Добавил:
Upload Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:
Скачиваний:
35
Добавлен:
17.04.2015
Размер:
59.74 Кб
Скачать
/*********************************************************************/
TDUMP.EXE Utility Program

This file contains information on:

1. TDUMP: The file inspecting utility
2. Understanding "Undefined Symbol" Error Messages
3. Resolving Undefined Symbol linker messages
4. Borland Open Architecture: Name Mangling


===================================================================
1. TDUMP: The file inspecting utility
===================================================================
TDUMP.EXE is a utility that you use to examine the structure and
contents of files.

TDUMP organizes the output display according to the extension of the
file you're dumping. If the file extension is recognizable, TDUMP
displays the file's components according to the file type. TDUMP
recognizes many file extensions, including .EXE, .OBJ, and .LIB files.
If TDUMP doesn't recognize an extension, it produces a hexadecimal dump
of the file. You can control the output format by using command-line
options when you start the TDUMP (these command-line options are
described later).


TDUMP syntax
------------
The syntax for TDUMP is:

TDUMP [<options>] <Inputfile> [<Listfile>] [<options>]

o <options> stands for any of the TDUMP options discussed in the
next section. For a list of the available command-line options,
type TDUMP, then press <Enter> at the DOS prompt.

o <Inputfile> is the file whose structure you want to display (or
"dump").

o <Listfile> is an optional output file name. Note that you can
also use the standard DOS redirection command ">" to create an
output file.


TDUMP command-line options
--------------------------
You can use several option switches at a time with TDUMP. Because you can start
each option with either a hyphen or a forward slash, the following two commands
are equivalent:

TDUMP -el -v demo.exe

TDUMP /el /v demo.exe


The -a and -a7 options
- - - - - - - - - - - -
TDUMP automatically adjusts its output display according to the file extension
type. However, you can force an ASCII file display using either the -a or -a7
options:

o -a produces an ASCII file display that shows the offset and the
contents using the displayable ASCII characters. TDUMP displays
non-displayable characters (like control characters) as a period.

o -a7 converts high-ASCII characters to their low-ASCII
equivalents. This is useful if the file you are dumping sets
high-ASCII characters as flags (WordStar files do this).


The -b# option
- - - - - - - -
The -b# option allows you to display information beginning at a specified offset.
For example, if you wanted a dump of MYFILE starting from offset 100, you would
use the following command:

TDUMP -b100 MYFILE


The -d option
- - - - - - -
The -d option causes TDUMP to dump any Borland 32-bit debug information found in
the .OBJ file. If you do not specify this option, TDUMP displays raw data only.


The -e, -el, -er and -ex options
- - - - - - - - - - - - - - - - -
All four options force TDUMP to display the file as an executable (.EXE) file.

An .EXE file display consists of information contained within a file that is used
by the operating system when loading a file. If symbolic debugging information is
present (Turbo Debugger or MicrosoftCodeView), TDUMP displays it.

TDUMP displays information for DOS executable files, NEW style executable files
(Microsoft Windows and OS/2 .EXEs and DLLs), and Linear Executable files.

o -el suppresses line numbers in the display.

o -er suppresses the display of the relocation table.

o -ex prevents the display of New style executable information.
This means TDUMP will only display information for the DOS
"stub" program.


The -h option
- - - - - - - -
The -h option displays the dump file in hexadecimal (hex) format. Hex format
consists of a column of offset numbers, 16 columns of hex numbers, and their ASCII
equivalents (a period appears where no displayable ASCII character occurs).

If TDUMP doesn't recognize the input file's extension, it displays the file in hex
format (unless an option is used to indicate another format).


The -l and -li options
- - - - - - - - - - - -
The -l option displays the output file in library (.LIB) file format. A library
file is a collection of object files (see the -o option for more on object files).
The library file dump displays library-specific information, object files, and
records in the object file.

The -li option tells TDUMP to display a short form of "impdef" records when
dumping import libraries. You can also specify a search string using the following
syntax:
-li=<string>

For example, the command
TDUMP -li=codeptr import.lib
results in the following output:
Impdef:(ord) KERNAL.0336=ISBADCODEPTR
This output shows that the function is exported by ordinal, whose ordinal value is
336 (decimal). In addition, the output displays the module and function name.

If you give the command
TDUMP -li=walk import32.lib
TDUMP displays:
Impdef:(name) KERNEL32.????=HEAPWALK
This shows the output of a function exported by name.


The -m option
- - - - - - -
The -m option leaves C++ names occurring in object files, executable files, and
Turbo Debugger symbolic information files in "mangled" format. This option is
helpful in determining how the C++ compiler "mangles" a given function name and
its arguments.


The -o, -oc, -oi, and -ox options
---------------------------------
The -o option displays the file as an object (.OBJ) file. An object file display
contains descriptions of the command records that pass commands and data to the
linker, telling it how to create an .EXE file.

The display format shows each record and its associated data on a record-by-record
basis.

o -oc causes TDUMP to perform a cyclic redundancy test (CRC) on
each record encountered. The display differs from the -o display
only if an erroneous CRC check is encountered (the TDUMP CRC
value differs from the record's CRC byte).

o -oi<id> includes only specified record types in the object module
dump. Replace <id> with the name of the record to be displayed.
For instance,

TDUMP -oiPUBDEF MYMODULE.OBJ

produces an object module display for MYMODULE.OBJ that displays
only the PUBDEF records.

o -ox<id> excludes designated record types from the object module
dump. Replace <id> with the record name not to be displayed. For
instance,

TDUMP -oxPUBDEF MYMODULE.OBJ

produces an object module display for MYMODULE.OBJ that excludes
the PUBDEF records.

o The -ox and -oi options are helpful in finding errors that occur
during linking. By examining the spelling and case of EXTDEF and
PUBDEF symbols, you can resolve many linking problems. For
instance, if you receive an "unresolved external" message from
the linker, use "TDUMP -oiEXTDEF" to display the external
definitions occurring in the module causing the error. Then, use
"TDUMP -oiPUBDEF" on the module containing the public symbol
which the linker could not match.

