Tomasz, is "\" a token or not?





According to documentation expression is valid and should be taken from next line.

So, 4 cases:
1. wrong in documentation and "\" as ";" must be not in list of special symbols +-/*=<>()[]{}:?!,.|&~#`\
2. there is a bug in fasmg itself contrary to the documentation

3. my example is bad and "\" work with eval in different example.
4. there is another way to include to source line symbo."\" after end of parsing stage?

Concerning documentation: the use of quotes disables tokenization for the contents of the string, the entire quoted string is always a single token. So even though backslash usually is a standalone token, it is not when it occurs inside a quoted string.

As for your actual problem: the arguments to EVAL define a string and only then this string is evaluated. The "\" can be a part of this string and then the line concatenation occurs within the data contained in it: You can use some concatenation macros to create long, multi-line strings for EVAL.

entry conditions:
Source code - minimal strict form of program that consist of only a sequences of tokens, where each token and sequence separated from each other with single separator symbol (for ex. whitespace and/or CRLF).
Source text - decorated form of program that allows comments,indentation (and so on...) to improve human readability.
Any character is perceived as a token if it can cross parsing stage from source text to source code and become part of source code, and than can be cathed with match directive as standalone one.

demonstration:
source:

output:

Proved:symbol "\" is trully token.

By the way, why behavior of "\" is so that only last "\" is line appender symbol? Why its behavior not like ";" with erasing all text in line that follows that symbol? What for symbol "\" is stayed as token?

The "\" in a middle of a line is left available for other uses (you can MATCH it, so you can use is for your own syntax when needed). Note that in fasm 1 it had other uses natively (as symbol escaping).