dos_compilers/Borland Turbo Pascal v55/DOC/HELPME!.DOC
2024-07-02 06:49:04 -07:00

454 lines
18 KiB
Plaintext
Raw Permalink Blame History

This file contains invisible Unicode characters

This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

TURBO PASCAL 5.5: ANSWERS TO COMMON QUESTIONS
---------------------------------------------
1. Can I build programs bigger than 64K?
The total size of a program's code is only limited by the
memory you have available; but each unit (module) can be
no larger than 64K, since it has to have its own code
segment.
The data segment is still no more than 64K, but the heap
is unlimited just as in 3.0. In fact, we've rewritten the
heap manager to make it much more efficient. There's no
waste when allocating memory (in 3.0, all blocks were
rounded up to a factor of 8), and you can install a heap
error routine that gets called if an allocation request
fails. All in all, 5.5's heap manager is much faster than
version 3.0.
2. Can Turbo Pascal run on generic MS-DOS machines?
TPC.EXE will run on generic machines when you use the /Q
option. The System, Overlay, Dos, and Printer standard
units will operate correctly on MS-DOS generic machines.
Generated .EXE's are MS-DOS compatible as long as you
don't use the special PC units (such as Crt, Graph, and
Graph3).
3. Does Turbo Pascal 5.5 support large integers?
Yes, TP 5.5 has virtually every incarnation of 8-, 16-, and
32-bit integers: shortint, integer, longint, byte, and
word.
4. Will the toolboxes for 4.0 work with 5.5?
Yes, all 4.0 versions of the toolboxes will work with
Turbo Pascal 5.5. In a few cases, minor changes to
compiler directives are recommended. Refer to the Turbo
Pascal README file for more information.
5. Does Turbo Pascal version 5.5 support conditional
compilation?
Yes, Turbo 5.5 includes conditional compilation support.
You use {$DEFINE ...} and {$UNDEF ...} for symbols and
{$IFDEF ...}. Using the {$IFOPT ...} conditional
directive, you can even test the settings of compiler
directives like R-, N+, and others. For the command-line
compiler, you can define symbols with the /D directive. In
the integrated compiler, you can also define symbols via
the Options/Compiler/Conditional Defines menu command.
6. How much of the 64K in the data segment is actually
available to my program?
The amount of data segment used by the run-time library
depends on which standard units you use in your program.
Here is the data segment usage (in bytes) for each unit:
UNIT Data Size
---- ---------
System 664
Overlay 24
Crt 20
Dos 6
Printer 256
Graph 1070
Turbo3 256
Graph3 0
=========
2282
The total size of the data segment is 65,520 bytes. If you
used only the System unit, the amount of data segment
space left over would be
65520 - 664 = 64856 bytes
7. What is the largest global data structure you can
allocate?
The maximum size of a single variable that can be
allocated on the heap is 65,521 bytes.
8. How do I find out how much code and data were generated by
the compiler for a program or unit?
If you are using the integrated environment, build your
program or unit and then use the Get Info command in the
Compile menu. This will bring up a window of information
that includes the size of code and data.
If you are using the command-line compiler, the size of
generated code and data is displayed on the screen at the
end of compilation.
9. Are the .OBJ files generated by Turbo C and Turbo
Assembler compatible with 5.5?
You can write Turbo C or Turbo Assembler routines and link
the .OBJ files into your Turbo Pascal programs by using
{$L} compiler directives. Turbo Pascal 5.5 generates .TPU
(Turbo Pascal Unit) files, not .OBJ files. We've made that
decision for many reasons:
A. TP 5.5's .TPU files are smaller than .OBJ's, and they
contain symbolic information important to the support
of Pascal's strict type conventions (types, constants,
etc.).
B. .TPU files allow "smart linking" - elimination of
unused code and data on a procedure-by-procedure
basis.
C. .TPU's allow built-in project management through
version 5.5's Make and Build commands.
D. .TPU's allow faster compilation speeds (34,000 lines
per minute on a PS/2 Model 60).
10. Will the $L compiler directive work for compiler object files
other than assembler?
That depends on the language. TURBO requires all the code
in the .OBJ to be in *one* CODE segment, and all the data
to be in *one* DATA segment. With assembly language that's
easy, but it may not work with some high-level language
compilers. You can use Turbo C to generate .OBJ files for
use by Turbo Pascal programs. An example, CPASDEMO.PAS is
included on the distribution disks.
11. Does the built-in linker eliminate unused data?
Yes. Unused code AND data are stripped when you compile to
disk.
12. If two units use a third unit, does the third unit get
included twice in my program?