Another use for the -oi switch is to check the names and sizes of the segments
generated in a particular module. For instance,

TDUMP -oiSEGDEF MYMODULE.OBJ

displays the names, attributes, and sizes of all of the segments in MYMODULE.

Note: To get a list of record types for -0i and -ox, use the command-line
options -oi? and -ox?.


The -R option
- - - - - - -
The -R option causes TDUMP to dump relocation tables from 32-bit PE (Win32) format
images. The default is to suppress these dumps.


The -v option
- - - - - - -
The -v option is used to create a verbose display. If you use -v with an .OBJ or
.LIB file, TDUMP produces a hexadecimal dump of the record's contents without any
comments about the records.

If you use TDUMP on a Turbo Debugger symbol table, it displays the information
tables in the order in which it encounters them. TDUMP doesn't combine information
from several tables to give a more meaningful display on a per-module basis.


===================================================================
2. Understanding "Undefined Symbol" Error Messages
===================================================================
One of the most common error messages seen by developers using a C or
C++ compiler is "undefined symbol." This document provides a general
description of what causes undefined symbol error messages, as well as
instructions on solving specific undefined symbol errors.

The following error message are treated in order:
UNDEFINED SYMBOL AT COMPILE TIME
UNDEFINED SYMBOL AT LINK TIME
OTHER UNDEFINED SYMBOL ERRORS
o UNDEFINED SYMBOL WHEN LINKING A BORLAND EXAMPLE
o UNDEFINED SYMBOL WHEN TLINKING FROM DOS COMMAND LINE
o UNDEFINED SYMBOL LINKING C/C++ AND ASSEMBLY MODULES
o UNDEFINED SYMBOL LINKING C++ WITH C OR ASSEMBLY MODULES
o UNDEFINED SYMBOL: '_main' IN MODULE C0.ASM
o UNDEFINED SYMBOL LINKING A DLL
o UNDEFINED SYMBOL: A PSEUDO REGISTER (ie. _AX)
o UNDEFINED SYMBOL: 'FIWRQQ'
o UNDEFINED SYMBOL: AN IOSTREAM CLASS MEMBER
o UNDEFINED SYMBOL: 'abort()'
o UNDEFINED SYMBOL: '_exitclean()'
o UNDEFINED SYMBOL: LLSH or SCOPY or FMUL or FDIV
o UNDEFINED SYMBOL: STATIC POINTER TO A CLASS MEMBER FUNCTION
o UNDEFINED SYMBOL: '_WSPRINTF'
o UNDEFINED SYMBOL: 'fidrqq'
o UNDEFINED SYMBOL IN WINDOWS.H
o UNDEFINED SYMBOL USING TCLASDLL.LIB
o UNDEFINED SYMBOL USING SELECTORS PROVIDED BY WINDOWS
o UNDEFINED SYMBOL: 'ChangeSelector'
o UNDEFINED SYMBOLS USING THE OBJECTWINDOWS LIBRARY (OWL)
o UNDEFINED SYMBOL: 'Object::new(unsigned int)'
GETTING A CLOSER LOOK
o USING TDUMP TO RESOLVE UNDEFINED SYMBOLS
o USING IMPDEF TO RESOLVE UNDEFINED SYMBOLS IN A DLL


UNDEFINED SYMBOL AT COMPILE TIME
--------------------------------
An undefined symbol at compile time indicates that the named
identifier was used in the named source file, but had no definition in
the source file. This is usually caused by a misspelled identifier
name, or missing declaration of the identifier used.

EXAMPLE 1:
int main(void)
{
test = 1;
return 0;
}

The code shown in Example 1 generates an undefined symbol error if the
variable "test" was not declared in either an included header file or
in the actual source file itself.

EXAMPLE 2:
int main(void)
{
int test;
Test = 1;
return 0;
}

The code shown in Example 2 causes an undefined symbol error message
to be displayed because the variable "Test" was not spelled as it was
declared. Remember that C and C++ are case-sensitive languages.


UNDEFINED SYMBOL AT LINK TIME
-----------------------------
When linking multi-file projects, the linker must resolve all
references to functions and global variables shared between modules.
When these references cannot be resolved, the linker generates
an "undefined symbol" error message. This means that after searching
all of the object files and libraries which are included in the link,
the linker was unable to find a declaration for an identifier you
used. This can be caused by:

o Forgetting to include a needed object module or library in
your link (project file, response file, or command line).

o Misspelling the name of the undefined symbol either where it
was used or where it was declared.

o Accidentally declaring a global variable as "extern."

o Forgetting to use extern "C" to disable name mangling when you're
mixing C++ with C or Assembly modules. See the specific entry on
this subject elsewhere in this document or consult the
HELPME!.DOC file included with the product.

o Turning OFF Generate Underbars in one of the modules you're
linking.

If all else fails, use TDUMP to dump both object modules and note any
difference between the symbols used. This will usually give enough
insight to resolve the problem. For more information on using TDUMP to
resolve undefined symbol errors, see the "Getting a Closer
Look "section in this document.


OTHER UNDEFINED SYMBOL ERRORS
------------------------------
The following list provides solutions to some of the more common
causes of undefined symbol errors:

o UNDEFINED SYMBOL WHEN LINKING A BORLAND EXAMPLE
Almost all of Borland's examples come with project files. You
must use the project file when building the example to ensure
that all necessary modules are linked and all the necessary
settings are defined.

o UNDEFINED SYMBOL WHEN TLINKING FROM DOS COMMAND LINE
The TLINK command line must have the libraries in the following
order ( GRAPHICS.LIB + <user libs> + EMU.LIB + MATH (S,T,C,M,L -
for model) + C (S,T,C,M,L - for model)

o UNDEFINED SYMBOL LINKING C/C++ AND ASSEMBLY MODULES
There are several sources of undefined symbol errors when trying
to link assembly with C or C++ modules:

o Turbo Assembler generates all upper-case symbols unless you
specify /ml or /mx on the assembly command line. Since C
modules are, by default, case sensitive, failing to do this
will result in undefined symbols for all symbols that are
not completely upper case in the C module(s).

o The symbols in the assembly file being referenced from a C
module must be declared using a PUBLIC directive. TLINK does
not consider symbols which are not declared PUBLIC when it
attempts to resolve an undefined symbol.

o All symbols in the assembly module that are referenced in
the C module must be prototyped/declared as extern in the C
module -- the compiler generates undefined symbol errors for
all symbols not declared in this manner. In addition, all
symbols in the assembly module that are referenced from a C
module must have an underscore prefix. This naming
convention must be used in the assembly module. You can
either do this explicitly (_symbol) or you can use:

.model <memory model used>, C

to specify this implicitly for all symbols.

