8bitfiles.net/archives/archives.tar/archives/compuserve-file-archive/05 Programming/EA2.DOC

Text File | 2019-04-13 | 24KB | 700 lines

Documentation for EA - Part Two of Two.

EA is an editor-assembler.  Copyright 1984 by Lew Lasher.

Overview of assembler features:

EA is an assembler for the Commodore 64.  An assembler is a
program used to translate machine language programs from
person-readable form (the source program) to a machine-usable
form (the object program).  The assembler allows you to write,
maintain, and modify a large, complicated machine language
program.

Operation:

EA requires the source file to be put in one or more text files
on a floppy disk.  After you have created the source files, and
written them to disk (using the C= W command), you can run the
assembler by typing C= A.

EA asks you for the name of the source file.  If you type
"blurfo" for the name, EA first looks for a file named "blurfo".
If it cannot find "blurfo", then it looks for "blurfo.a".  If
it cannot find either one, it prints an error message and returns to the editor.
If it finds either one, it creates an object file called
"blurfo.o".  The filename you type must not be longer than 14
characters, to make room for the extension ".o".

Then EA asks you if you want to make a list file.  You should
type "y" or "n".  The list file, which in the current example
would be named "blurfo.l", lists the numeric value of every byte
in the object file, alongside the address at which the byte will
be loaded and the source line corresponding to that byte.  List
files are useful in debugging programs.

All these files must be on a disk drive with device number 8
and, if a dual drive, drive number 0.

After optionally creating the list file, EA begins the first
pass.  Unless there is an error on the disk drive, you will not
see anything on the screen during the first pass.  No output is
written to either the object file or the optional list file
during the first pass.  The only purpose of the first pass is to
figure out the addresses where the program will be located, and
thereby assign numeric values to each label and symbol in your
program (See below for more about symbols and labels).

After the first pass is completed, EA prints the message
"Beginning 2nd pass" and does just that.  During the 2nd pass,
you should see on the screen a listing of your source program,
but with various numbers over in the left margin, and various
error messages preceding the erroneous source lines.  If you
chose to make a list file, the text you see on the screen will
be the text you get in the list file.  Meanwhile the object
program is being written to the object file.

Finally you will see the message "Errors detected: 0" which
indicates that EA did not detect any of the errors in your
program.  Note that any errors that EA detects are, in a sense,
"mere warnings", in that (except for I/O errors), you still are
given an object file which can be loaded and run.

At any point in either the first or second pass, you may cut off
the assembly by hitting the RUN/STOP key.  You should delete the
incomplete object and list files which will have been created.

After EA exits, you can load and run the object file by typing:

        load "blurfo.o",8,1
        sys n

where n is the address of the start of your program.

Miscellaneous specifications:

EA will not work with disk drives other than a Commodore 1541 or
an MSD SD-2.  If changes are made to the ROM's in these models,
it may not even work with them.

Differences from other assemblers:

There are numerous differencs between EA and other assemblers
for the C-64.  Many of these are "added features" of EA which
will not interfere with the assembly of programs written for
another assembler.  The following are the major differences
which must be taken into account to modify a program written for
another assembler:

1.  Every label must be followed by a colon (":").

2.  The directive * = n is not supported.  In other assemblers,
this directive is used primarily for two purposes.  The first
purpose is to establish the address for the first byte in the
object program.  For that purpose, replace:

        * = 2049

with:

        .origin 2049

The second common purpose is to allocate an area of memory
without specifing the contents of that memory.  For that
purpose, replace:

        * = * + 1

with:

        .blkb 1

3.  Commodore's assembler has a directive called ".byt" with can
be used to generate both numeric and text data.  In EA, this
directive is called ".byte", and only generates numeric data,
while the .ascii directive generates text data.  Therefore,
replace:

        .byt 15, 'Text string', 0

with:

        .byte 15
        .ascii 'Text string'
        .byte 0

4.  EA lacks the macro directives, conditional assembly
directives, and other specialized directives found in other
assemblers.

5.  EA does not require or allow the ".end" directive which many
assemblers require at the end of a source file.  Actually, EA
ignores the .end as an unrecognized directive, but you should
delete the .end to avoid the error mesage.

Format of source file:

The source file should be a sequential file, with each line
terminated by a RETURN characater.  Lines are limited to 255
characters, not including the RETURN character.  Blank lines are
legal, and spaces can be put in liberally to improve readability
(exceptions: spaces may not be put in the middle of a numeral or
an identifier).  Upper and lower case letters are treated
identically except in .ascii or .asciz directives and in
character literals.  A semicolon (";") is used to indicate the
beginning of a comment.  All text after the semicolon, and the
semicolon itself, is ignored (exceptions: text strings in .ascii
and .asciz directives and character literals).

