Allow longer strings

[Rangi42]

Right now MAXSTRLEN is 255. I think that a larger limit, like 1024 or even 4096, would be helpful.

• PRINTV/PRINTI/PRINTF emitting trailing newlines (PRINTV, PRINTI, and PRINTF emit a trailing newline #624) mean that PRINTT with STRFMT (Implement STRFMT and more printf-like format specifiers for string interpolation #646) (or just {formatted:interpolation}) will be the obvious choice for complex printing. These strings could get pretty long, especially if formatted values use wide padding.
• Now that macros can't be nested, the "official" solution to define one macro within another is by putting the whole inner MACRO ... ENDM definition in an EQUS. These are likely to be multi-line strings, so even if 255 is a reasonable length limit per line, the whole macro could be longer than that.

[ISSOtm]

What we really need is to remove that hardcoded limit. Raising the limit would cause lvalues to become stupidly large; note also that strings get strcpyd every time they change parser contexts. This is utterly stupid, and perhaps a source of performance loss.
Compared to malloc? I dunno. I believe there are known solutions to that, though...

[Rangi42]

With longer strings, test/asm/quine.asm could be reduced further to 373 bytes:

MACRO N
FOR R,1,_NARG+1
PRINT STRSUB("\n\"\\ NRT1,ABCDEFGILMnOPSU_+#()<>",\<R>,1)
ENDR
PRINTLN"\#"
ENDM
 N 19,10,12,6,21,4,5,1,15,21,6,4,6,9,8,9,25,5,10,6,16,26,8,1,22,6,17,5,7,4,23,7,6,23,24,11,28,2,3,20,3,2,3,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,2,9,3,30,6,31,9,8,29,1,14,5,13,6,1,22,6,17,5,7,18,5,2,3,27,2,1,14,5,13,19,1,4,5,4

However, \# would be 265 characters.

[ISSOtm]

By the way, according to some perf profiling, copying lvalues is currently the largest bottleneck in RGBASM by far. A WIP implementation of heap-stored strings can be found in my fork.