that typedef 'short'. 'char' (which was previously used) because char
may be unsigned and ((char)EOF) != EOF if that is the case. That was
causing the (char)EOF (0xff) pushed back in main to be interepreted as
a character, and, in some cases, to be written to the output. 'short'
was used rather than 'signed char' because if the latter is used,
0xff characters in the input would confuse m4. (No point in introducing
(more?) 8-bit lossage.)
quotes. E.g. if left quote was "``" and right quote was "''", then
"``foo ``bar'' baz''" would yield "foo `bar' baz" when it should yield
"foo ``bar'' baz".
it uses pbstr() instead of doing the characters individually (in
reverse order) with putback().
Fix dodiv() so that divert(<out-of-range stream>) doesn't cause "divnum" to
return 0, but rather the stream number specified to divert(). Macros
like PUSHDIVERT/POPDIVERT rely on this. The new behaviour is compatible
with vendor m4 implementations.
in `The M4 Macro Processor', by Kernighan and Ritchie.
* `|' and `||' are both logical or. There is no binary or.
* `&' and `&&' are both logical and. There is no binary and.
* `!' has a much higher precedence.
* The equal and relation operators have the same precedence.
* `**' is implemented, and `^' is for exponentation (not xor).
* Unary `+' in implemented.
I've coded exponentiation to return what appears to be the same results
that other versions of m4 return.
bug:if there are multiple files on the command line, file index is not
reset properly, ie. it falls off the side. sigh. also, somewhere along
the line, i again lost the proper m4wrap/autodiversion semantics
[sorry, eric. :-|] but fixed with this patch.
cgd