof expression are both permissible. 519 Size incompatibility -- An attempt was made to cast a pointer to a pointer of unequal size. This could occur for example in a P model where pointers to functions require 4 bytes whereas pointers to data require only 2. This error message can be circumvented by first casting the pointer to an integral quantity (int or long) before casting to a pointer. 520 Highest operator or function lacks side-effects -- The first expression of a for clause should either be one of the privileged operators: assignment, increment, decrement or call to an impure function or one modifying its argument(s). 521 Highest operator or function lacks side-effects -- The third expression of a for clause should either be one of the privileged operators: assignment, increment, decrement or call to an impure function or one modifying its argument(s). 522 Highest operator or function lacks side-effects -- If a statement consists only of an expression, it should either be one of the privileged operators: assignment, increment, decrement or call to an impure function or one modifying its argument(s). For example if operator * is the built-in operator, the statement *p++; draws this message but p++; does not. This is because the highest operator is '*' which has no side effects. The definition of pure and impure functions and function calls which have side effects are given in the discussion of the pure semantic in section 4. Semantics 524 Loss of precision (Context) (Type to Type) -- There is a possible loss of a fraction in converting from a float to an integral quantity. Use of a cast will suppress this message. 525 Negative indentation from Location -- The current line was found to be negatively indented (i.e., not indented as much) from the indicated line. The latter corresponds to a clause introducing a control structure and statements and other control clauses and braces within its scope are expected to have no less indentation. If tabs within your program are other than 8 blanks you should use the -t option. 526 'Symbol' (Location) not defined -- The named external was referenced but not defined and did not appear declared in any library header file nor did it appear in a Library Module. This message is suppressed for unit checkout (-u option). Please note that a declaration, even one bearing prototype information is not a definition. See the glossary at the beginning of this chapter. If the Symbol is a library symbol, make sure that it is declared in a header file that you're including. Also make sure that the header file is regarded by PC-lint/FlexeLint as a Library Header file. Alternatively, the symbol may be declared in a Library Module. 527 Unreachable code at token Symbol -- A portion of the program cannot be reached. 528 Symbol 'Symbol' (Location) not referenced -- The named static variable or static function was not referenced in the module after having been declared. 529 Symbol 'Symbol' (Location) not subsequently referenced -- The named variable was declared but not referenced in a function. 530 Symbol 'Symbol' (Location) not initialized -- An auto variable was used before it was initialized. 531 Field size too large for 'Symbol' -- The size given for a bit field of a structure exceeds the size of an int. 532 Return mode of function 'Symbol' inconsistent with Location -- A declaration (or a definition) of a function implies a different return mode than a previous statement. (The return mode of a function has to do with whether the function does, or does not, return a value). A return mode is determined from a declaration by seeing if the function returns void or, optionally, by observing whether an explicit type is given. See the fdr flag for a further explanation of this. See also the fvr and fvo flags. 533 function 'Symbol' should (not) return a value (see Location) -- A return statement within a function (or lack of a return at the end of the function) implies a different return mode than a previous statement at Location (The return mode of a function has to do with whether the function does, or does not, return a value.) 534 Ignoring return value of function 'Symbol' (compare with Location) -- A function that returns a value is called just for side effects as, for example, in a statement by itself or the left-hand side of a comma operator. Try: (void) function(); to call a function and ignore its return value. See also the fvr, fvo and fdr flags. 537 Repeated include file 'FileName' -- The file whose inclusion within a module is being requested has already been included in this compilation. The file is processed normally even if the message is given. If it is your standard practice to repeat included files then simply suppress this message. 538 Excessive size -- The size of an array equals or exceeds 64K bytes. 539 Did not expect positive indentation from Location -- The current line was found to be positively indented from a clause that did not control the line in question. For example: if( n > 0 ) x = 3; y = 4; will result in this warning being issued for y = 4;. The Location cited will be that of the if clause. 540 Excessive size -- A string initializer required more space than what was allocated. 541 Excessive size -- The size of a character constant specified with \xddd or \xhhh equalled or exceeded 2**b where b is the number of bits in a byte (established by the -sb option). The default is -sb8. 