A valid line can be any of the following:

1.  A blank line.

2.  A symbol definition, e.g.:

        cursordown = 17

3.  An instruction for the 6510 microprocesor, e.g.:

        lda # 0

4.  An assembler directive, e.g.:

        .byte 5

5.  Any of the above (1-4) preceded by one or more labels, e.g.:

        start: startloop: lda # 0
        length: .byte 0
        blankline: ; comment

Blank lines:

Blank lines are useful for the following purposes:  to make the
source or list file more readable, to hold a label, and to hold
comments.  I apologize to people who would not consider a line
having a label and/or a comment to be a "blank line".

Symbol definitions:

A symbol definition is used to define a name to represent a
numeric value.  One use of symbols is to refer to addresses not
located within the object file, for example, KERNAL routines,
page zero, the special addresses used by the video, sound, and
I/O chips, and vectors:

        chrout = 65490
        current.key = 197
        color.memory = $D800
        ADSR1  = 54277
        cia1.data.port.A = $dc00
        irq.vector = 788

Symbols can be used for various constants in your program, to
make the program easier to understand and to modify, e.g.:

        space.char = 32
        disk.devicenum = 8
        pointer = 251

Sometimes you may use a symbol to refer to a memory location
within the object file.  In general, labels are preferred over
symbols for this purpose.  However, note that the following are
very similar in effect:

        location:       ; label
        location = *    ; symbol

The difference is that a symbol can be redefined many times
within the program without error, while a label is supposed to
refer to only one location.  The one instance in which it makes
sense to use a symbol is to mark the beginning of a loop:

        loop = *
            ; body of loop
            dex
            bne loop

Since the assembler uses the most recently-assigned value of the
symbol, this pattern can be used for many loops, without having
to think of distinct names for the start of each loop.  But note
that this pattern will fail miserably for a forward branch, or
for a loop within a loop.

Because the first pass of the assembler is devoted to defining
labels and symbols, you can refer to a symbol or label that is
defined later in the source file.  This is called a "forward
reference".  More precisely, you are allowed one level of
forward reference.  For example:

        a = b + 40
        b = 1024

is legal, but:

        a = b
        b = c
        c = 100

is not legal.  You should not, however, use even a single-level
forward reference in a .blkb or .blkw directive.  (These
directives are described below).  The assembler uses these
directives in the first pass to figure out how much memory to
allocate.  If a symbol used in a .blkb or .blkw directive is
undefined on the first pass, then all subsequent labels will be
defined incorrectly.  Symbols that refer to page zero should be
defined before they are used.  Since instructions referring to
page zero usually take 2 bytes, as opposed to 3 bytes for
addresses on any other page, the assembler needs to know during
the first pass whether an instruction refers to page zero or
not.  If the address is specified by an undefined symbol, the
assembler assumes that the address is NOT on page zero.  If this
assumption proves incorrect, all subsequently-defined labels
will be defined incorrectly.  Therefore, all symbols referring
to page zero MUST be defined before the instructions that refer
to the symbols.

Since it is an error to redefine a label, it is also an error to
define a symbol with the same name as a label.

Identifiers:

An "identifier" is the name used for a symbol or label.  An
identifier must start with a letter, and can have letters,
numeric digits, or periods for subsequent characters.  The
length of an identifier is not limited other than by the maximum
length (255 characters) of a source line.  Upper and lower case
letters are treated identically in identifiers.  The names A, X,
and Y should not be used, even though legal, because other
assemblers forbid their use.  This assembler allows their use,
except that a lone A as an operand to a 6510 instruction will be
confused for a reference to the 6510's accumulater.

Examples of legal identifiers:     and illegal identifiers:

            Q                       8
            A3333333                3AAAAAAA
            a.b.c                   .a.b.c

Machine instructions:

(After all, the whole point of an assembler is to let you enter
machine instructions.)  EA recognizes the standard 56
instruction mnemonics, which you may type in either upper or
lower case.  EA uses the standard punctuation for the various
addressing modes:

        clc         ; implied
        adc # 1     ; immediate
        asl A       ; accumulator
        ldx Z       ; zero page
        ora Z,X     ; zero page, X
        and Z,Y     ; zero page, Y
        cpx E       ; absolute
        ldy E,X     ; absolute, X
        ldx E,Y     ; absolute, Y
        beq E       ; relative
        eor (Z,X)   ; (indirect,X)
        sbc (Z),Y   ; (indirect),Y
        jmp (Z)     ; absolute indirect

where Z is an expression whose value is between 0-255, and E is
an expression whose value is not necessarily between 0-255.
(The exact definition of an "expression" will be given later.)
Note that it is often not the punctuation, but the value of the
expression, which distinguishes certain addressing modes.

You may, but need not, put in spaces before or after a "(", ")",
",", or "#".  You must put in a space after the name of the
instruction if the operand starts with a letter.

