Skip to content

Gc enabled hashtable fsduo #1866

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.

Sign up for GitHub

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

Jump to bottom
Merged
merged 48 commits into from
Nov 18, 2025
Merged

Gc enabled hashtable fsduo #1866

merged 48 commits into from
Nov 18, 2025

Conversation

@andrew-johnson-4
Copy link
Collaborator

@andrew-johnson-4 andrew-johnson-4 commented Nov 18, 2025 •
edited
Loading

Describe your changes

Features:

  • it is now a hard error if any phi type is ever dropped
  • typechecking can now be called multiple times on the same term
  • many many new tests for phi features
  • starting to create tests for phi + GC promises

The new promises testing framework has allowed for development to continue with few regressions despite fairly large rearchitecting of GC / phi interactions. Those features will be in promises soon, but first we need to address the technical limitations that even make those features possible.

In this commit, there is only 1 single regression. Looking back at that code, I am not sure how it was even working based on how broken some of its dependencies were.

Issue ticket number and link

#1775

Checklist before requesting a review

  • [ x ] I have performed a self-review of my code
  • [ x ] If it is a new feature, I have added thorough tests.
  • [ x ] I agree to release these changes under the terms of the permissive MIT license (1).
  1. https://github.com/andrew-johnson-4/lambda-mountain/blob/main/LICENSE

@andrew-johnson-4
open up failing test case, now there is a double-free somehow
0d38f4d
@andrew-johnson-4
set[] needs to intentionally leak that variable
1da2203
@andrew-johnson-4
nope, misunderstanding. retain/release were correct as default
cd42289
@andrew-johnson-4
dropping phi types needs to be a hard error
f7c6a3e
@andrew-johnson-4
focus on this bug
092e866
@andrew-johnson-4
No more dropped phi variables
21acfb1
@andrew-johnson-4
lets just eliminate dropped phi and call it a day
a4eb76b
@andrew-johnson-4
dead-on-arrival linear variable introduced but not eliminated
092821b
@andrew-johnson-4
the correct problematic linear variable is showing up in release now
e6813c6
@andrew-johnson-4
Ok, does this fix missing phi?
5bb91ee
@andrew-johnson-4
Ok, pinpointed the problem now need to think of a better solution
cbd1db3
@andrew-johnson-4
the moved variable is not dead, the live variable is
8ba7c5d
@andrew-johnson-4
Vector OK!
0f6ada0
@andrew-johnson-4
remove logging
70c3969
@andrew-johnson-4
bisecting to find, close
18cfc4d
@andrew-johnson-4
wow that was rotten, ascript-force has never been working
8b1a781
@andrew-johnson-4
OK
e0495d0
@andrew-johnson-4
there are still some missing phi types
b8e1ee9
@andrew-johnson-4
phi fresh working, MustNotFresh not so much
a3880ac
@andrew-johnson-4
reorganize lib2 with a lower entry point
cf29f69
@andrew-johnson-4
fixed a bug that needs weakening, but now a type is missing in C backend
649a2b2
@andrew-johnson-4
OK, some phi working
abb2926
@andrew-johnson-4
point all phi tests to baremetal
368ebe5
@andrew-johnson-4
must retain
4ef803e
@andrew-johnson-4
the test was wrong
d5bcc9b
@andrew-johnson-4
error moved slightly
823bee3
@andrew-johnson-4
progress
d33648b
@andrew-johnson-4
found a way for dead-on-arrival phi variables to not impact phi looku…
8880762 
…p for live variables but still be present for dead ones
@andrew-johnson-4
l
5bd706c
@andrew-johnson-4
where is this release coming from
642279e
@andrew-johnson-4 andrew-johnson-4 merged commit 86a0fb3 into Lambda-Mountain-Compiler-Backend:main Nov 18, 2025
1 check passed
@andrew-johnson-4 andrew-johnson-4 deleted the gc-enabled-hashtable-fsduo branch November 18, 2025 01:24
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@andrew-johnson-4