Hello,
I am getting a lot of error message when running tests manually:

# ./qbe -t rv64 test/sum.ssa
qbe:test/sum.ssa:10: invalid instruction

        %c =w cslew %n1, 0

but it is not consistent to cslew

# ./qbe -t rv64 test/_fix3.ssa
qbe:test/_fix3.ssa:11: invalid instruction

function w $test() {
@start
         %x =w copy 100
         %s =w copy 0
@l
         %c =w cslew %x, 10 <-- passes
         jnz %c, @a, @b
@a
         %s =w add %s, %x
         %x =w sub %x, 1
         jmp @c    <-- line 11



I do not have issues with the same environment on qbe-1.0

# ./qbe -t rv64 test/sum.ssa
.text
.globl sum
sum:
	sd fp, -16(sp)
	sd ra, -8(sp)
	add fp, sp, -16
	add sp, sp, -16
	li t0, 0
...


I am not able to run the tools/test.sh script because I am in a 
completely different environment than expected

# uname -a
Minix 10.0.2.15 3 1.6 i686
# pcc -v
Portable C Compiler 1.2.0.DEVEL 20190328 for i686-pc-minix3


Is there a reason some of these tests would fail for qbe-1.1 when run 
manually but succeed in qbe-1.0?

Thank you,
tim

-- 
No one is born hating another person because of the color of his skin,
or his background, or his religion.  People must learn to hate, and if
they can learn to hate, they can be taught to love, for love comes more
naturally to the human heart than its opposite.
		-- Nelson Mandela

My fault.  I realized it was the lexical parser issue that I had fixed 
to get qbe-1.0 to work on Minix 3.1.6.

All is good, sorry for the noise.

tim
-- 
You can play almost anything in between, but ultimately, you have to 
come back to 'The One.'

		-- Bootsy Collins

One thing though, I highly recommend you use the
master branch for your work. It's got important
bug fixes over 1.1.