542 Excessive size for bit field -- An attempt was made to assign a value into a bit field that appears to be too small. The value to be assigned is either another bit field larger than the target, or a numeric value that is simply too large. You may cast the value to the generic unsigned type to suppress the error. You may get this message unexpectedly if the base of the bit field is an int. For example: struct { int b : 1 } s; s.b = 1; /* Warning - - requires 0 or -1 */ The solution in this case is to use 'unsigned' rather than 'int' in the declaration of b. 544 endif or else not followed by EOL -- The preprocessor directive #endif should be followed by an end-of-line. Some compilers specifically allow commentary to follow the #endif. If you are following that convention simply turn this error message off. 545 Suspicious use of & -- An attempt was made to take the address of an array name. At one time such an expression was officially illegal (K&R C [1]), was not consistently implemented, and was, therefore, suspect. However, the expression is legal in ANSI C and designates a pointer to an array. For example, given int a[10]; int (*p) [10]; Then a and &a, as pointers, both represent the same bit pattern, but whereas a is a pointer to int, &a is a pointer to array 10 of int. Of the two only &a may be assigned to p without complaint. If you are using the & operator in this way, we recommend that you disable this message. 546 Suspicious use of & -- An attempt was made to take the address of a function name. Since names of functions by themselves are promoted to address, the use of the & is redundant and could be erroneous. 547 Redefinition of symbol 'Symbol' conflicts with Location -- The indicated symbol had previously been defined (vis #define)to some other value. 548 else expected -- A construct of the form if(e); was found which was not followed by an else. This is almost certainly an unwanted semi-colon as it inhibits the if from having any effect. 549 Suspicious cast -- A cast was made from a pointer to some enumerated type or from an enumerated type to a pointer. This is probably an error. Check your code and if this is not an error, then cast the item to an intermediate form (such as an int or a long) before making the final cast. 550 Symbol 'Symbol' (Location) not accessed -- A variable (local to some function) was not accessed. This means that the value of a variable was never used. Perhaps the variable was assigned a value but was never used. Note that a variable's value is not considered accessed by autoincrementing or autodecrementing unless the autoincrement/decrement appears within a larger expression which uses the resulting value. The same applies to a construct of the form: var += expression. If an address of a variable is taken, its value is assumed to be accessed. An array, struct or union is considered accessed if any portion thereof is accessed. 551 Symbol 'Symbol' (Location) not accessed -- A variable (declared static at the module level) was not accessed though the variable was referenced. See the explanation under message 550 (above) for a description of "access". 552 Symbol 'Symbol' (Location) not accessed -- An external variable was not accessed though the variable was referenced. See the explanation under message 550 above for a description of "access". 553 Undefined preprocessor variable 'Name', assumed 0 -- The indicated variable had not previously been defined within a #define statement and yet it is being used in a preprocessor condition of the form #if or #elif. Conventionally all variables in preprocessor expressions should be pre-defined. The value of the variable is assumed to be 0. 555 #elif not K&R -- The #elif directive was used and the K&R preprocessor flag (+fkp) was set. Either do not set the flag or do not use #elif. 556 indented # -- A preprocessor directive appeared indented within a line and the K&R preprocessor flag (+fkp) was set. Either do not set the flag or do not indent the #. 557 unrecognized format -- The format string supplied to printf, fprintf, sprintf, scanf, fscanf, or sscanf was not recognized. It is neither a standard format nor is it a user-defined format (see printf_code and scanf_code). 558 Too few arguments for format (Integer missing) -- The number of arguments supplied to printf, sprintf, fprintf, scanf, fscanf or sscanf was inconsistent with the number expected as a result of analyzing the format string. 559 size of argument number Integer inconsistent with format -- The given argument (to printf, sprintf, or fprintf) was inconsistent with that which was anticipated as the result of analyzing the format string. Argument counts begin at 1 and include file, string and format specifications. For example, sprintf( buffer, "%f", 371 ) will show an error in argument number 3 because constant 371 is not floating point. 560 argument no. Integer should be a pointer -- The given argument (to one of the scanf or printf family of functions) should be a pointer. For the scanf family, all arguments corresponding to a format specification should be pointers to areas that are to be modified (receive the results of scanning). For the printf family, arguments corresponding to %s or %n also need to be pointers. Argument counts begin at 1 and include file, string and format specifications. For example scanf( "%f", 3.5 ) will generate the message that argument no. 2 should be a pointer. 561 (arg. no. Integer) indirect object inconsistent with format -- The given argument (to scanf, sscanf, or fscanf) was a pointer to an object that was inconsistent with that which was anticipated as the result of analyzing the format string. Argument counts begin at 1 and include file, string and format specifications. For example if n is declared as int then: scanf( "%c", &n ) will elicit this message for argument number 2. 