Aztec C86
Release notes
Beta version 5.2a
17 Nov 1992
This release document describes version 5.2a of Aztec C86.
It's organized into the following sections:
1. Installation instructions, which describes how to install
the Aztec C files from the distribution disks onto your
disks;
2. Disk contents, which describes what's on the distribution
disks;
3. New features, which describes what's changed in going
from version 4.2b to the current release;
4. New features, which describe what's changed in going from
the 9 April beta release to the current release.
5. New features, which describe what's changed in going from
the 4 Oct beta release to the current release.
After installation, additional documentation is also
available in file README2.TXT in the DOC subdirectory that's
within your root Aztec directory.
To report problems with this beta release, please call Manx
Tech Support at 908-542-1795, between 2:30pm and 5pm Eastern
time.
1. Installation
To install Aztec C86, put the first distribution disk in a
floppy drive, make that drive the current drive, and type
install
The install program will display a dialog box containing
text fields, check box fields, and button fields that you can
adjust to customize the installation. To move around within
this dialog box, you can use a mouse, or the tab key. You can
get help by pressing the F1 key.
Once you've made your selections, press the "Proceed" button
to continue with the installation, or press the "Abort" button
to terminate the installation. After you press the "Proceed"
button, insert the distribution disks in the source drive as
prompted by the program.
When the files are all installed to the hard disk, the
program will return to DOS.
After the installation program is complete, the bin
subdirectory of your main Aztec directory will contain a batch
file named AZ86.BAT. The commands in this file will initialize
your system for use with Aztec C86. You may wish to place the
contents of this file directly in your AUTOEXEC.BAT file.
Release notes page 1
Aztec C86, beta v5.2a 11/17/92
The following paragraphs describe the fields in the
installation program's dialog box.
1.1 The 'Source drive' text field
In the 'source drive' text field, type the name of the
floppy drive in which you will place the distribution disks.
This field defaults to the A: drive.
1.2 The 'Destination directory' text field
In the 'destination directory' text field, type the name of
the drive and directory into which you want the installation
program to place the Aztec C86 files. This field defaults to
C:\AZ86.
The installation program will place files in the following
subdirectories of the destination directory:
BIN The executable programs;
INCLUDE Include files;
LIB Library source, object modules, and libraries;
MON86 Code for RDB monitor program.
1.3 The 'Install programs, includes, and source' check boxes
Check these boxes depending on what you want installed.
You toggle the setting of a box by pressing the space bar,
or by clicking the mouse in the box.
1.4 The 'Object libraries' check boxes
The Aztec C86 distribution disks contain the following
categories of object module libraries:
C libraries Non-floating point functions;
IEEE math libraries Floating point functions, which
perform calculations using software;
8087 math libraries Floating point functions, which
perform calculations using an 8087;
Sensing math libraries Floating point functions, which
perform calculations using an 8087 if
present or using software if not;
Screen libraries Functions for accessing the screen
and keyboard via the ROM BIOS;
Graphics libraries Graphics functions.
Check the boxes that define the libraries you want
installed.
1.5 The 'Memory models' check boxes
Check on the 'library memory models' boxes to define the
memory models of the libraries that you want installed.
Release notes page 2
Aztec C86, beta v5.2a 11/17/92
For example, if you want the C libraries installed that use
the small code and small data memory model, you would check
the 'C libraries' and the 'small code, small data' boxes.
1.6 The 'Optional components' check boxes
If you've purchased the optional Level 2 or Level 3
Extensions to Aztec C86, check the appropriate boxes to
specify which of the Extension components you want installed.
Don't check these boxes if you don't have the extensions.
1.7 Bypassing the installation program
Most of the files on the distribution disks are in standard
LHARC format. So if you want, you can bypass the installation
program altogether and dearchive the archive files that are on
the distribution disks using the LHARC program that is
provided with Aztec C. This archive program is installed in
the Aztec root directory's BIN subdirectory; this program, and
documentation on how to use it, is also on distribution disk
1. The LHARC program is in file LHA.EXE, and documentation is
in file LHA.DOC.
1.8 Re-entering the installation program
The installation program can be run more than once. For
example, during one execution you could install the programs,
include files, and object libraries; and during a second
execution you could install the library source files.
2. Contents of distribution disks
v5.2a of Aztec C86 is packaged into a base package and two
levels of optional extensions. The components of these
packages are:
* Base package:
+ Compiler, assembler, DOS linker
+ Include files
+ Object module utilities
+ Object libraries
+ SDB source level debugger.
* Level 2 extensions:
+ Library source;
+ HEX86, startup routines, and heap-management
routines, for ROMming programs that have non-complex
segmentation requirements.
