Hello Hector Peraza,

while M80.COM handles 8bit references correctly (see attachment testm80.prn), ZSM4 generates the error "R" (see testzsm4.prn). The listing test2.prn shows the second file in which the variable "ValNum" is defined, which is referenced by TESTM80.MAC and TESTZSM4.MAC. When I link TESTM80.REL and TEST2.REL to TESTX.COM/TESTX.SYM with L80.COM, ZSID.TXT shows the correct use of "ValNum" (value 10H).

It seems that ZSM4 version 4.7 does not correctly handle the use of 8Bit (LOW) from the Extrn defined 16Bit variable "ValNum", which however the M80.COM/L80.COM and also the SLR180.COM/SLRNK.COM can do correctly.

Otherwise, I would like to thank you for your valuable work on ZSM4, it is a great contribution to the open source community. I use ZSM4 under my Z80 emulator "YAZE-AG" (Yet Another Z80 Emulator by AG, www.yaze-ag.de and also here on guthub.com) and on my real Z280 machine "CPU280" by
test2.txt
testm80.txt
testzsm4.txt
zsid.txt
Tilmann Reh (Germany).

Best regards
Andreas Gerlich (agl7, IRCNet: agl, Libera.Chat: agl)
(University of Ulm, Germany)

When the source contains one of the following typos (probably more, but haven't tested), then zsm4 does the strangest kind of things (i.e. crashing, exiting without compiling, writing its name in an endloss loop etc.) :

A>type test.mac
        .z180

        out0    (66h)
        out0    66

        end

When the instructions are typed correctly, then everything works out fine.

Tested with zsm4 v4.7 on an actual Z180-Computer and in a z80-emulator.

I have Z80 code containing the "LD I,A" instruction. The instruction generates an "O" error when I specify the /U (undefined symbols) option to ZSM4 but compiles correctly if I leave that option off. See below:

You have an .ODD and .EVEN… directive it would be convenient to have an ".ALIGN ". See https://github.com/jamesots/maz for example.

HI Hector,

Was updating the copy of ZSM from 41 to 42 in the ROMWBW distribution but on checking, the final .BIN file for the forth rom after linking was all 0's. There are no errors generated from assemble but rel files are slightly different. Linker is DR LINK 1.31. No changes to build process. REL files attached.

This is the link to directory for the files being assembled with ZSM

https://github.com/wwarthen/RomWBW/tree/dev/Source/Forth

The file in particular is camel80,azm which uses .phase/.dephase

The build code is

zx zsm =camel80.azm -/l
zx link -CAMEL80.BIN=CAMEL80

Forth.zip

Link output from 41
LINK 1.31

ABSOLUTE 0000
CODE SIZE 1700 (0100-17FF)
DATA SIZE 0000
COMMON SIZE 0000
USE FACTOR 00

Link output from 42
LINK 1.31

ABSOLUTE 16B5 (0000-16B4)
CODE SIZE 1700 (0100-17FF)
DATA SIZE 0000
COMMON SIZE 0000
USE FACTOR 00

Please let me know if you need any more information.

Regards Phil

Path to Linux utilities

ZXCC = zxcc

Path to CP/M utilities

ZSM4 = ../../Tools/cpm/zsm4.com
DRLINK = ../../Tools/cpm/drlink.com
TKB = ../../Tools/cpm/tkb.com

It seems to me that the INT routine does not parse correctly the string of characters, enabling some annoying errors to escape unnoticed.

Example: the line:

ld hl,1v2

is assembled as:
0000' 21 00FC ld hl,1v2

,without indicating the obvious syntax error: 1v2 is not a valid number! and 0FCH is not the correct value !

This way, you risk assembling a source containing such syntax errors without being warned about these errors!

The M80 assembler by Microsoft has the assembler directive

    ds      35,"A"

so that 35 bytes are reserved and are initialized with the value of "A" (or 65 or 41h).

zsm4 v4.7 only accepts the first part of the directive

    ds      35

but not a value for initialization.

I would like to request, that zsm4 can also initialize the reserved bytes with an arbitrary value.