New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix #6444: pruner issues with reference stealing functions #6446
Conversation
but to acquire the reference.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for the patch, think this will fix the issue. Is there a test that can be added to demonstrate?
|
||
|
||
Quirks | ||
------ | ||
Quirks of old optimization pass |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Qualify "old" again? Also perhaps this should be called "legacy"?
Co-authored-by: stuartarchibald <stuartarchibald@users.noreply.github.com>
about the renaming
Close/reopen was because Azure got irretrievably stuck due to a deprovisioning request. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for the fixes, a typo and a query else looks good.
Since the pre-0.52.0 `refcount optimization pass <nrt-refct-opt-pass_>`_ | ||
requires LLVM function optimization pass, the pass works on the LLVM IR as | ||
text. The optimized IR is then materialized again as a new LLVM in-memory | ||
bitcode object. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm not sure what this is trying to convey? Is it that the old pass works on textual IR, so the bitcode repr has to be optimised by LLVM whilst as bitcode, then converted to text, optimisations run on this text and then materialised back to bitcode?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes. I agree it wasn't written clearly. I didn't actually change it though, just shuffled the line-breaks around to fit.
(the lazy part of me doesn't want to change it since it's going away very soon)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I agree on the basis that only core developers adjusted it and it's being retired. Apart from adding the missing "the" above, let's just leave it.
Co-authored-by: stuartarchibald <stuartarchibald@users.noreply.github.com>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Few typos else looks good, thanks for the fixes.
Since the pre-0.52.0 `refcount optimization pass <nrt-refct-opt-pass_>`_ | ||
requires LLVM function optimization pass, the pass works on the LLVM IR as | ||
text. The optimized IR is then materialized again as a new LLVM in-memory | ||
bitcode object. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I agree on the basis that only core developers adjusted it and it's being retired. Apart from adding the missing "the" above, let's just leave it.
Co-authored-by: stuartarchibald <stuartarchibald@users.noreply.github.com>
13b9569 fixes a typo. |
Assuming CI passes this is good to go, thanks for fixing this! |
/AzurePipelines run |
Azure Pipelines successfully started running 1 pipeline(s). |
Fix numba#6444: pruner issues with reference stealing functions
Fixes #6444