* Level 3 extensions:
+ RDB debugger and MON86 monitor, for remote source-
level debugging of ROM- and DOS-based programs;
+ ALN linker, ALOC locator, AHEX hexer, and startup
routines, for ROMming programs that have complex
segmentation requirements
The following paragraphs describe the files that are on the
Aztec distribution disks.
Release notes page 3
Aztec C86, beta v5.2a 11/17/92
2.1 Disk 1 contents
File Contents
INSTALL.EXE Installation program
LHA.EXE LHARC program (this replaces the ARCV
and MKARCV programs that are described
in the manual)
LHA.HLP LHARC help
LHA.DOC LHARC documentation
2.2 Disk 2 contents
File Contents
CC.LZH Compiler
AS.LZH Assembler
LN.LZH Linker
INCLUDE.LZH include files
UTIL.LZH Object module utilities, HEX86, and
miscellaneous utilities
UNITOOLS.LZH Unix-compatible programs
2.3 Disk 3 contents
File Contents
SDB.LZH SDB source level debuger
2.4 Disk 4 contents
File Contents
C.LZH Small code, small data c & m libraries
CL.LZH Large code, large data c & m libraries
CLC.LZH Large code, small data c & m libraries
CLD.LZH Small code, large data c & m libraries
M87.LZH 8087 math libraries (all memory models)
MS.LZH Sensing math libraries (all models);
G.LZH Graphics libraries
S.LZH Screen libraries
2.5 Disk 5 contents
Disk 5 contains the optional Level 2 extensions.
File Contents
86SUP.LZH LIB\86SUP source files
DOS.LZH LIB\DOS source
GRAPHICS.LZH LIB\GRAPHICS source
INP.LZH LIB\INP files
LIB.LZH LIB files
MISC.LZH LIB\MISC source
MATH.LZH LIB\MATH source
ROM86.LZH LIB\ROM86 source
SCREEN.LZH LIB\SCREEN source
STDIO.LZH LIB\STDIO source
STDLIB.LZH LIB\STDLIB source
STRING.LZH LIB\STRING source
SYSIO.LZH LIB\SYSIO source
TIME.LZH LIB\TIME source
UTIL2.LZH Programs for level 2 extensions
2.6 Disk 6 contents
Disk 6 contains the optional Level 3 extensions.
File Contents
Release notes page 4
Aztec C86, beta v5.2a 11/17/92
LNKLOC.LZH Advanced linker, locator, and hexer
programs
RDB.LZH Remote source level debugger
3. New Features since v4.2a
This section describes functional changes that have been
made in going from the v4.2b release to the 9 April v5.2a beta
release.
3.1 Compiler changes
The following paragraphs describe the changes that have been
made to the compiler in going from version 4.2b to 5.2a.
3.1.1 New compiler options
-J Enable trigraph support.
-F Generate prototypes for static functions, without
generating executable code. Send prototypes by default to
a file whose name is derived from input file, by changing
the extension to .PRO. The -O option can be used to
select the output file name. This option replaces the
PROTO program.
-X Generate prototypes for non-static functions, without
generating executable code. Send prototypes by default to
a file whose name is derived from input file, by changing
the extension to .PRO. The -O option can be used to
select the output file name.
-W Preprocess source. Send preprocessed C source by default
to a file whose name is derived from input file, by
changing the extension to .PRE. The -O option can be used
to select the output file name. This option replaces the
CPP program.
3.1.2 New predefined symbols
__VERSION Predefined with the version of the compiler. It
has the value 520;
_LARGE_CODE Predefined with value 1 if program is compiled
to use the large code memory model;
_LARGE_DATA Predefined with value 1 if program is compiled
to use the large data memory model;
3.1.3 New language features
Support for the following language features has been added
to the v5.2 compiler:
* Initialization of auto aggregate variables;
* Trigraphs.
3.1.4 New support for 8086 segmentation
The compiler now allows you to define, using the new
"#pragma segment" directive, the segments into which variables
are placed. This directive has the following form:
Release notes page 5
Aztec C86, beta v5.2a 11/17/92
#pragma segment [segname]
[class classname] [at addr] [near | far]
sym1, sym2 ...
where the parameters have the following meanings:
sym1, sym2, ...
Symbols that are to be placed in a
specified segment;
segname
'segname' is the segment that's to hold
the symbols.
class classname
'classname' is the class in which the
segment is located. If not specified, the
class defaults to 'DATA'.
at addr 'addr' is the segment's starting address;
near | far A NEAR segment is always accessible via
the DS segment register, and a FAR segment
is not. To access a FAR segment, the ES
register must first be loaded with the
segment's paragraph address. Defaults to
FAR.
