al@escom.com (Al Donaldson) (10/12/90)
I thought I'd pass on the following things I observed about asld
and its 1.5 documentation. Note -- My asld is built from Unipress
source and has a minor local change that should not affect any
of the following items. However, your mileage may vary..
- as I reported earlier, the .org directive is broken.
When I use .org, asld generates an empty a.out file.
- I couldn't get the .even directive to work, either.
asld just seemed to ignore it.
- There's a couple of inaccuracies in the asld extended man pages:
First, the sto[bw] statement stores AL/AX in ES:DI instead
of DS:DI as commented.
Second, the arguments for the calli statement are reversed.
Section 9.1.13 shows it as "calli segment, offset" but it
should actually be "calli offset, segment". I seem to
remember that the same applies to the jmpi instruction.
- Section 1.1.12 made a big deal about the difference between
8-bit constants ("*FF") and 16-bit constants ("#FFFF").
I found this didn't make much difference; I could use #
for everything, and asld figured out the context.
The extended man pages are pretty sparse. I found it absolutely
necessary to have the Intel Programmer's Reference Manual
available in addition to the man pages. The error messages
from asld (syntax problems) are fairly good, though.
I was quite disappointed in the asld documentation from Unipress.
I expected it to be a lot more thorough with respect to the
instruction set and all its intricacies. However, it spent most
of its pages describing the EM instruction set and, consequently,
wasn't much help in writing an assembler program. Maybe I'm
missing the point of this documentation.
Since the "New and Improved" version of asld is in the works,
these comments are probably not of much use. I certainly hope
that the assembler directives such as .org and .even will be fixed.
Al