by jmillikin on 7/13/25, 1:12 AM with 42 comments
by adinisom on 7/13/25, 5:04 AM
typedef struct blinky_state {
size_t pc;
uint64_t timer;
... variables that need to live across YIELDs ...
} blinky_state_t;
blinky_state_t blinky_state;
#define YIELD() s->pc = __LINE__; return; case __LINE__:;
void blinky(void) {
blinky_state_t *s = &blinky_state;
uint64_t now = get_ticks();
switch(s->pc) {
while(true) {
turn_on_LED();
s->timer = now;
while( now - s->timer < 1000 ) { YIELD(); }
turn_off_LED();
s->timer = now;
while( now - s->timer < 1000 ) { YIELD(); }
}
}
}
#undef YIELD
Your company is probably using hardware containing code I've written like this.
What's especially nice that I miss in other languages with async/await is ability to mix declarative and procedural code. Code you write before the switch(s->pc) statement gets run on every call to the function. Can put code you want to be declarative, like updating "now" in the code above, or if I have streaming code it's a great place to copy data.
by astrobe_ on 7/13/25, 8:02 AM
That's because you need a state machine when your control flow is not linear. They are represented by graphs, remember? This is actually a case where using gotos might be clearer. Although not drastically better because the main problem is that written source code is linear by nature. A graph described by a dedicated DSL such as GraphViz has the same problem, although at least you can visualize the result.
But control flow is only one term of the equation, the other being concurrency. One typically has more than one state machine running; sometimes one use state machines that are actually essentially linear because of that. Cooperative multitasking. I would question trying to solve these two problems, non-linearity and concurrency. Sometimes when you try too hard to kill two birds with one stone you end up with one dead bird and a broken window.
One lecturer of the conference announced earlier [1] made that point too that visualization helps a lot, and that reminded me of Pharo's inspection tools [2]. Seeing what's going on under the hood is more important that one usually thinks.
One issue with state machines is that they are hardly modular: adding a state or decomposing a state into multiple states is more work than one would like it to be. It is the inverse problem of visualization: what you draw is what you code. A good tool for that would let the user connect nodes with arrows and assign code to nodes and/or arrows; it would translate this into some textual intermediate language to play nice with Git, and a compiler would transform it to C code for integration in the build system.
[1] https://bettersoftwareconference.com/ [2] https://pharo.org/features
by mikepurvis on 7/13/25, 1:32 AM
That said, if I was stuck rolling this myself, I think I’d prefer to try to do it with “real” codegen than macros. If nothing else it would give the ability to do things like blocks and correctness checks, and you’d get much more readable resulting source when it came to stepping through it with a debugger.
by userbinator on 7/13/25, 4:23 AM
That tends to just make the project eventually implement an approximation of one... as what appears to have happened here.
How I'd solve the given problem is by using the PWM peripheral (or timer interrupts if no PWM peripheral exists) and pin change interrupts, with the CPU halted nearly 100% of the time. I suspect that approach is even simpler than what's shown here.
by syncurrent on 7/13/25, 7:27 AM
by Neywiny on 7/13/25, 1:34 AM
It's my same argument that there should be no maximum number of lines to a function. Sometimes, you just need to do a lot of work. I comment the code blocks, maybe with steps/parts, but there's no point in making a function that's only called in one place.
But anything is better than one person I met who somehow was programming without knowing how to define their own functions. Gross
by user____name on 7/13/25, 11:25 AM
by jonhohle on 7/13/25, 7:17 PM
by throwaway81523 on 7/13/25, 1:40 AM
by codr7 on 7/13/25, 1:49 PM
This is an alternative I wrote for my C book:
by moconnor on 7/13/25, 8:04 AM
by Asooka on 7/13/25, 6:40 PM
by Nursie on 7/13/25, 5:41 AM
I’m not sure this is so much hacking as an accepted technique from the old-old days which has somewhat fallen out of favour, especially as C is falling a little outside of the mainstream these days.
Perhaps it’s almost becoming lost knowledge :)
by webdevver on 7/13/25, 9:16 PM
by Agyemang on 7/13/25, 9:56 AM
by joshlk on 7/13/25, 12:06 PM