No. All your units are "linked" together when you compile
your program. Only one copy of each procedure and function
used is generated. There is NO duplication of run-time
code. In fact, Turbo Pascal 5.5 has "smart linking," which
eliminates any unused code and data from the final .EXE.
13. What happens if you attempt to link another unit in which the
compiler directives are set differently?
Compiler directives are local to the unit they are
declared in. Thus, the compiler directives in one unit, or
in the main program, have no effect on the directives set
in another unit.
14. Can I create my own .TPL file?
Yes, but Turbo Pascal will only use the TURBO.TPL library
file. If you want to add your own units to the TURBO.TPL
file, you can use the unit mover program (TPUMOVER.EXE).
For example, you might want a customized version of
TURBO.TPL for each of the programs you're developing. A
corresponding configuration file for Turbo Pascal would
specify a different Turbo directory and thus fetch the
appropriate .TPL file for each of your projects.
15. What rules should I follow when writing an interrupt
handler?
The following is a list of rules to keep in mind when
writing an interrupt handler:
A. Use GetIntVec and SetIntVec to install/uninstall
interrupt handlers
B. Use the interrupt directive
C. Be careful about reentrancy. Don't use any calls to
DOS or to Turbo Pascal's overlay or heap management
routines in your interrupt handler
D. Interrupt procedures and functions must use the far
call model (use the {$F+} option)
E. Be proficient with the BIOS and assembly language
before attempting to write an interrupt handler
F. Make sure your interrupt handler is not in an
overlaid unit.
G. Neither static nor virtual methods can be used as
interrupt handlers.
16. Does a procedure or function in a program have to be of a
near or far call model?
If you are using overlays or procedural variables, you
should probably turn {$F+} on for all units and the main
program (the extra overhead of always using far calls is
usually quite small).
Otherwise, Turbo Pascal automatically selects the correct
call model. A routine is always a near call model unless
1) It is declared in the interface section of a unit
2) You override the default call model by using the {$F+}
compiler option
3) It is an object method. Methods are always far calls.
You should also use the {$F+} option to override the
default call model if you are writing interrupt handlers,
error handlers, exit procedures, or procedures or
functions that will be called via a procedural variable.
17. How do I write reentrant code in Turbo Pascal?
If a routine follows these rules, it is reentrant:
A. All data is allocated on the stack.
B. The routine doesn't use any global variables.
C. The routine can be interrupted at any time without
affecting the execution of the routine.
D. The routine doesn't call any other routines that are
not reentrant (e.g., DOS I/O).
18. What is the best approach to taking advantage of the new IEEE
floating-point types?
The new IEEE floating-point types are available when you
compile your program with {$N+} and you have a math
coprocessor; they are also available if you don't have a
coprocessor, but specify {N+,E+}. The 8087 emulator has
greater precision, but is significantly slower than the
fast, 6-byte, software-only reals. When developing
programs that will be compiled and run on machines without
the 8087 coprocessor, consider the trade-offs of speed
(built-in reals) vs. precision (8087 hardware/emulation)
and make the appropriate choice.
19. What type is Comp? What is it useful for?
The Comp type is a cross between an integer and a real
type and is available when 8087 code is generated {$N+}.
If no math coprocessor is available, specify {$N+,E+} and
the emulator will support the Comp type.
The compiler treats it as a real type without an exponent.
Thus Comp is useful when you need to store extremely large
numbers but don't need a decimal point. For example, you
might use variables of type Comp to store amounts in cents
and divide the value of the variable by 100 to determine
what the value in dollars and cents would be.
20. How many significant digits do the 8087 floating-point types
provide?
Type Digits of precision
-------- -------------------
single 7-8
double 15-16
extended 19-20
comp 19-20
21. Are the intermediate results of real number expressions
stored in the 8087 registers?
No. The user (8086) stack is used to store intermediate
results of real number expressions.
22. How does rounding work with IEEE floating point?
The 8087 math coprocessor uses a different method for
rounding numbers than what you may be used to. In order to
achieve a more even distribution of values, the 8087 uses
a method sometimes called "Banker's Rounding." This method
dictates that a number will always be rounded to the
nearest EVEN number. Note that this is quite different
than always rounding UP. Here are a couple of examples:
Round(0.5) = 0
Round(1.5) = 2
23. How do you do I/O redirection?
If you want to do DOS I/O redirection when running an .EXE
file generated by Turbo Pascal, DON'T use the Crt unit. If
you do, make sure you assign a null file name to the
standard Output file handle:
Assign(Output, ''); { Assign a null file name Output }
ReWrite(Output); { Open the file for output }
Any Write statement that does not specify a file variable
will be redirected to the DOS standard output file.
24. How do you go about upgrading version 3.0 programs with
lots of chain files?