IMPORTANT NOTE: If you put underscores in front of your
assembly routines and also use the .model (memory model)
C directive, the public symbol will be generated with two
underscores; consequently an undefined symbol error is
generated.

If all else fails, TDUMP both object modules and note any
difference between symbols. This will usually give enough
insight to resolve the problem. For more information on
using TDUMP to resolve undefined symbol errors see the
"Getting a Closer Look" section in this document.

o UNDEFINED SYMBOL LINKING C++ WITH C OR ASSEMBLY MODULES
C++ is a strongly typed language. In order to support type-safe
linkage (as well as function overloading), Borland C++ must
attach information to the symbols generated for function names
and variables. When this is done, the symbol will no longer match
the standard C style function name. In order to link correctly
with C or assembly modules, the compiler must be notified that
the symbol is to be in the standard C style (non-encoded) rather
than employing C++ name-mangling (encoded). This is done by
prototyping the function as type extern "C". Here is a quick
example:

extern "C" int normal_c_func( float, int, char );

For an additional example, you may want to look at the header
files which came with the product. One such header file is
stdio.h.

o UNDEFINED SYMBOL: '_main' IN MODULE C0.ASM
Every DOS C program must contain a function called main(). This
is the first function executed in your program. The function name
must be all in lower case. If your program does not have one,
create one. If you are using multiple source files, the file that
contains the function main() must be one of the files listed in
the project.

Note that an underscore character '_' is prepended to all
external Borland C++ symbols.

In addition to an absent, misspelled or mis-cased symbol main,
there are two additional common causes:

o The "generate underbars" option is disabled.

o The Pascal calling Convention rather than the C calling
convention is selected.


o UNDEFINED SYMBOL LINKING A DLL
It is relatively simple to link a DLL to your source:

1) Create a .LIB from the .DLL using Borland's implib utility.

2) Include the .LIB in your project if you're using the IDE or
in your TLINK command if you're using the command-line
compiler or linker.

3) Turn case sensitive link ON.

4) Turn case sensitive exports ON.

The issues of linking C++ with C, assembly, or any other language
still apply. See the sections on linking C++, C, and assembly in
this document.

If the link still fails, the techniques in the section "Getting A
Closer Look" should help you resolve the problem.

o UNDEFINED SYMBOL: A PSEUDO REGISTER (i.e. _AX)
Pseudo registers are only allowed in the Borland C++ and ANSI
modes of the compiler. You can change this setting in the
Options | Compiler | Source menu.

o UNDEFINED SYMBOL: 'FIWRQQ'
Your program uses floating point routines directly (or
indirectly) and you have NONE selected for floating point. Or,
you are using TLINK and have forgotten to include EMU.LIB or
FP87.LIB on the command line.

o UNDEFINED SYMBOL: AN IOSTREAM CLASS MEMBER
If you are using the Integrated Development Environment, simply
turn off Options | Compiler | Code Generation | Unsigned
Characters.

If you are using the command-line compiler, simply remove the
'-K' option.

o UNDEFINED SYMBOL: 'abort()'
The sole purpose of abort is to print the error message

"Abnormal Program Termination"

and exit the program with an error code of 3. This function is
located in the startup code C0.ASM. Linker errors indicating that
abort() is an undefined symbol are only possible if the standard
startup code is not being linked into a project. Although this is
not a common C/C++ development practice, it is to be expected
when linking in code written other languages such Microsoft
Fortran, Clipper, or in cases where embedded systems are being
developed. To resolve the undefined symbol, extract the abort()
function from the startup code and make a separate object out of
it to be linked into the project.

o UNDEFINED SYMBOL: '_exitclean()'
There is a function called _exitclean which is new to Turbo C++.
Users moving from Turbo C 2.0 to Turbo C++ may encounter
_exitclean() as an undefined symbol at link time. _exitclean() is
defined in the Turbo C++ startup code. Users creating embedded
system (ROMable code), who do not use the standard Turbo C++
startup code, are likely to encounter _exitclean() as an
undefined symbol. These users can strip the function from the
C0.ASM file and create a separate .OBJ file which can be linked.
Another option would be to purchase the Borland C++ RTL source and make
the necessary adjustments.

o UNDEFINED SYMBOL: LLSH or SCOPY or FMUL or FDIV
The helper functions have changed their names from Turbo C 2.0 to
Turbo C++. This can lead to many undefined symbol issues.

When LLSH or SCOPY or FMUL or FDIV (note no underscores here)
appear as undefined symbols at link time, it is likely that an
object module or library has code generated a call to some helper
function from the Turbo C 2.0 libraries. The solution is to
simply recompile all objects from source. You can do this by
choosing Compile | BuildAll from the menu in the IDE.

o UNDEFINED SYMBOL: STATIC POINTER TO A CLASS MEMBER FUNCTION
Any static member of a class must be initialized--if not that
static member generates an undefined symbol error. The following
is an example of how to initialize a static pointer to class
member function of a class is initialized:

// When testing static member initialization, you must
// declare an instance of the class in a main function;
// otherwise, the linker has no reference which it must
// try to resolve, and the undefined symbol error will
// not be seen - thus you won't know that your
// initialization was in error.