Expressions:

An expression is a bunch of text that represents a numeric
value.  The following are legal expressions:

1.  A decimal numeral (0 to 65535):

        1024

2.  A hexadecimal numeral, preceded by a dollar sign:

        $ff
        $DC00

3.  A binary numeral, preceded by a percent sign:

        %10000001

4.  A character literal, preceded by an apostrophe:

        'A

5.  An identifier:

        loop
        color.memory
        disk.devicenum

    The numeric value of an identifier is given by its most
    recent definition as a symbol or as a label.

6.  The "location counter", either a period or asterisk:

        .
        *

    The numeric value of the location counter is the address in
    the object program corresponding to the beginning of the
    current source line.

7.  An expression preceded by a "unary operator":

        < E         ; low-order byte of E
        > E         ; high-order byte of E
        - E         ; negative E (two's complement)
        ? E         ; logical (one's) complement of E

8.  An expression followed by a "binary operator" followed by a
second expression:

        E1 + E2     ; addition
        E1 - E2     ; subtraction
        E1 * E2     ; multiplication
        E1 / E2     ; integer division
        E1 % E2     ; modulo (remainder after integer division)
        E1 & E2     ; logical AND
        E1 ! E2     ; logical OR
        E1 ^ E2     ; arithmetic shift
                    ; left shift if E2 is positive
                    ; right shift if E2 is negative

If more than one operator is used in the same expression, they
are done in the following order: ? - (negative) * % / -
(subtraction) + & ! ^ < >.  If the same operator appears twice,
evaluation is from right to left.  You can override the normal
order of evaluation by using brackets ("[" and "]") as
parentheses.

Note, however, that all these arithmetic operations may be done
only on constant expressions.  That is, you cannot use the above
operators to add the contents of various memory locations.  The
usefulness of elaborate assembler expressions is primarily to
define one symbol in terms of another, so that if you change one
symbol, various other symbols are automatically redefined.

Assembler directives:

The assembler directives perform various miscellaneous functions:

.byte
.byte Z
.byte Z1, Z2, ... , Z3

    The .byte directive allocates one or more bytes in the
    object file.  If an expression is given, the expression is
    put in as the initial value of the byte.  If more than one
    expression is given, separated by commas, than a byte is
    allocated for each expression.  If no expression is given, a
    single byte is allocated with an initial value of zero.  Only
    the low-order byte of the expression is used.

.word
.word E
.word E1, E2, ... , E3

    The .word directive is very similar to the .byte directive,
    but two bytes are allocated for each expression, the first
    for the low-order byte and the second for the high-order
    byte.  If no expression is given, then two bytes are
    allocated, both initialized to zero.

    Note that

            .word E

    is exactly equivalent to:

            .byte < E, > E

.blkb
.blkb E

    The .blkb directive is used to allocate memory in the object
    program without specifying the initial values.  If an
    expression is given, that number of bytes is allocated.  If
    no expression is given, 1 byte is allocated.  But if the
    value of the expression is zero, no bytes are allocated.  If
    the expression is negative or greater than 32767, an error
    message is generated and no bytes are allocated.

.blkw
.blkw E

    The .blkw directive is similar to the .blkb directive,
    except that the number of bytes to be allocated is
    multiplied by two.

.ascii "Text string"

    The .ascii directive is used to allocate memory in the
    object program for a text string.  One byte is allocated for
    each character in the text string, and each byte is
    initialized with the character code (not the code for screen
    memory) for the character from the source line.  The text
    string can be delimited by any character not found within
    the text string, with the exception that a semicolon (";"),
    a space, or a RETURN character may not be used as the
    delimiter.  Note that the character codes are not true ASCII
    but the Commodore variant, sometimes called PETSCII.

.asciz /Text string/

    The .asciz directive is similar to the .ascii directive,
    except that a zero byte is allocated after the text string.

.origin E

    The .origin directive is used at the beginning of the source
    file to specify the address to be used for the first byte in
    the object file.  In fact, the .origin directive need not be
    literally on the first line of the source file, but it must
    come before any source lines that allocate memory in the
    object file.  If you do not include a .origin directive, the
    default origin is 49152 ($C000 hexadecimal).

.print E text

    The .print directive is used to print messages on the
    terminal and into the optional list file.  The expression
    following the .print is evaluated.  The numeric value of
    that expression is printed, in decimal, to the screen and
    into the optional list file, followed by any other text on
    the source line.  The main use of the .print directive is to
    produce a memory map rather than a full listing of every
    line of the program.  The output from the .print directive
    is performed even if the .nolist directive is in effect.

    Examples:

            .print subroutine subroutine
            subroutine:
                ...
                ...
                rts

            .print * - subroutine, length of subroutine

            .print * End of file
            .print * - start, Size of file

.nolist

    The .nolist directive turns off the listing of every source
    line to the screen and to the optional list file.  The
    output of the .print directive is still performed, however.

