[Libre-soc-bugs] [Bug 849] binutils: parse macros in SVP64 and refactor code

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=849

--- Comment #26 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
(In reply to Dmitry Selyutin from comment #22)
> (In reply to Luke Kenneth Casson Leighton from comment #21)
> > (In reply to Dmitry Selyutin from comment #20)
> > bleuch, far too many random characters.
> 
> These characters are _not_ random.

i know: i mean, i'm counting the number, ignoring what they actually are, and
9 characters is exceeding a threshold i feel comfortable with.
(10 including /)

/ 1
sm= 3
1<< 3
%r3 3

total of 10, it's way too much.

> These characters come from the way this
> assembly work in PPC. My point is, if 1<<r3 is the most obvious and evident
> choice, the way it's described in GNU assembly is 1<<%r3, or via -mregnames
> flag. I suggest at least keeping this as an alias for u3 (which, in my
> opinion, is not obvious at all).

agreed: i just can't think of a good short mnemonic.

does 1<<%3 work?

how about dropping the 1 and using one "<"? sm=<%3

or "^"? sm=^r3 sm=^%3

would a special-case pattern in the lexer work (if "^" has not
been used),

i am wondering if the unary-bit form is going too far in the
macro substitution.  if it should be excluded as a special
case.  if you want sm=1<<r3 you *know* you want it, and you
cannot use 1<<r10 or 1<<r31 so macros will not help you.

-- 
You are receiving this mail because:
You are on the CC list for the bug.