retro html formatted by nom
comments/suggestions to: mjb at nomlang dot org
When you tell a story, try to have point! "Dumb and dumber"
Set the automatic character counter register to zero (i.e. to nochars*)
Everytime that a character (Unicode we hope) is read from the input stream,
(by the nom commands read or while or whilenot or
until ) the character counter register in the pep virtual machine
is incremented by one.
The “nochars” command sets the automatic character counter to zero. This is useful for making the character number relative to the line number. This is often more helpful than having a character count that starts from the beginning of the input stream. For example at the beginning of a script, we can put
read; [\n] { nochars; } print; clear;
chars; "60" { clear; add "\n"; print; }
clear;
read;
"\n" { nochars; }
"z" {
add "\n[Found a 'z' at line: "; lines; add ", char: "; chars;
add "]\n";
}
print; clear;
The pep machine also has a built-in automatic line counting register
called the lines register. It is automatic because it is
automatically incremented every time a newline character is encountered
in the input-stream. Like the chars register it cannot be
decremented but it can be set = 0 with the nolines command.
The reason that these 2 registers exist (and are automatic) is that pretty much every language compiler or “syntax checker” or data-format checker will try to provide a somewhat helpful error message when an error is encountered in the source. That means, as far as possible, giving a line and (relative) character number as a pointer to where the error occured.
In recent nom scripts I have actually developed the technique of creating
and push ing an error token on the stack, and that token
will print the message that is in the token “attribute” (stored in the
corresponding cell of the machine tape ) This avoid duplicating
the line and character printing code in every error message.