#include <iostream.h>

// used to allow global initialization of static member pointer
typedef void (*fptr)();

// declare class containing static members
class First
{
public:
static fptr statptr;
};

// initialize static members of class First
fptr First::statptr = NULL;

int main(void) {
First fVar;

if (fVar.statptr == NULL)
cout << "fVar.statptr is NULL: " <<
fVar.statptr << endl;

return 0;
} // end of main()

o UNDEFINED SYMBOL: '_WSPRINTF'
Turn off the "Case-sensitive exports" and "Case-sensitive link"
options. If you are using the command linker, don't use the /c
switch. If you are invoking the linker from the BCC(x) command
line, use the -lc- switch. If you are using the IDE, go to the
linker options dialog box and turn off the case sensitivity
switch.

o UNDEFINED SYMBOL: 'fidrqq'
You will get an undefined symbol fidrqq when using the Integrated
Development Environment if you have the Options | Compiler | Code
Generation | More | Floating Point Option set to NONE and you are
using floating point arithmetic in your program. In order to best
solve this problem you must set the IDE option for Floating Point
to either the emulation choice or to the 80x87 choice. Note that
if you choose an 80x87 setting the application generated by the
compiler will require an 80x87 chip to be present at run-time.
Use this setting only when truly appropriate.

o UNDEFINED SYMBOL IN WINDOWS.H
Make sure you are using the windows.h file that came with Borland
C++, NOT the windows.h that came with the Microsoft Windows SDK.
If you include the Microsoft windows.h file you will get many
undefined symbols including WINMAIN in caps and translatemessage
in lower case. Use our windows.h file instead of Microsoft's when
you are using our compiler.

o UNDEFINED SYMBOL USING TCLASDLL.LIB
To use the DLL version of the container class library you must do
ALL of the following:

o Use the large memory model

o Use Smart Callbacks

o Turn case sensitive link ON

o Turn case sensitive exports ON

o Use the DLL version of the RTL

o Define _CLASSDLL

o UNDEFINED SYMBOL USING SELECTORS PROVIDED BY WINDOWS
If you are using _C000h or other selectors provided by Windows
and they are coming up as undefined symbols, perhaps you are
compiling in C++ mode and forgot to extern "C" them.

Programming in C:
extern WORD _C000h

Programming in C++:
extern "C" WORD _C000h

o UNDEFINED SYMBOL: 'ChangeSelector'
The Windows API function ChangeSelector() has the wrong name in
KERNEL.EXE for Windows 3.0, and therefore in IMPORT.LIB. The name
given to this function (and this is NOT a joke) is
PrestoChangoSelector().

Use PrestoChangoSelector() in your program in place of
ChangeSelector() and all will be well.

o UNDEFINED SYMBOLS USING THE OBJECTWINDOWS LIBRARY (OWL)
If you get any of the following the undefined symbols:

TApplication(unsigned char far *, unsigned int, unsigned int,
unsigned char far *, int )

and

TWindow::TWindow(TWindowsObject near *, unsigned char far *,
TModule near *)

and are using the DLL versions of OWL and the Class Library,
you must define _CLASSDLL in Options | Compiler | Code
Generation | defines combo box.

It could be because you have forced unsigned characters. The
functions in the .lib take signed characters and thus if you
compile to use unsigned characters, the symbols will not match.

Using the Integrated Development Environment be sure to turn
off Options | Compiler | Code Generation | Unsigned Characters.

If you are using the command line compiler be sure to remove
the -K option to solve this problem.

o UNDEFINED SYMBOL: 'Object::new(unsigned int)'
You forgot to link with the TCLASDLL.LIB file where it is
defined!

Basically the problem is that you are mixing both STATIC and
DYNAMIC LINK libraries into the application. You must use only
one or the other. If you are working in the IDE, change the
LINKER options section for libraries.

If you are using the dynamic link library, remember to set
_CLASSDLL and Build All.

If you are using the command line compiler and linker, just be
sure to specify the correct set of library files. For specific
information on the "correct set of library files" please see the
documentation included with the product as library names tend to
change from version to version.


GETTING A CLOSER LOOK
---------------------
Borland provides tools that you can use to determine exactly what the
linker is seeing when it is trying to match symbols: TDUMP and IMPDEF.
This section provides some simple techniques for using these utilities
to resolve undefined symbol errors.

o USING TDUMP TO RESOLVE UNDEFINED SYMBOLS
TDUMP can be used to list the symbols in a .OBJ or a static .LIB
that the linker is having trouble matching.

First, TDUMP the module that is trying to reference the symbol.
For example, if main.cpp is trying to access a function, myfunc()
in myfuncs.cpp and is getting "Undefined symbol myfunc() in
module main.cpp",

tdump -m -oiEXTDEF main.obj > main.ext

Then, TDUMP the module in which the symbol is defined.

tdump -m -oiPUBDEF myfuncs.obj > myfunc.pub

Using a text editor find the symbol associated with the error in
each file. If they are not the same, then you have verified that
the linker is correct in generating the error. You must check
your code and verify that the compiler is seeing the same
declaration for the symbol when each module is being compiled.

You can use TDUMP to look at a static .LIB file the same way you
look at an .OBJ. file.

tdump -m -oiPUBDEF mystatic.lib > mystatic.pub

To use TDUMP with an implib,

tdump -m -oiTHEADR mydll.lib > mydll.pub

You can also use IMPDEF to view the symbols exported by a DLL.


o USING IMPDEF TO RESOLVE UNDEFINED SYMBOLS IN A DLL
If you are trying to link a Borland generated DLL with another
language or product and are getting undefined symbol errors, you
should verify that the names or the ordinals that are being
exported by the DLL are the same as your application expects.

This can be done by generating a .DEF file with the utility
IMPDEF. For example, to create a .DEF file for MY.DLL,

impdef my.def my.dll

The .DEF file will have the exported symbol name and its ordinal
number. Remember that C++ mangles names. Your application must
expect the mangled symbol name to call the function in the DLL
properly. This can be a problem if the application automatically
uppercases the symbol name before trying to call the function. If
this is so, you must change the declaration of the function and
re-build your DLL.


===================================================================
3. Resolving Undefined Symbol linker messages
===================================================================
This section provides an overview of the Linking process and helps to
identify causes of 'unresolved external symbols'.