562 Ellipsis (...) assumed -- Within a function prototype a comma was immediately followed by a right parenthesis. This is taken by some compilers to be equivalent to an ellipsis (three dots) and this is what is assumed by PC-lint/FlexeLint. If your compiler does not accept the ellipsis but makes this assumption, then you should suppress this message. 563 Label 'Symbol' (Location) not referenced -- The Symbol at the cited Location appeared as a label but there was no statement that referenced this label. 564 variable 'Symbol' depends on order of evaluation -- The named variable was both modified and accessed in the same expression in such a way that the result depends on whether the order of evaluation is left-to-right or right-to-left. One such example is: n + n++ since there is no guarantee that the first access to n occurs before the increment of n. Other, more typical cases, are given in the manual. Volatile variables are also checked for repeated use in an expression. 565 tag 'Symbol' not previously seen, assumed file-level scope -- The named tag appeared in a prototype or in an inner block and was not previously seen in an outer (file-level) scope. The ANSI standard is dubious as to how this tag could link up with any other tag. For most compilers this is not an error and you can safely suppress the message. On the other hand, to be strictly in accord with ANSI C you may place a small stub of a declaration earlier in the program. For example: struct name; is sufficient to reserve a place for name in the symbol table at the appropriate level. 566 Inconsistent or redundant format char 'Char' -- This message is given for format specifiers within formats for the printf/scanf family of functions. The indicated character Char found in a format specifier was inconsistent or redundant with an earlier character found in the same format specifier. For example a format containing "%ls" will yield this error with the character 's' indicated. This is because the length modifier is designed to be used with integral or float conversions and has no meaning with the string conversion. Such characters are normally ignored by compilers. 567 Expected a numeric field before char 'Char' -- This message is given for format specifiers within formats for the printf/scanf family of functions. A numeric field or asterisk was expected at a particular point in the scanning of the format. For example: %-d requests left justification of a decimal integer within a format field. But since no field width is given, the request is meaningless. 568 nonnegative quantity is never less than zero. -- Comparisons of the form: u >= 0 0 <= u u < 0 0 > u are suspicious if u is an unsigned quantity or a quantity judged to be never less then 0. See also message 775. 569 Loss of information (Context) (Integer bits to Integer bits) -- An assignment (or implied assignment, see Context) was made from a constant to an integral variable that is not large enough to hold the constant. Examples include placing a hex constant whose bit requirement is such as to require an unsigned int into a variable typed as int. The number of bits given does not count the sign bit. 570 Loss of sign (Context) (Type to Type) -- An assignment (or implied assignment, see Context) is being made from a negative constant into an unsigned quantity. Casting the constant to unsigned will remove the diagnostic but is this what you want. If you are assigning all ones to an unsigned, remember that ~0 represents all ones and is more portable than -1. 571 Suspicious Cast -- Usually this warning is issued for casts of the form: (unsigned) ch where ch is declared as char and char is signed. Although the cast may appear to prevent sign extension of ch, it does not. Following the normal promotion rules of C, ch is first converted to int which extends the sign and only then is the quantity cast to unsigned. To suppress sign extension you may use: (unsigned char) ch Otherwise, if sign extension is what you want and you just want to suppress the warning in this instance you may use: (unsigned) (int) ch Although these examples have been given in terms of casting a char they will also be given whenever this cast is made upon a signed quantity whose size is less than the casted type. Examples include signed bit fields (a possibility in the new standard), expressions involving char, and expressions involving short when this type is smaller than int or a direct cast of an int to an unsigned long (if int's is smaller than long). This message is not issued for constants or for expressions involving bit operations. 572 Excessive shift value (precision Integer shifted right by Integer) -- A quantity is being shifted to the right whose precision is equal to or smaller than the shifted value. For example, ch >> 10 will elicit this message if ch is typed char and where char is less than 10 bits wide (the usual case). To suppress the message you may cast the shifted quantity to a type whose length is at least the length of the shift value. 