Specifying literals size about customasm HOT 5 CLOSED

This can be done with asserts, but I'd like it if there was shorter syntax that had the same effect.

assert(imm >= 0)
assert(imm <= 0xFF) ; assert that this unsigned(!) byte is within range.

assert(imm <=  0x7f)
assert(imm >= !0x7f) ; assert that this signed(!) byte is within range.

from customasm.

Oh, yeah! This was definitely in the plans. My planned syntax for this would be:

example {value: u8} -> 0x10 @ value[7:0] ; unsigned 8-bit constraint, accepts 0x00 through 0xff
example {value: s8} -> 0x10 @ value[7:0] ; signed 8-bit constraint, accepts -0x80 through 0x7f

(Perhaps even allowing you to suppress the [7:0] slice syntax)

These would throw errors similarly to the asserts moonheart08 posted above. But right now, you must use explicit asserts to get this behavior. I'll try to work on this feature as soon as possible!

from customasm.

Done! You can now use this syntax, and even elide slices on argument usage. There is a third parameter type too:

example {value: u8} -> 0x10 @ value ; unsigned 8-bit constraint, accepts 0x00 through 0xff
example {value: s8} -> 0x10 @ value ; signed 8-bit constraint, accepts -0x80 through 0x7f
example {value: i8} -> 0x10 @ value ; signed/unsigned 8-bit constraint, accepts -0x80 through 0xff

And, of course, you can use whichever number of bits you like, such as {value: s32} or {value: u3}.

from customasm.

This can be done with asserts, but I'd like it if there was shorter syntax that had the same effect.

assert(imm >= 0)
assert(imm <= 0xFF) ; assert that this unsigned(!) byte is within range.

assert(imm <=  0x7f)
assert(imm >= !0x7f) ; assert that this signed(!) byte is within range.

I didn't think of this one, that's pretty clever :)

Oh, yeah! This was definitely in the plans. My planned syntax for this would be:

example {value: u8} -> 0x10 @ value[7:0] ; unsigned 8-bit constraint, accepts 0x00 through 0xff
example {value: s8} -> 0x10 @ value[7:0] ; signed 8-bit constraint, accepts -0x80 through 0x7f

(Perhaps even allowing you to suppress the [7:0] slice syntax)

These would throw errors similarly to the asserts moonheart08 posted above. But right now, you must use explicit asserts to get this behavior. I'll try to work on this feature as soon as possible!

Oh that's great! This syntax is far better that I proposed.
If you need help on specific implementation points, I'd be happy to help if I can :)

from customasm.

Oh that was quick :D
Thanks a lot for this improvement!

from customasm.