Fix floating point unary minus and plus
commit05c9b76131e9d0a2e1b011eeb70ad7897efc9084
authorMichael Matz <matz@suse.de>
Sun, 12 Jan 2014 03:44:27 +0000 (12 04:44 +0100)
committerMichael Matz <matz@suse.de>
Sun, 12 Jan 2014 03:44:27 +0000 (12 04:44 +0100)
tree6bb541490c316505c2a9ffccc39507057b4a6b1c
parent9c6ddbfe903445763405d2c8bdec916c8a20f105
Fix floating point unary minus and plus

negate(x) is subtract(-0,x), not subtract(+0,x), which makes
a difference with signed zeros.  Also +x was expressed as x+0,
in order for the integer promotions to happen, but also mangles signed
zeros, so just don't do that with floating types.
tccgen.c
tests/tcctest.c