Macros apparently have higher precedence than numbers

[Y-Less]

#define xFF
new a = 0xFF;
printf("%d", a); // Prints 0

That probably shouldn't happen, but only if the symbol is a valid part of a number and used as such. I quite like this, that I saw in a recent include:

#define _Seconds * 1000
new a = 5_Seconds;
printf("%d", a); // Prints 5000

In that case, 5_ is not possibly a valid number so should be replaced; while things like 0x, 0b, or 5e could be so shouldn't be.

[Y-Less]

http://forum.sa-mp.com/showthread.php?p=3924669#post3924669
http://forum.sa-mp.com/showthread.php?t=640660

[VVWVV]

I guess this is not a bug. The preprocessor works first, and it does not analize the code. To fix the problem we need to rewrite the preprocessor.

[Y-Less]

I guess this is not a bug.

To fix the problem

Which is it? If I write:

#define xFF
new AxFF = 255;
new a = AxFF;
printf("%d", a); // Prints 255

That is fine, because macros don't match in the middle of symbols, so they probably shouldn't match in the middle of valid numbers either.

[VVWVV]

Sorry) Yeah, the preprocessor skips all non-alphabetic characters here: https://github.com/Zeex/pawn/blob/master/source/compiler/sc2.c#L1772-L1780

[Dayvison]

Don't skip if have a-z @ _ before the number

#define test -

0test
10test
_0test

Output:

0-
10-
_0test

[Y-Less]

_ and @ should be counted - they can start legitimate tokens. I literally have a macros called ___1 and @1.

[Dayvison]

Yeah is just a example, what i want to say is if as any character before number, don't match.

[Dayvison]

What i tried

#define hook%1(%2) forward UNIQUE_FUNCTION<@yH0%1@...>(%2);UNIQUE_FUNCTION<@yH0%1@...>(%2)
// Strip out extra spaces (nicely recursive)
#define @yH0%0\32;%1(%2) @yH0%1(%2)

#define tag:: tag_
hook tag::function()

Changed @yH_ to @yH0