The code for printf() is in a module in the run time library. When you
call printf() in a C/C++ module, the compiler creates a record
(referred to as EXTDEF - EXTernal DEFinition) that indicates the call
to an external function. The linker then looks at that OBJ, along with
all the other modules and libraries specified and attempts to find
another module (.OBJ or .LIB) which defines/provides the
symbolprintf().

If the linker cannot resolve the call to printf(), the linker
generates an error indicating that printf() is an undefined symbol.
The error message, however, is very often not the result of leaving
out the module containing the symbol being looked for, but rather a
discrepancy between the name used by the caller (the C/C++ module
calling printf() in the case mentioned above) and the supplier (the
LIBRARY containing the code to printf() ).

The *real* name of any symbol is almost always different from the
name/identifier used by the programmer. For example, the *real* name
(by *real* name we mean the identifier used/generated by the tools) of
strcpy() is: '_strcpy()'. The *real* name of a symbol depends on the
various settings and options. The relevant settings are list below:

Calling Conventions:
> cdecl
> pascal
> fastcall

Compiler Settings:
> generate underbars
> unsigned chars ( C++ only )

Optimizations:
> Object Data Calling ( C++ only )

Virtual Table:
> Far Virtual Tables

Language used:
> C
> C++
> Assembly

Furthermore there are two options which affect how the linker attempts
to match symbols:

> Case sensitive link
> Case sensitive exports ( Windows only )

The following is a discussion of how the above mentioned options
affect the *real* name of symbols, hence the resolution of symbols.


Calling Conventions
-------------------
Borland and Turbo C++ both allow you to specify the default calling
convention. This default can be overridden by using the
'pascal', '_fastcall' or 'cdecl' keywords. Whether set globally or on
individual function instances, the calling convention affects the name
of functions. By default, when the compiler encounters a function
declared as,

int Foo( int ); // or: int cdecl Foo( int );

the name generated is _Foo; that is, the resulting name has the same
case but is preceded with a leading underscore. The generation of the
underscore is the default behavior and is necessary when you link to
the run time libraries. There is no 'printf()' symbol in the RTL
(!),but there is a '_printf()'.

While the C calling convention implies 'Case Sensitivity' and
'Generation of Underbars', Borland/Turbo C++ provides separate
settings for the generation of underbars and the calling convention
Generation of Underbars can be controlled from the Options | Compiler|
Advanced Code Generation Dialog, or the -u option from the command
line ( -u- would turn it off, it is on by default). The 'Calling
Convention' can be modified via the Options | Compiler |
Entry/ExitCode Dialog.

If our function 'Foo' is declared with the pascal modifier, for
example:

int pascal Foo( int );

(or if the default 'Calling Convention' is set to 'PASCAL') the
resulting name will be FOO--that is, all upper-case with no
underscore. The '_fastcall' modifier is similar to 'cdecl' in regards
to Case Sensitivity but the underscore character is replaced with '@'.
Hence:

int _fastcall Foo( int );

will result in the '@Foo' symbol.

Therefore, mismatching the calling conventions might result in
'Undefined Symbols.' Watch for clues in the undefined symbol name
provided in the Linker error messages (look at the Case Sensitivity
and any leading characters) to spot cases of incorrect settings in the
'Calling Convention' and/or 'Generation of Underbars'.


NAME MANGLING:
--------------
The C++ language uses yet another naming convention as part of its
implementation of 'type safe linkage.' Imagine a function myfunc()
which take two longs [void myfunc( long, long );]. What if someone has
it incorrectly prototyped in a calling module as taking two floats,
for example void myfunc( float, float );. The results of such a call
will be unpredictable. When using the C language, the linker would
resolve such a call since the symbol the compiler uses to call the
function taking two floats will be '_myfunc()', and the name the
compiler used in the module which implements the function taking two
longs is also '_myfunc()'.

In C++, however, the name the compiler generates for a function is a
'mangled' name: it is 'encrypted' based on the parameters types the
function expects. In the scenario described in the prior paragraph, the
call to myfunc() will not be resolved since the compiler generates
different names for 'void myfunc( float, float )' and 'void myfunc(
long, long )'.

Because a C++ function's mangled real name depends on the types of its
parameters, if unsigned chars is used as a compiler option, it changes
the name of functions declared to take a char, or char *. Unsigned
chars is off by default, it is turned on under the Options |Compiler |
Code generation menu. Or by specifying the -K option with the command
line compiler. Watch out for potential 'Undefined Symbol' messages
caused by a mismatched of char vs. unsigned char.

The 'virtual mechanism' of C++ is implemented via a table commonly
referred to as the Virtual Table or the VMT (Virtual Method Table).
Various settings of the Compiler dictate whether the Table ends up in
the Default Data Segment or in a Far Segment (namely Memory
Model, '_export' and 'huge' class modifiers, Virtual Table Control
Optionsetc). To further enforce 'type- safe-linkage', the
Borland/Turbo C++ compilers include the 'distance' of the Virtual Table
as part of its 'Name-Mangling' logic. This prevents the linker from
resolving function calls which would crash at run-time because of
mismatched 'Virtual Table Control' settings. In the same token,
Borland provides the 'Object Data Calling convention' for improved
efficiency of C++ code. Once again, the 'Name-mangling' algorithm also
reflects the enabling of 'Object Data Calling'. This ensures that
function calls involving mismatched 'Object Data Calling' convention
between caller and callee will be caught at link time (instead of
resulting in erratic run-time behavior).

To illustrate the effect of 'Virtual Table Control' and 'Object
DataCalling,' let's create a simple class and look at the effects of
the various settings on the resulting names:

class Test
{
public:
virtual int Process( void );
};

int main( void )
{
Test t;
return t.Process();
}

The following table illustrates the effects of Compiler Settings on
the *actual* name generated for the member function 'int
Test::Process(void)'.