Chaining is not possible with .EXE files. Control can be
passed to another program by use of the EXEC procedure in
the DOS unit. You can also use 5.5's overlay manager to
build very large programs.
25. Are overlays supported in 5.5?
Yes! See the example program OVRDEMO.PAS and refer to the
Turbo Pascal manual for information on overlays.
26. Is there any support in Turbo Pascal 5.5 for file and record
locking?
There's a standard variable in the System unit called
FileMode, which you can use to assign an open mode for use
in all subsequent Resets. There are no record-locking
routines implemented in the standard version, but they are
easily implemented through MsDos calls.
27. Does Turbo 5.5 support procedural parameters?
Yes. See PROCVAR.PAS, DIRDEMO.PAS, and refer to the
Reference Guide for a complete description.
28. Can you use identifiers other than scalar in the case statement?
Case statements allow the following ordinal types:
Char, Boolean, ShortInt, Byte, Integer, Word, and
user-defined enumeration.
29. Is the run-time license policy the same as in version 3.0?
YES, there are no royalties!
30. C has static variables, is there anything similar in 5.5?
You can declare private global variables in the
implementation part of a unit. Such variables are only
visible within that unit. Like other globals, they retain
their values across calls.
Typed constant declarations declared within a procedure or
function also behave exactly like C's static variables.
They are local in scope but since they are allocated in
the data segment, they retain their values from call to
call.
31. What Turbo Pascal 3.0 code will cause the most problems
converting to version 5.5?
With our UPGRADE program (see appropriate Appendix in your
manual), it's not very difficult to port your code to 5.5.
It depends a lot on the type of programs you write.
The passing of parameters on the stack is done much more
efficiently now, so changes will have to be made to inline
machine code and assembly language. Most of the changes
are optional: using new types, breaking your program into
modules to take advantage of separate compilation. (The
UPGRADE program has a special option to help you "unitize"
your program too. It does all the "typing" for you.)
Some stricter type-checking is performed in version 5.5.
For example, the Dos unit now defines the often-seen
registers record type (AX, BX...); MsDos and Intr now take
this type. In this case, you can type-cast or change the
type identifier and recompile.
32. What books can I read that will help me with Turbo Pascal
5.5?
The Turbo Pascal Tutor is an excellent reference to Turbo
Pascal. Also, Osborne/McGraw Hill has a line of books
about Borland's products.
33. How do I use .BIN files provided by third-party vendors with
5.5?
We've included a utility on your distribution disk called
BINOBJ.EXE, which converts binary files into .OBJ files
that are linkable to your Turbo Pascal 5.5 programs. In
general this will only work if the binary files contain
data, not code. Contact your third-party vendor to see if
they also provide .OBJ versions of their programs.
34. Why does TURBO sometimes try to read from another drive
when I run it?
When you leave Turbo Pascal, it saves the name and path of
the file you were last editing in a pick list. The next
time you load Turbo, it checks this pick list and tries to
load the file you were last editing. If the file you were
last editing was in another drive, Turbo will try to read
from that drive. This also occurs if you have installed
another drive as your Turbo Directory.
35. Does Turbo Pascal 5.5 support EMS?
Yes, Turbo Pascal 5.5 will use up to 64K of EMS for
storing the edit buffer. In addition, you can instruct the
Overlay unit to place your overlaid units in EMS. Finally,
EMS.PAS on the distribution disk shows you how to access
EMS memory from a Turbo Pascal program.
36. How can I allocate my own I/O buffer for a text file?
You can use the procedure SetTextBuf to allocate your own
text file buffer either in the data segment or on the
heap.
37. Why aren't the new settings used after I install TURBO.EXE
using the TINST.EXE program?
You probably have a .TP file in the current or Turbo
directory being loaded and the settings in the .TP file
override the settings installed by TINST. Delete the .TP
file.
38. What is the largest string supported?
A string can be 255 characters long. You can write your
own routines to handle strings with lengths greater than
255 characters.
39. Can I still write to file 'Con' without changes?
The 'Con' file is no longer supported, but you can still
write to the screen with a simple Write with no file
variable. The file system has been completely redesigned
to allow you to write your own text file device drivers.
With these, you can implement a Pascal-like text-file
interface to any device, such as serial ports, windowing
systems, memory, etc.
40. What is constant merging?
For example, when you use the same string constant more
than once in a program block, the compiler only saves one
copy of this string. In the generated program, a pointer
is created that references the one copy of this string in
the generated .EXE file.
41. Have Turbo Pascal 3.0 run-time error codes changed in
Turbo Pascal 5.5?
Yes, error codes have changed; refer to Appendix I in the
Reference Guide. The Turbo3 unit contains a version 3.0
compatible IOResult function.