I wanted a program which ran once and halted, spinning forever until explicitly restarted by code. So I wrote this:
.program test
nop
.wrap_target
.wrap
And I got this:
#define test_wrap_target 1
#define test_wrap 0
This is technically correct. After the instruction at address 1 (the nop) is complete, the program counter is moved to... the address after. Which is correct because that's where I put the .wrap_target, but is meaningless.
What's happened here is that .wrap_targetlooks like it's an instruction which gets jumped to, but it's not. It just labels the next instruction, and the end of the program is not an instruction. This was a mistake on my part, but ideally the assembler should have caught this and complained that what I was trying to do wasn't valid. I only realised what was going on once I looked at the generated code.
Feature request: please spot when I do this and error out.
I wanted a program which ran once and halted, spinning forever until explicitly restarted by code. So I wrote this:
And I got this:
This is technically correct. After the instruction at address 1 (the nop) is complete, the program counter is moved to... the address after. Which is correct because that's where I put the
.wrap_target, but is meaningless.What's happened here is that
.wrap_targetlooks like it's an instruction which gets jumped to, but it's not. It just labels the next instruction, and the end of the program is not an instruction. This was a mistake on my part, but ideally the assembler should have caught this and complained that what I was trying to do wasn't valid. I only realised what was going on once I looked at the generated code.Feature request: please spot when I do this and error out.
Thanks!