+----------------------------------------------------------+
| Object Call. | Far V. Tbl. | Huge Md. | [ REAL NAME ] |
|--------------+-------------+-----------------------------+
| No | No | No > @Test@Process$qv |
|--------------+-------------+-----------------------------+
| No | Yes | No > @Test@0Process$qv |
|--------------+-------------+-----------------------------+
| Yes | No | No > @Test@1Process$qv |
|--------------+-------------+-----------------------------+
| Yes | No | Yes > @Test@2Process$qv |
+--------------+-------------+-----------------------------+

NOTE: Using the '_export' or 'huge' keyword when defining a class
results in Far Virtual Tables for the class.

'Undefined Symbol Messages' caused by mismatching Virtual Table
Controls or Object Data Calling conventions may be hard to identify; it
is often useful to use TDUMP.EXE to find the actual names of the
unresolved symbols (however, watch out of any '0', '1' or '2' following
the '@ClassName@' portion of the real names).


LANGUAGE USED
-------------
By default assemblers (including TASM) do not modify public
names--they merely convert symbols to upper case. With TASM, the /mx
option forces the assembler to treat public symbols with case
sensitivity. Without /mx, a call to _myfunc from an assembly module
looks like _MYFUNC to the linker (this causes undefined symbol errors
when linking C and assembly).

NOTE: TASM has an extension which causes the automatic generation of
underscores. See. .MODEL <model>, Language directives in the TASM
User's Guide.

As previously mentioned in the section about 'Name Mangling,' the
C++ language uses a different naming convention than does the C
language. This can result in undefined symbols when calling C from C++
(or vice-versa). C++ modules should use the 'extern "C"' syntax when
interfacing with C modules (see the Name Mangling section of
Programmer's Guide for the proper syntax).


LINKER SETTINGS
---------------
By default, the linker treats _myfunc and _MYFUNC as different
symbols. However, you can control whether the linker pays attention to
Case Sensitivity via the Options | Linker | Settings dialog (IDE), or
the /c option with TLINK (/c: Enables Case Sensitivity [default],
/c-turns the option off).

For example, if the option is disabled, a call to _myfunc could be
resolved to _MYFUNC.

When creating a Windows application, not only can you link to 'static'
modules (.OBJs or .LIBs which are a collection of .OBJs), but you can
also link to dynamic libraries where the resolution of the call is
completed by Windows at load time. Functions residing in DLLs and
called from an .EXE are said to be imported. Functions that are coded
in an .EXE or .DLL, and are called by either Windows, .EXEs, or .DLLs
are said to be exported.

Functions are imported in two ways: by listing them in the IMPORTS
section of the .DEF file, or by linking to an import library. Functions
can be exported by two methods: by using the _export keyword in the
source code or listing the functions in the EXPORTS section of the
.DEF file.

Suppose your application calls the symbol _myfunc which is in a
DLL. The linker can treat symbols coming from an import library, or
IMPORTS section of the .DEF file with or without case sensitivity,
(determined by the setting of case sensitive exports under the Options
| Linker | Settings Dialog or /C option on the TLINK command line). If
this setting is NOT enabled, then the Linker treats symbols in import
libs or IMPORTS sections as all uppercase. It then considers upper
case symbols during the link phase. At that point it is doing normal
linking using the setting of the case sensitive link option. If we are
importing both _myfunc and _MYFUNC without the /C option, the linker
can only resolve the call to _MYFUNC.

If you are calling _myfunc (a cdecl function) and are performing a
case sensitive link, but do not have case sensitivity on
EXPORTS, _myfunc will show up as undefined.

> Imported cdecl functions and C++ names will link when /c
> and /C are both enabled, or neither are enabled.

C++ names are always generated with lowercase letters. When importing
or exporting C++ names, it's recommended that you use both the /c
and/C options.

Now let's apply the above to some common scenarios and provide
possible diagnostics and suggestions:


PROBLEM:
All the functions in a 3rd party library are undefined!

SOLUTION:
3rd party libraries must be explicitly linked in. To explicitly
link to a 3rd party library from the IDE, open a project file and
insert the .LIB file into the project file. The project file also
needs to have all of your source code files listed in it. From
the command line, insert the .LIB on your command line to TLINK.


PROBLEM:
All the functions in the RTL are undefined!

SOLUTION:
You need to link in Cx.LIB, where x is the memory model. A
feature of the IDE in Turbo C++ and Borland C++ v2.x is that if
you put a .LIB in the project file which starts out as Cx where x
is a memory model, the new library overrides the normal run time
library, and the latter will not be linked in (for example, if
you're using a library named CSERVE.LIB). Rename any such
libraries, then the normal Cx.LIB will automatically be linked
in. (Borland C++ 4.x has a dialog for specifying which Run Time
Libraries should be linked in).


PROBLEM:
When mixing C and C++ modules (.c and .cpp source) symbols are
undefined.

SOLUTION:
Because of name mangling (see above) the symbol the linker sees
being called from a C++ module will not look like the symbol in
the Cmodule. To turn name mangling off when prototyping
functions:

// SOURCE.CPP

extern "C" {
int Cfunc1( void );
int Cfunc2( int );
}

NOTE: You can also disable name-mangling for functions written in
C++ and called from C.

A C++ compile will happen if the source code has a .CPP extension, or
Options | Compiler | C++ options use C++ compiler is set to always.


PROBLEM:
randomize and other macros are coming up as undefined symbols.

SOLUTION:
Turn keywords to Borland C++. Since some macros are not ANSI
compatible, the header files will not define them if compiled
with ANSI or UNIX keywords on.


PROBLEM:
min and max are coming up undefined.

SOLUTION:
These macros are only included in a C compile, and will not be
seen by the compiler if compiling in C++. In C, you must
#include <stdlib.h> to use them.


PROBLEM:
I cannot get my assembly modules to link with my C/C++ program.

SOLUTION:
For C++, see above. Otherwise, the .ASM must be assembled with
case sensitivity on public symbols (/mx for TASM). It must also
match the C naming convention, which will have an underscore in
front of the name. So given the following code in a C module,

int myfunc( void );

you need to

call _myfunc

from the assembly module. (NOTE: TASM has extensions which will
automatically generate underscores for you). Also, make sure the
.OBJ which has the assembly code is listed in the project file,
or on the tlink line.


