allow negative numbers in #emit & changes to public variable behaviour

[YashasSamaga]

allows negative numbers in #emit

fix for #125

[VVWVV]

Yes, you are allow to use negative numbers in #emit, but this is only for integer numbers, not floating.
For example, I did this: VVWVV@880fd10

[YashasSamaga]

I don't think someone will ever need to pass a rational value to an assembly instruction.

I am not sure how the compiler stores floating point values but amx_ctof seems to do a simple direct cast...

#define amx_ftoc(f)   ( * ((cell*)&f) )   /* float to cell */
#define amx_ctof(c)   ( * ((float*)&c) )  /* cell to float */

so what you did there (OR operation) should be correct.

[Y-Less]

I guess they would if they wanted to do:

#emit PUSH.C 0.5
#emit PUSH.C 7.7
#emit PUSH.C 8
#emit SYSREQ.C floatmul

But that seems more likely to cause confusion and mistakes since there is no tag checking so a stray . won't be flagged and will vastly alter the result.

[Zeex]

@YashasSamaga Mind fixing Appveyor build and indentation?

[YashasSamaga]

1. AppVeyor requires the code to abide by the C89 declaration rules. Moved the s2 declaration to the start of the block.

2. Indentation Fixed. I have a doubt with the indentation for line 755 in sc6 ('else'). I am not sure how you prefer to indent else with single statement. You can edit it.

[Zeex]

Great, thanks.

Yes, Visual Studio / MSVC doesn't support C99, and in fact the whole codebase is written in C89, so I try to stick to it too for consistency (I think it was originally done for maximum portability).