Previous versions of the compiler allowed you to specify
that variables were to go into their own segments, but the
compiler itself chose the names of these segments. With the
ability to choose segment names yourself, and with the new ALN
linker and ALOC locator, you can now easily position variables
in special areas of memory. For example, if your system has
memory-mapped I/O, you can create structure that matches the
memory-mapped I/O section of memory, place that structure in
its own segment, and position the segment at the memory-mapped
I/O's section of memory.
Previous versions of the compiler did not allow you to put
near data items into special segments. With this capability,
and with the new ALN linker and ALOC locator, you can now
easily take advantage of battery-backed-up RAM. For example,
you can place variables that should not be initialized on
system startup in a DS-accessible near segment, and position
that segment at the beginning of RAM. The other standard
Aztec-created, DS-accessible, near segments can be placed
after the special segment in RAM, where they can safely be
initialized on system startup.
3.2 Assembler changes
* ORG directive now fully supported;
* The "AT addr" clause of the SEGMENT directive is now
supported.
3.3 DOS linker changes
* None.
3.4 Library changes
* Libraries are now ANSI compatible (previously they were
K&R compatible). Probably the main change you will have
to make to your programs is to change your calls to
fopen(), to support fopen's TEXT and BINARY modes.
Release notes page 6
Aztec C86, beta v5.2a 11/17/92
* The time() function now returns the current time as the
number of seconds that have elapsed since Jan 1, 1970.
Previously it returned the current time in the format
used for file directory time fields.
3.5 Include file changes
* Include files are now completely ANSI compatible
(previously they were K&R compatible).
3.6 Level 2 extensions for creating ROMable code
The optional Level 2 extensions contain components for
creating ROMable code of systems whose segmentation
requirements are not too demanding. For example, a typical
system might have its code in one set of ROM chips and its
data in one set of RAM chips.
These components are:
* The HEX86 utility that was in v4.2b, for converting a
program into Intel hex code;
* BEGIN-SD.ASM and HEAP-SD.ASM, the startup and heap
management routines for programs that use the small data
memory model;
* BEGIN-LD.ASM and HEAP-LD.ASM, the startup and heap
management routines for programs that use the large data
memory model;
The heap management routines are new; the routines in 4.2b
and earlier releases were for DOS-based systems.
BEGIN-SD.ASM and BEGIN-LD.ASM have replaced the ROM86.ASM
startup routine that was in 4.2b. They were changed so that
they now initialize the heap as required by the new heap
management routines.
3.7 Level 3 extensions for creating ROMable code
The new, optional, Level 3 extensions to Aztec C86 contain
the following utility programs and startup code for creating
multi-segmented romable code:
* The ALN linker. This program links object modules
together, and outputs a relocatable object module. The
input modules can be in Aztec object format, Microsoft
OMF format, or Intel OMF format. The output module is in
Intel OMF format.
* The ALOC locator. This program 'locates' an object module
that is in relocatable Intel OMF format; that is, defines
the starting memory addresses of the module's segments
and converts relocatable items to absolute.
* The AHEX hexer. This program converts an object module
that is in absolute Intel OMF format into Intel hex code
or binary code.
* The ABEGINSD and ABEGINLD startup routines, which gain
control when a program created with ALN, ALOC, and AHEX
is activated.
Release notes page 7
Aztec C86, beta v5.2a 11/17/92
For heap management, programs that use the Level 3
extensions should use the heap management routines that are in
the Level 2 extensions.
ALN, ALOC, and AHEX are described at the end of this
document. ABEGIN is described in the following section.
3.8 The ABEGIN startup routine.
ABEGIN is the startup routine that's used by programs that
have been generated using the ALN, ALOC, and AHEX programs.
3.8.1 ABEGIN files
To support different memory models, ABEGIN is provided in
two source files and four object module files, in directory
LIB\ROM86. Source files:
ABEGINSD.ASM Source for programs that use small data
memory model;
ABEGINLD.ASM Source for programs that use large data
memory model;
Object files:
ABEGINSD.O Object module for programs that use small
code, small data memory model. Generated
from ABEGINSD.ASM;
ABEGINSD.OLC Object module for programs that use large
code, small data memory model; Generated
from ABEGINSD.ASM;
ABEGINLD.OLD Object module for programs that use small
code, large data memory model; Generated
from ABEGINLD.ASM;
ABEGINLD.OL Object module for programs that use large
code, large data memory model. Generated
from ABEGINLD.ASM;
3.8.2 Notes about ABEGIN
* By default, ABEGIN initializes a program's DATA class in
RAM from a copy in ROM. The copy must immediately follow
the program's CODE class in ROM.