PROBLEM:
wsprintf is coming up undefined.

SOLUTION:
In Borland C++ 2.0, to use wsprintf when case sensitive exports
is on, you need to reverse a define in windows.h via the
following:

#ifdef wsprintf
#undef wsprintf

#define wsprintf wsprintf
extern "C" int FAR cdecl wsprintf( LPSTR, LPSTR, ... );
#endif

To call wsprintf (or any cdecl imported function ) with case
sensitive exports off, you need to match an upper case name. Thus
windows.h #defines wsprintf to be WSPRINTF. wsprintf is one of the
cdecl functions from windows, so the compiler will generate a
lower case symbol for when calling it.


PROBLEM:
FIWRQQ and FIDRQQ are undefined

SOLUTION:
These symbols are in the EMU or FP87 library. You must link it
in explicitly when using TLINK, or set the IDE to link it in under
the Options | Compiler | Advanced Code Generation Floating point
box.


PROBLEM:
Warning attempt to export non-public symbol ...

SOLUTION:
The exports section of the .DEF file has a symbol which does not
match one the compiler generated in the source code. This happens
if:

o The source was compile in C++ (the symbol name is mangled).
Resolve by exporting with the _export keyword, compiling in C,
or by declaring the function as extern "C".

o Case sensitive exports is ON, you are exporting a PASCAL
function, and exporting it like: WndProc. Resolve by
exporting as WNDPROC or by turning case sensitive exports off.

o You are exporting a cdecl function. If declared as
int myfunc( void );
export as _myfunc and turn case sensitive exports on (or
just use the _export keyword).

NOTE: When using the '_export' keyword, it must be used in
the prototype of the function. For example:
int FAR _export myfunc( int );


PROBLEM:
C++ and DLL linking problems.

SOLUTION:
Classes declared in the DLL need to be declared as the following:

class _export A
{
...
};

When defined in the EXE, the same must be prototyped as:

class huge A
{
...
};
// see User's Guide for more information

Then link with /c and /C on (both case sensitive link and case
sensitive exports ENABLED) when building BOTH the .DLL and the
calling.EXE.


PROBLEM:
OWL and undefined symbols.

SOLUTION:
If you're linking to the static libraries:
- with BC 2.0, link in owlwx.lib, tclasswx.lib, and
sallocwx.lib. (You don't need sallocwx.lib with BC
v 3.x ).

- do NOT define _CLASSDLL in the code generation
dialog, or before including owl.h.

- link with /c and /C. (from IDE, under linker
options, case sensitive link and case sensitive
exports).

- Do NOT compile with -K or unsigned char's on. You
will get several undefined symbols in this case.

If you're linking to the OWL .DLL, DO define _CLASSDLL before
including OWL.H, and use /c and /C linker options (both Case
Sensitive Link and Case Sensitive Exports ENABLED).


PROBLEM:
With an OWL application, wsprintf is undefined in module when
linking to the static libraries.

SOLUTION:
Link with /C (case sensitive exports ENABLED).


PROBLEM:
_main is an undefined symbol.

SOLUTION:
main is the entry point for every DOS C/C++ program. Make sure
you write a function called main (all lowercase) in your program.
If you have a project file loaded, make sure your source code
file (.c or .cpp file) which has main in it is listed in the .prj
file. Make sure generate underbars is turned on.


PROBLEM:
iostream members, like << operator are undefined

SOLUTION:
Turn options | compiler | code generation | unsigned chars off
(do not use -K on the command line).


PROBLEM:
Getting undefined symbols LLSH, SCOPY FMUL, FDIV, etc

SOLUTION:
Many of the helper functions have changed names from Turbo C
2.0 to Borland C++. This functions were called from Turbo C 2.0.
The solution is to recompile any .OBJ or .LIB them with Borland
C++.


SNOOPING AT THE REAL NAMES:
---------------------------
An .OBJ file is a collection of records. When you call a function or
reference a symbol not defined in your module, the compiler generates
an external definition record in the OBJ. This external definition
record has the symbol which the linker must resolve. When you define a
function, or make storage for some data, the compiler generates a
public definition record for that module (unless you declared the item
as static, that makes it private to that module). One of the tasks of
the Linker is to match Public Definitions and External
Definitions(PUBDEFs and EXTDEFs).

To see the symbols the LINKER has to deal with, use TDUMP.EXE provided
with the Borland C++ package. For example:

tdump -m -oiEXTDEF some.obj

The above shows all the EXTDEF (external definition) records in an .OBJ
file. Be sure to add the -m option when coding in C++.

tdump -m -oiPUBDEF some.obj

The above will display all the PUBDEF (public definition) records.

Let's assume you've purchased a third party library and a symbol
provided the library is unresolved; the possible steps in
identifying the problem could include:

- Create a Listing of symbols in the Library using TLIB. For
example:
TLIB NEWLIB.LIB, NEWLIB.LST

- TDUMP the .OBJ file which was created from the .C/.CPP
module calling the desired function.
TDUMP -m -oiEXTDEF MYCODE.OBJ MYCODE.LST

- Attempt to find any discrepancies between the name in
NEWLIB.LST and the one in MYCODE.LST and ascertain that
the Library does indeed provide the desired function.

Windows Programmers will find the IMPDEF.EXE utility (in addition to
TDUMP and TLIB) a very useful tool to help identify unresolved symbols
when DLLs and/or Import Libraries are involved.


===================================================================
4. Borland Open Architecture: Name Mangling
===================================================================
There are four basic forms of encoded names in Borland C++:

1. @className@functionName$args

This encoding denotes a member function Name belonging to
class Name and having arguments args.

Class names are encoded directly. The following example shows a
className in an encoded name:

@className@...

The class name may be followed by a single digit; the digit value
contains the following bits (these can be combined):

0x01 : the class uses a far vtable

0x02 : the class uses the -po calling convention

0x04 : the class has an RTTI-compatible virtual table; this bit is
only used when encoding the name of the virtual table for
the class

The digit is encoded as an ASCII representation of the bit mask
value, with 1 subtracted (so that, for example, the class prefix for a
class 'myfunc' that uses far vtables would be '@myfunc@0').

