B<nl>=2;

caused the first P<> rule to fail.  Had to fix it by adding

P<dummy>=3;

and never calling it.

Eventually worked more cleanly when I included all 3 B<> rules,

B<EOF> = 0;
B<CH> = 1;
B<NL> = 2;

even though only EOF and NL were used.


The skeleton-Makefile still had a reference to mnemosyne which I had to
remove - dependencies on both mnemosyne.h and mnemosyne.o which had to be
commented out. (and the mnemosyne.o from the link)



Looks like excess spaces are not being slipped despite WHITESPACE_WANTED in the .h file


Default build doesn't handle '%' stropping in Imp so temporarily I've set
keywords to be lower case and variables to be upper case.  Need to add proper
stropping code.  Compare to algol60 parser?


Discovered when writing a parser of .OBJ files for the Vectrex 3D project: (20250614)

A '{' in a string in the initial (and possible later) code section of a .g file
throws off the bracket matching and causes an error at the 'E' instruction.


===============

parser:
  in the {} C code attached to a rule, I need to be able to write code that is executed during the building of the G_ tuple tree as well as in the building of the P_ tuple tree.  Should I use ifdefs or some special syntax?

My documentation of the demandload_ch() calls in the regexp code is wrong.  Redo it after rereading the source again.  It's more complex and has more levels including the atom() array.

imp80 to C: handling of "!" comments - can it be done in the parser rather than during line reconstruction to handle the A ! !B! ! C ambiguity.

(there may have been some other note that I wanted to make and forgot before I could get to a keyboard :-( )

takeon:
add a nullable() test for regexp grammar rules - try test a rule against a null string and see if it matches!

c:
handling typedef comments and int16_t declarations in recent tests - not sure if
the problem is in the header files I'm using, or in my C parsing code.  (That was why
I grabbed a copy of diet libc and put it in the c/include subdirectory)

Which program was it that I was running the C parser on? Was it just the imp80 to C
translator or was there some other C that I needed to parse as well?