Add support for unit tests via $[test] attribute
#928
Conversation
Still TODO: * Documentation * Verify the type is `unit -> unit` for functions where this is applied. * Tests * Other backends?? * Output formatting isn't quite right; not sure why.
Test Results 12 files ± 0 24 suites ±0 0s ⏱️ ±0s For more details on these errors, see this check. Results for commit 59d29e3. ± Comparison against base commit e7f1262. ♻️ This comment has been updated with latest results. |
|
Yes, that looks like the right way to go. |
| | None -> () | ||
| end; | ||
| (* Check $[test] functions have type unit -> unit. *) | ||
| (* TODO: Does the annotation go on the type declaration or the function definition? *) |
There was a problem hiding this comment.
I think in general we put these annotations on the val type declarations. That's where $[property] and $[counterexample] are expected to be.
If you have a function with an inline annotation:
$[annot]
function f(x : int) -> unit = ...
it'll get expanded to
$[annot]
val f : int -> unit
$[annot]
function f(x) = ...
with the annotation on both.
| let end_extern_cpp = separate hardline (List.map string [""; "#ifdef __cplusplus"; "}"; "#endif"]) in | ||
| let hlhl = hardline ^^ hardline in | ||
|
|
||
| (* TODO: Formatting here isn't quite right. What are the arguments to jump? *) |
There was a problem hiding this comment.
It's explained a bit here https://github.com/fpottier/pprint/blob/767ef4ae0d5de95b21e0d650e8bb752830d75a47/src/PPrint.mli#L163
|
I think this change has probably broken this PR: #960 I can fix if you like? |
|
Ah that's a nice change. Don't worry, I'll try and figure it out. You have more important things to do! |
…ms-project#928) From section 12.2.1 of the priv spec: > For implementations that make satp.MODE read-only zero (always Bare), attempts to execute an SFENCE.VMA instruction might raise an illegal-instruction exception.
2 participants
Draft for #919, just checking if this approach makes sense.
Still TODO:
unit -> unitfor functions where this is applied.