* ABEGIN can optionally initialize a program's FAR_DATA
class in RAM from a copy in ROM. This is the class in
which the compiler by default places a program's far
data. This class must immediately follow the program's
DATA class in ROM.
* ABEGIN contains the target system's reset code, in the
RESET_CODE segment. By default, AHEX will position this
segment so that it resides at the system's reset vector
location.
* ABEGIN contains code that will initialize an 80186. To
enable generation of this code, assemble ABEGINxx.ASM
with the symbol CPU_186 defined.
* The small data version of ABEGIN (ABEGINSD.ASM) puts a
program's stack within the program's uninitialized data
area, beginning at the location defined by symbol
'cstack'. The number of paragraphs in the stack area is
defined by the symbol 'stacksize', which defaults to 80h
paragraphs.
Release notes page 8
Aztec C86, beta v5.2a 11/17/92
* The small data version of ABEGIN puts a program's heap
right after the program's uninitialized data area. The
number of paragraphs in the heap is defined by the symbol
'heapsize', which defaults to 80h paragraphs.
* The large data version of ABEGIN (ABEGINLD.ASM) puts a
program's stack right after the program's uninitialized
data area. The symbol 'stacksize' defines the number of
paragraphs in this area, and defaults to 80h paragraphs.
* The large data version of ABEGIN puts a program's heap
right after the program's stack area. The symbol
'heapsize' defines the number of paragraphs in this area,
and defaults to 80h paragraphs.
3.9 SDB changes
* There are now two versions of SDB: the regular SDB; and
RDB, which allows remote source level debugging of a
program that resides in another MSDOS/PCDOS- or ROM-based
system.
* Both SDB and RDB have a new windowed interface, and
hypertext help.
* INITSDB is no longer needed for configuring SDB.
* RDB and SDB occupy significantly more RAM memory than
v4.10d SDB. This is not a problem for RDB, since the
program being debugged resides in a separate machine. But
it is a problem with SDB since it and the program being
debugged must reside in the same system. Until we resolve
this problem with SDB, which we will probably do by
moving SDB up into extended memory, we are providing the
v4.10d SDB, in file SDB410D.EXE, and INITSDB (for
configuring the SDB410D).
3.10 Z changes
* The command :se 50=1 can be used on VGA screens to
display 50 lines.
3.11 Other program changes
Other new programs in v5.2 are these:
* XREF, which generates a cross-reference listing of symbol
references and definitions in object modules.
Programs that are no longer provided:
* PROTO, which has been replaced with the new compiler
options -F and -X;
* CPP, which has been replaced with the new compiler option
-W;
* ENGTOC;
* CTOENG;
* ARCV.
3.12 Building libraries
Library source is in the optional Level 1 extensions to
Aztec C86. The INSTALL program will create a LIB subdirectory
of the root Aztec directory and several subdirectories of LIB,
and then install the source in these subdirectories.
Release notes page 9
Aztec C86, beta v5.2a 11/17/92
You can build one or more libraries by CD'ing into the LIB
directory and then invoking MAKE with codes that identify the
libraries you want made. The codes and their associated
libraries are:
ALL All libraries
C_ALL All C libraries
C C.LIB
CL CL.LIB
CLC CLC.LIB
CLD CLD.LIB
M_ALL All M libraries
M M.LIB
ML ML.LIB
MLC MLC.LIB
MLD MLD.LIB
M87_ALL All M87 libraries
M87 M87.LIB
M87L M87L.LIB
M87LC M87LC.LIB
M87LD M87LD.LIB
MS_ALL All MS (8087-sensing) math libraries
MS MS.LIB
MSL MSL.LIB
MSLC MSLC.LIB
MSLD MSLD.LIB
S_ALL All S (screen) libraries
S S.LIB
SL SL.LIB
SLC SLC.LIB
SLD SLD.LIB
G_ALL All G (graphics) libraries
G G.LIB
GL GL.LIB
GLC GLC.LIB
GLD GLD.LIB
To build libraries, MAKE will invoke a second copy of MAKE,
passing to it parameters that define what's to be made.
For each library that it's to make, this second MAKE will do
the following:
* CD into each required subdirectory of the LIB
subdirectory and invoke a third copy of MAKE. This third
MAKE will generate object modules for the library, from
the source that's in the subdirectory.
* CD into the INP subdirectory and sort the names of the
object modules that will go into the library, using the
ORD utility;
* Build the library into the LIBS subdirectory of the LIB
directory, using the LB utility.