See the next section on the encoding of function names and argument
types.

2. @functionName$args

This form of encoding denotes a function functionName with arguments
args.

3. @className@dataMember

This form of encoding denotes a static data member dataMember
belonging to class className. Names of classes and data members are
encoded directly. The following example shows a member myMember in
class myClass:

@myClass@myMember

4. @className@

This name denotes a virtual table for a class className. As mentioned
previously, class names are encoded directly.


Encoding of nested and template classes
---------------------------------------
The following form encodes a name of a class lexically nested within
another class:

@outer@inner@...

A template instance class encodes the name of the template class, along
with the actual template arguments, in the following way:

%templateName$arg1$arg2 ..... $argn%

Each actual argument starts with a letter, specifying the kind of
argument it is:

o t type argument

o i nontype integral argument

o g nontype nonmember pointer argument

o m nontype member pointer argument

The first letter is followed by the encoded type of the argument. For
a type argument, this code also represents the argument's actual
value. For other kinds of arguments, the type code is followed by $and
the argument value, encoded as an ASCII number or symbol name. An
instance of template<class T, int size> whose name is
vector<long,100>is encoded as shown in the following example:

%vector$tl$ii$100%


Encoding of function names
--------------------------
The encoded function Name might denote either a function name, a
function such as a function such as a constructor or destructor, an
overloaded operator, or a type conversion.


Ordinary functions
------------------
Ordinary function names are encoded directly, as shown in the
following examples:

foo(int) --> @foo$qi sna::foo(void) --> @sna@foo$qv

The string $qi denotes the integer argument of function myfunc();'$qv'
denotes no arguments in sna::myfunc.


Constructors, destructors, and overloaded operators
---------------------------------------------------
The following information covers argument encoding in more
detail. Constructors, destructors, and overloaded operators encoded
with a $bcharacter sequence, followed by a character sequence from the
following table:

Character Meaning Sequence
_________________________________

ctr constructor
dtr destructor
add +
adr &
and &
arow ->
arwm ->*
asg =
call ()
cmp ~
coma ,
dec --
dele delete
div /
eql ==
geq >=
gtr >
inc ++
ind *
land &&
lor ||
leq <=
lsh <<
lss <
mod %
mul *
neq !=
new new
not !
or |
rand &=
rdiv /=
rlsh <<=
rmin -=
rmod %=
rmul *=
ror |=
rplu +=
rrsh >>=
rsh >>
rxor ^=
sub -
subs []
xor ^
nwa new[]
dla delete []
__________________________________

The following examples show how arguments are encoded with character
sequences, add, ctr, and dtr from the previous table:

operator+(int) --> @$badd$qi plot::plot() --> @plot@$bctr$qv
plot::~plot() --> @plot@$bdtr$qv

The string $qv denotes no arguments in the plot constructor or
destructor.


Type conversions
----------------
Encoding of type conversions accomplished with the $o character
sequence, followed by distinguishing return type of the conversion as
part of function name. The return type follows the rules for argument
encoding, explained later. The lack of arguments in one version is
made explicit in the mangling by adding $qv the end of the encoded
string.

Example:

myfunc::operator int() --> @myfunc@$oi$qv myfunc::operator char *()
--> @myfunc@$opzc$qv

The i following $o in the first example denotes int; the pzc in the
second example denotes a near pointer to an unsigned char.


Encoding of arguments
----------------------
The number and combinations of function arguments make argument
encoding the most complex aspect of name mangling.

Argument lists for functions begin with the characters $q. Type
qualifiers are then encoded as shown in the following table:
_____________________________

Character Meaning Sequence
_____________________________

up huge
ur _seg
u unsigned
z signed
x const
w volatile
____________________________

Encoding of built-in types follows that for applicable type
qualifiers, in accordance with the following table:

______________________________

Character Meaning Sequence
______________________________

v void
c char
s short
i int
l long
f float
d double
g long double
e ...
______________________________

Encoding of non-built-in types follows that for applicable type
qualifiers, in accordance with the following table:

______________________________

Character Meaning Sequence
______________________________

<a digit> (an enumeration or class name)
p near *
r near &
m far &
n far *
a array
M member pointer (followed by class and
base type)
______________________________

The appearance of one or more digits indicates that an enumeration or
class name follows; the value of the digit(s) denotes the length of
the name, as shown in the following examples:

foo::myfunc(myClass near&) is mangled as @foo@myfunc$qr7myClass
foo::myfunc(anotherClass near&) is mangled as
@foo@myfunc$qr12anotherClass

A character x or w may appear after p, r, m, or n to denote a constant
or volatile type qualifier, respectively. The character q appearing
after one of these characters denotes a function with arguments the
follow in the encoded name, up to the appearance of a $ character, and
finally a return type is encoded. The following example show how these
encoding rules are applied:

@foo@myfunc$qpxzc is mangled as foo::myfunc(const char near*)
@func1$qxi is mangled as func1(const int) @foo@myfunc$qpqii$i is
mangled as foo:myfunc(int (near*)(int,int))

Array types are encoded as a, followed by a dimension encoded as an
ASCII decimal number and a $, and finally the element type, as shown
in the following example.

myfunc( int (*x)[20] ) is mangled as @myfunc$qpa20$i

Encoded arguments are concatenated in the order of appearance in the
function call. The character t followed by an ASCII character encodes
the arguments when a number of identical non-builtin types are
function arguments. The ASCII character, ranging from ASCII 31H -
39Hand 61H - 7FH (1 to 9 and a onward), denotes which argument type to
duplicate, as shown in the following example:

@plot@func1$qdddiiilllpzctata is unmangled to plot::func1(double,
double, double, int, int, int, long, long, long, char near*,
char near*, char near*)

The two duplicate ta character sequences at the end of the encoded
name denote the tenth argument, encoded as pzc.

/*************************************************************************/

Соседние файлы в папке DOC