#: 8927 S12/OS9/68000 (OSK)
    28-Dec-90  10:28:06
Sb: #8926-#68000 ASM Language
Fm: Ed Gresick 76576,3312
To: Ed Gresick 76576,3312 (X)




The offending code was the 'bne' - it should have been 'bmi' - he wanted to
test the 'n' flag.  Try to find it in about 12,000 lines of poorly commented
code.  So, is there a way for the assembler to pick the correct test?

Please stay away from C-like constructs - they can get messy and are not always
clear.  There is no reason to stay with traditional mnemonics per se. Yes, in
some cases what was selected may turn out to be the best but, more often than
not, it' not the best today.  One question, who is your intended user?  If
you're targeting the experienced C/assembler programmer, I suspect you'll get a
lot of resistance.  On the other hand, the new or casual user (really one and
the same) will appreciate an assembler that is easy to use and does not require
all kinds of 'tools' to increase programmers' productivity. And I fall into
this category.

Can the assembler use ordinary English words?  In order to ease the
requirements of the parser, a preprocessor might be used to tokenize these
words and if the tokens are sensible, users will probably learn them.

Another task the assembler should do is select whether the branch or jump is a
long or short.  Get tired of putting '.s's in only to get back error messages
telling me I can't do that - especially if its because I added some code.





There are 2 Replies.