Uh oh!
There was an error while loading. Please reload this page.
GH-118095: Handle RETURN_GENERATOR in tier 2 - #118180
Conversation
gvanrossum
left a comment
There was a problem hiding this comment.
Why not add a test? E.g.
deftest_return_generator(self):
defgen():
yieldNonedeftestfunc(n):
foriinrange(n):
gen()
res, ex=self._run_with_optimizer(testfunc, 20)
self.assertIsNotNone(ex)
self.assertIn("_RETURN_GENERATOR", get_opnames(ex))| get_code(_PyUOpInstruction *op) | ||
| { | ||
| assert(op->opcode == _PUSH_FRAME || op->opcode == _POP_FRAME); | ||
| assert(op->opcode == _PUSH_FRAME || op->opcode == _POP_FRAME || op->opcode == _RETURN_GENERATOR); |
There was a problem hiding this comment.
It's a bit inelegant that all places that treat _PUSH_FRAME and _POP_FRAME special now also have to check for _RETURN_GENERATOR. Not sure what to do about it. :-(
There was a problem hiding this comment.
Once #118095 is done we can add a flag for all uops that push or pop frames.
| ctx->frame->stack_pointer = stack_pointer; | ||
| frame_pop(ctx); | ||
| stack_pointer = ctx->frame->stack_pointer; | ||
| OUT_OF_SPACE_IF_NULL(res = sym_new_unknown(ctx)); |
There was a problem hiding this comment.
Am I correct that this is the only line that differs from _POP_FRAME? The duplication of so much code is unfortunate. At the same time I don't see a better solution. Maybe add a comment to both explaining they need to be kept in sync?
There was a problem hiding this comment.
A good candidate for refactoring, but I'd like to get #118095 done first so that we more easily see the common patterns.
| def testfunc(n): | ||
| for i in range(n): | ||
| gen() | ||
| res, ex = self._run_with_optimizer(testfunc, 20) |
There was a problem hiding this comment.
I'd assert the value of res as well.
markshannon
commented
Apr 25, 2024
The four failing tests are the usual suspects. |
Uh oh!
There was an error while loading. Please reload this page.