573 Signed-unsigned mix with divide -- one of the operands to / or % was signed and the other unsigned; moreover the signed quantity could be negative. For example: u / n where u is unsigned and n is signed will elicit this message whereas: u / 4 will not, even though 4 is nominally an int. It is not a good idea to mix unsigned quantities with signed quantities in any case (a 737 will also be issued) but, with division, a negative value can create havoc. For example, the innocent looking: n = n / u will, if n is -2 and u is 2, not assign -1 to n but will assign some very large value. To resolve this problem, either cast the integer to unsigned if you know it can never be less than zero or cast the unsigned to an integer if you know it can never exceed the maximum integer. 574 Signed-unsigned mix with relational -- The four relational operators are: > >= < <= One of the operands to a relational operator was signed and the other unsigned; also, the signed quantity could be negative. For example: if( u > n ) ... where u is unsigned and n is signed will elicit this message whereas: if( u > 12 ) ... will not (even though 12 is officially an int it is obvious that it is not negative). It is not a good idea to mix unsigned quantities with signed quantities in any case (a 737 will also be issued) but, with the four relationals, a negative value can produce obscure results. For example, if the conditional: if( n < 0 ) ... is true then the similar appearing: u = 0; if( n < u ) ... is false because the promotion to unsigned makes n very large. To resolve this problem, either cast the integer to unsigned if you know it can never be less than zero or cast the unsigned to an int if you know it can never exceed the maximum int. 575 enumeration constant exceeds range for integers -- For many compilers the value of an enumeration constant is limited to those values that can fit within a signed or unsigned int. 577 Mixed memory model (option 'String') -- The indicated option requested a change to the memory model after part or all of another module was processed. The memory model option should be specified before any module is processed. The most common cause of this error is specifying the memory model after having specified the standard library. This would be a natural error to make if the standard library file were specified via a LINT environment variable. 578 Declaration of symbol 'Symbol' hides symbol 'Symbol' (Location) -- A local symbol has the identical name as a global symbol ( or possibly another local symbol). This could be dangerous. Was this deliberate? It is usually best to rename the local symbol. 579 parameter preceding ellipsis has invalid type -- When an ellipsis is used, the type preceding the ellipsis should not be a type that would undergo a default promotion such as char, short or float. The reason is that many compilers' variable argument schemes (using stdarg.h) will break down. 580 Redeclaration of function 'Symbol' (hiding Location) causes loss of prototype -- A declaration of a function within a block hides a declaration in an outer scope in such a way that the inner declaration has no prototype and the outer declaration does. A common misconception is that the resulting declaration is a composite of both declarations but this is only the case when the declarations are in the same scope not within nested scopes. If you don't care about prototypes you may suppress this message. You will still receive other type-difference warnings. 581 Option 'String' is obsolete and should no longer be used -- This message is issued whenever we encounter an option that appears to do more harm than good. 'String' is the option in question. 582 esym (or emacro) name 'String' should not contain '(' -- The name provided to esym should not contain a (. For example, to suppress message 534 when calling f(int) use the option -esym(534,f) even if f is overloaded. 583 Comparing type 'Type' with EOF -- The message is issued when some form of character is compared against the EOF macro. EOF is normally defined to be -1. For example: while( (ch = getchar()) != EOF ) ... If ch is defined to be an int all is well. If however it is defined to be some form of char, then trouble might ensue. If ch is an unsigned char then it can never equal EOF. If ch is a signed char then you could get a premature termination because some data character happened to be all ones. Note that getchar returns an int. The reason it returns an int and not a char is because it must be capable of returning 257 different values (256 different characters plus EOF, assuming an 8-bit character). Once this value is assigned to a char only 256 values are then possible -- a clear loss of information. 584 Trigraph sequence (??Char) detected -- This message is issued whenever a trigraph sequence is detected and the trigraph processing has been turned off (with a -ftg). If this is not within a string (or character) constant then the trigraph sequence is ignored. This is useful if your compiler does not process trigraph sequences and you want linting to mirror compilation. Outside of a string we issue the Warning but we do translate the sequence since it cannot make syntactic sense in its raw state. 585 The sequence (??Char) is not a valid Trigraph sequence -- This warning is issued whenever a pair of '?' characters is seen within a string (or character) constant but that pair is not followed by a character which would make the triple a valid Trigraph sequence. Did the programmer intend this to be a Trigraph sequence and merely err? Even if no Trigraph were intended, it can easily be mistaken by the reader of the code to be a Trigraph. Moreover, what assurances do we have that in the future the invalid Trigraph might not become a valid Trigraph and change the meaning of the string? To protect yourself from such an event you may place a backslash between the '?' characters. Alternatively you may use concatenation of string constants. For example: pattern = "(???) ???