.list

    The .list directive turns back on the listing of every
    source line, after a previous .nolist directive.  In this
    way, you can make a list file of only those particular
    routines which you are interested in debugging.

.include identifier

    The .include directive is used to incorporate another source
    file as though it had been merged with the principal source
    file.  This allows you to have a library of common
    subroutines without having to make a separate copy of the
    subroutines in each program in which you use them.  It is
    also useful to make several, slightly different versions of
    the same program.  It is also useful to divide up a large
    program for more convenient editing.  It is also useful for
    incorporating a list of common symbol definitions, such as
    the addresses of the KERNAL routines.  The assembler first
    looks for a file whose name is given by the identifier
    (converted to lower case).  If that file cannot be found, it
    adds ".a" to the filename and looks for that name.  After
    the assembler has finished processing the included source
    file, it returns to the principal source file and continues
    from the line following the .include directive.  The
    included file can include a "third-level" source file, but
    there is a maximum of four levels of inclusion.

Error messages:

File name may be 1 to 14 chars long

    If, when the assembler initially asks for the name of the
    source file, you hit the RETURN key with no name, or give a
    name longer than 14 characters, you get this message.

Couldn't create object file:
Couldn't create list file:
Error reading source file:
Error writing to list file:
Error writing to list file:
Error closing object file:
Error closing list file:
Error re-reading source file

    Also prints the error message given by the disk drive

Couldn't open control or random channel

    May indicate that you ran EA while file channels were still
    opened from another program.

Line too long

    Source line longer than 255 characters.

Assembly terminated by STOP key

Unrecognized statement

    The first non-blank character in the source line was neither
    a letter nor a period.  May indicate that the source line
    contains a non-printing character.

Unrecognized opcode

    May indicate a misspelled opcode, or the accidental omission
    of the colon after a label or of the = following a symbol
    definition.

Extra input ignored

    May indicate the omission of the semicolon preceding a
    comment.  May also indicate the omission of an operator from
    an expression.

No character after '

    Indicates that an apostrophe was the last character in the
    source line.

Malformed expression

Warning - extra [

Warning - extra ]

Warning - arithmetic overflow

    Indicates an arithmetic operation whose result required more
    than 16 bits.

Warning - division by zero

Internal assembler error

    This message indicates that the assembler was not able to
    evaluate an expression properly.

Second argument to ^ must be -15 to 15

Undefined symbol: identifier

Symbol table full

    Indicates that the assembler was not able to create a label
    or symbol because of insufficient available memory.

Warning - redefining label:

    This may indicate that the source file has two different
    labels with the same name, or a symbol and a label with the
    same name.  It also may indicate a duplicate name from an
    include file.  If, however, you get this error message for
    every label in the source file, it means that the value of
    the labels changed from the first pass to the second pass.
    This, in turn, usually means that a symbol was used in a
    .blkb or .blkw directive before it was defined, or that a
    symbol referring to page zero was used before it was
    defined.

Unrecognized addressing mode

Improper addressing mode for opcode

    Not all the possible addressing modes can be used for any
    particular opcode.  See the C-64 Programmer's Reference
    Manual, or any other book about the 6502, for a description
    of the permissible addressing modes for each opcode.

Expression truncated

    An expression used with an opcode in immediate mode or used
    in a .byte directive was not within the range -256 to 255.

Branch out of range

    The various branch instructions for the 6510 may only be
    used to branch to a location less than 254 bytes before the
    branch instruction or less than 253 bytes after the branch
    instruction.  This error message is also often given if the
    expression specifying the address uses an undefined
    identifier.

Unrecognized directive

    A source line began with a period, but was not one of the
    assembler directives described above.

Warning - extra comma

    A .byte or .word directive had a trailing comma after the
    last expression.

Warning - comma inserted

    A .byte or .word directive had two expressions not separated
    by a comma.

Argument must be positive

    The argument to .blkb or to .blkw was negative, or greater
    than 32767.

Missing string

    An .ascii or .asciz directive did not have a text string, or
    used a semicolon as the delimiter.

Unterminated string

    An .ascii or .asciz directive was followed by a text string
    which was not followed by the delimiter character preceding
    the text string.

Origin can only be set once at start

    The source file contained two .origin directives, or
    contained an .origin directive after a source line which
    allocated memory.  May indicate that an include file had an
    .origin directive.

Filename must start with a letter

    A .include directive was not followed by an identifier.

Already at maximum depth of includes

    Only four levels of source files are allowed.

Couldn't open include file:

    Prints the error message from the disk drive

File number not opened
File number out of range
File number already in use
Directory not updated

    These messages indicate internal errors in the file-handling
    routines used by the assembler.

For more information, call Lew Lasher at [phone]