Release notes page 10
Aztec C86, beta v5.2a 11/17/92
4. New Features since the 4/9/92 beta
This section describes changes that have been made since the
9 April 92 beta release. It first describes functional changes
that have been made, and then bug fixes.
4.1 Compiler changes
* The new -M option causes the compiler to issue a warning
if a function is called without a prototype being in
scope.
* The new +T option causes $$var statements (used to access
far data) to be put into the code segment instead of the
data segment.
4.2 Z changes
* On a VGA display, the command :se 50=1 can be used to
display 50 lines.
4.3 Library changes
* The new TimeFile2TM() function converts a time value
that's in the format used for file directory time fields
to the format that's used by the time() function.
* The new TimeTM2File() function converts a time value
that's in the format used by the time() function to file
system format.
* The graphics functions now support VGA modes 0x10, 0x11,
0x12, and 0x13.
* The movblock() function has been changed. A length
argument of 0 now means don't copy anything; used to mean
copy 64kbytes.
* The scdir() function has been changed. When passed a
NULL, it resets itself so that the next call can specify
a new pattern.
4.4 SDB/RDB changes
* The DI command causes sdb/rdb to display information
about the debugger's heap usage.
5. New features since the 10/4/92 beta
5.1 SDB/RDB changes
* The name of the default configuration file for SDB and
RDB has been changed to SDB.SET, from SDB.INI.
* The sequence in which SDB/RDB search for SDB.SET has been
changed. On startup they look in the following places:
1. The current directory;
2. The directory whose name is assigned to the AZTEC
environment variable;
3. If SDB.SET is not found and AZTEC environment
variable doesn't exist, SDB.SET is created in the
current directory.
5.2 CNM changes
* The -U option has been added, to display information
about a module's unnamed symbols.
Release notes page 11
Aztec C86, beta v5.2a 11/17/92
5.3 Library changes
* The following functions have been added:
+ fnsplit(), which breaks a file name up into drive,
path, filename, and extension components; Calling
sequence:
int fnsplit(char *name, char *drive, char *dir,
char *fname, char *ext);
where
- name points at the name to be broken down;
- drive points at the area where the name's drive
component is placed;
- dir points at the area where the name's
directory component is placed;
- fname points at the area where the name's
filename component is placed;
- ext points at the area where the name's
extension is placed.
+ fnmerge(), which creates a file name from drive,
path, filename, and extension components. It has the
same arguments as fnsplit(); a complete file name is
built up from the drive, dir, fname, and ext args,
and stored in the area pointed at by name.
+ strdup(), which makes a copy of a string in the
heap. Calling sequence:
char *strdup(char *str);
+ lfind(), which performs a linear search of an array
for a specified item. Calling sequence:
void *lfind(void *key, void *base, size_t *num,
size_t width, int (*fcmp)(void *, void *));
where:
- key points at the value to be located;
- base points at the array to be searched;
- num points at the number of array elements;
- width is the width in bytes of each array
element;
- fcmp points at a function that compares two
keys. The function returns 0 if the items are
identical, and non-zero if different.
lfind() returns a pointer to the matching array
element, or NULL if not found.
+ lsearch(), same as lfind(), except it appends the
item if not found. The calling sequence is the same
as for lfind().
+ min(), returns the lesser of two ints. Calling
sequence:
int min(int a, int b);
where a and b are the items to be compared.
+ max(), returns the greater of two ints. Calling
sequence:
int max(int a, int b);
where a and b are the items to be compared.
* The following symbols were added to the ABEGINSD.ASM and
ABEGINLD.ASM startup routines:
+ NEAR_DATA_INIT. If 1, the initialized near data
segment (dataseg) is initialized from a copy in ROM.
If 0, this initialization is not done. Default
value: 1.
Release notes page 12
Aztec C86, beta v5.2a 11/17/92
+ FAR_DATA_INIT. If 1, the initialized far data class
FAR_DATA is initialized from a copy in ROM. If 0,
this initialization is not done. Default value: 0.
+ FAR_BSS_INIT. If 1, the uninitialized fat data class
FAR_BSS is cleared. If 0, FAR_BSS is not cleared.
Default value: 0.
+ CPU_186. Set to 1 if processor is an 80186/80188, or
0 if not. Default value: 0.
5.4 The AI program
AI is a new program that displays information about Aztec
C86 programs. For information, type AI.
5.5 Bug fixes
Numerous bugs have been fixed, in programs, library
functions, and include files. These fixes are described
in the CHANGES.* files that are stored in the BIN,
INCLUDE, and LIB subdirectories of the main Aztec
directory. Once you've finished examining these files,
you can delete them.
Release notes page 13