-????"; // warning 585 pattern = "(?\?\?) ?\?\?-?\?\?\?"; // no warning #define Q "?" pattern = "(" Q Q Q ") " Q Q Q "-" Q Q Q Q // no warning 586 String 'Name' is deprecated. String -- The Name has been deprecated by some use of the deprecate option. See Section 2.5.7 -deprecate. The first String is one of the allowed categories of deprecation. The trailing String is part of the deprecate option and should explain why the facility has been deprecated. 587 Predicate 'String' can be pre-determined and always evaluates to String -- The predicate, identified by the first String, (one of greater than, greater than or equal, less than, less than or equal, equal, or not equal), cannot possibly be other than what is indicated by the second String parameter. For example: unsigned u; ... if( (u & 0x10) == 0x11 ) ... would be greeted with the message that '==' always evaluates to 'False'. 588 Predicate 'String' will always evaluate to String unless an overflow occurs -- The predicate, identified by the first String, cannot possibly be other than what is indicated by the second String parameter unless an overflow occurred. For example: unsigned u; ... if( (u + 2) != 1 ) ... would be greeted with the message that '!=' always evaluates to 'True'. See also Message 587. 589 Predicate 'String' will always evaluate to String assuming standard division semantics -- The predicate, identified by the first String parameter, cannot possibly be other than what is indicated by the second String parameter assuming standard signed integer division semantics. For example: int n; ... if( n % 2 == 2 ) ... would be greeted with the message that '==' always evaluates to 'False'. By standard integer division semantics we mean truncation toward zero so that, for example, -1/4 has quotient 0 and remainder -1 and not quotient -1, remainder 3. Although the current C standard [4] has endorsed this, the current C++ standard [10] regards integer division involving negative numbers to be 'implementation defined'. See also Message 587. 590 Predicate 'String' will always evaluate to String assuming standard shift semantics -- The predicate, identified by the first String parameter, cannot possibly be other than what is indicated by the second String parameter assuming standard signed integer shift semantics. For example: int n; ... if( (5 << n) < 0 ) ... would be greeted with the message that the less-than always evaluates to 'False'. It is true that if you shift 5 left by 29 bits (or 31 bits) you will have in many cases (probably most cases) a negative number but this is not guaranteed. According to the C Standard [4], shifting a positive signed integer (5 in this case) left by a number of places (n in this case) is only valid if the type can accomodate 5*(2**n). This is equivalent to not shifting a one into or through the sign bit. As another example: int n; ... if( (n >> 5) >= 0 ) ... would always be regarded as true. This is because shifting a negative number to the right yields results that are implementation defined. See also Message 587. 591 Variable 'Symbol' depends on the order of evaluation; it is used/modified through function 'Symbol' via calls: String" -- The indicated variable (given by the first Symbol) was involved in an expression that contained a call of a function (given by the second Symbol) that would use or modify the variable. Further, the order of evaluation of the two is not determinable. For example: extern int n; void f() { n++; } int g() { f(); return 1; } int h() { return n + g(); } // Warning 591 The above code, on the second pass, will elicit the following warning: Warning 591: Variable 'n' depends on the order of evaluation; it is modified through function 'g(void)' via calls: g() => f() If the function g() is called and then n is added, you will obtain a different result than if n were first evaluated and then the call made. The programmer should generally rewrite these expressions so that the compiler is constrained to use the intended order. For example if the programmer wanted to use the n prior to the call on g() it can alter h() to the following: int h() { int k = n; return k + g(); } This analysis requires two passes; the first pass builds the necessary call trees. 592 Non-literal format specifier used without arguments -- A printf/scanf style function received a non-literal format specifier without trailing arguments. For example: char msg[100]; ... printf( msg ); This can easily be rewritten to the relatively safe: char msg[100]; ... printf( "%s", msg ); The danger lies in the fact that msg can contain hidden format codes. If msg is read from user input, then in the first example, a naive user could cause a glitch or a crash and a malicious user might exploit this to undermine system security. Since the unsafe form can easily be transformed into the safe form the latter should always be used. 593 Custodial pointer 'Symbol' (Location) possibly not freed or returned -- This is the 'possible' version of message 429. A pointer of auto storage class was allocated storage and not all paths leading to a return statement or to the end of the function contained either a free or a return of the pointer. Hence there is a potential memory leak. For example: void f( int n ) { int *p = new int; if( n ) delete p; } // message 593 In this example an allocation is made and, if n is 0, no delete will have been made. Please see message 429 for