Skip to content
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

Move usage annotations from values to traces #208

Merged
merged 99 commits into from
Sep 15, 2019
Merged

Move usage annotations from values to traces #208

merged 99 commits into from
Sep 15, 2019

Conversation

rolyp
Copy link
Collaborator

@rolyp rolyp commented Sep 15, 2019

No description provided.

rolyp and others added 30 commits September 3, 2019 13:26
@rolyp
Some attempt at fixing watching.
04a02ce
@rolyp
Drop Annotated type from identifiers in environments and traces. Add …
ea36d3c 
…Annotated type to Expl.
@rolyp
Move karma-webpack workaround plugin to app config.
881f6b8
@rolyp
Closures and primitives in traces now have (empty) explanations to ca…
176ef73 
…rry their usage information.
@rolyp
ExplValue -> Expl.
56acb49
@rolyp
Start migrating tests: arithmetic.
50e82f8
@rolyp
Concentrate on Fwd tests: filter.
5c0e9bd
@rolyp
Progress on migrating fwd tests.
0e7460e
@rolyp
Fwd tests migrated.
53eefcc
@rolyp
Oops, forgot bar-chart.
25811ed
@rolyp
Delete old fwd tests.
9649938
@rolyp
Can't migrate bwd tests until I fix Var traces.
5b90438
@rolyp
Match.apply takes Expl_.
a870021
@rolyp
zipWith; have fieldExplValues return Expl_<Value>, not [Expl, Value].
b5a35aa
@rolyp
Push Expl_ rather than Annotated<Value> through match code...
3f5608f
@rolyp
Got a cyclic dependency between Expl_ and DataValue to break :-/
8808b53
@rolyp
fieldValues -> children.
ffe193e
@rolyp
Extract DataValue construction machinery from eval; reuse in true_/fa…
9ae6b94 
…lse_ helpers to ensure that primitive operations make well-formed explained values.
@rolyp
MatchPrefix now maintains Expl values.
b8ede23
@rolyp
Weird..tried switching Env over and got infinite loop...one step at a…
5d10503 
… time..
@rolyp
Rename v to tv...
7dae300
@rolyp
change return type of get...
5c540c7
@rolyp
Now fix up type of tv field of Env.
a00ae2c
@rolyp
Memoise setall as (temporary?) solution to trace-annotation-clearing …
b3c6e86 
…problem. Might be moot once we move away from datatype-generic annotations anyway.
@rolyp
Propagate into subcomputations via traces, not values...
19e1bb3
@rolyp
match_fwd/bwd -> apply_fwd/bwd.
6a5b3f3
@rolyp
Summary of problem: when I setall annotations on the root env to true…
9864115 
…, I don't recurse into the __expl object of any DataValues I encounter.
@rolyp
Can hack that in as special behaviour for now, but need to revisit id…
186363e 
…ea that Expls and Values are bundled together, I think.
@rolyp
Delete unused type parameter.
c829366
@rolyp
Fwd slicing plausibly working through traces.
bb170b5
@rolyp
Onto simplifying test helpers.
d192f24
@rolyp
expl_needed/notNeeded.
a5e9ae6
@rolyp
Some progress on purging _expl field from DataValue and separating va…
0491c1c 
…lues from traces again.
@rolyp
Ok, node navigation is now completely type-unsafe - I've lost all the…
a854333 
… benefit of mapping object-language types to ES6 classes :((((
@rolyp
To be fair, I had lost that anyway, as soon as I move annotations fro…
9c84ed5 
…m values to traces - because I can no longer just interact directly with a value, I need the related metadata.
@rolyp
Remaining problems are in GraphicsRenderer, which is the only 'typed'…
dfe6c67 
… interface to the language. Tests are less heavily reliant on types. Wonder if we can pull Expl_ out of the main system entirely.
@rolyp
Running with the idea of separating out Expl from DataValue. Typed in…
228efe7 
…terface to this setup is only required for app (although tests could be much nicer too).
@rolyp
EXPERIMENT: switch GraphicsRenderer to use Cursor interface for now.
8f867bd
@rolyp
Better idea: drop idea of (imperative) cursor. In an imperative langu…
ffdad0b 
…age, it's mostly overkill.
@rolyp
Switch GraphicsRenderer to (immutable) ExplCursor interface.
10933d3
@rolyp
Factor joinα, meetα through setα.
22cb861
@rolyp
Some renaming.
30dfc0b
@rolyp
Better use of types for annotations.
f6afc23
@rolyp
Start capturing slice deltas. Factor all annotation setting through s…
e3e9eb5 
…etter.
@rolyp
Ah. ExplCursor.notNeed was setting to true, not false. Now there are …
5ea2a8b 
…6,174 annotations set during the bar-chart forward-slicing test.
@rolyp
Rename.
7cc9b6e
@rolyp
Progress on making (expr) cursor immutable.
eb41ba5
@rolyp
Fixes.
ccf6311
@rolyp
Weirdly had tabs instead of spaces...
fc5ddb7
@rolyp
Immutable cursor.
07da6f7
@rolyp
Remove cursor fields from FwdSlice/BwdSlice.
0995808
@rolyp
Remove cursor fields from FwdSlice/BwdSlice.
5a781bf
@rolyp
Consolidate cursors.
1655cf6
@rolyp
Consolidate non-terminal trace forms.
c7e8651
@rolyp
Implement explChildren in terms of explChild.
0302dff
@rolyp
Not going to sweat the point about primitives being terminal forms ju…
5e010e7 
…st yet.
@rolyp
Minor corrections to GraphicsRenderer. Fix ExplCursor.to to assert cl…
b87eeaa 
…ass.
@rolyp
Yah, need to clear delta map in app before each bwd slice, since we a…
b9d9d92 
…dd about 2.5 million entries per click...
@rolyp rolyp merged commit d537773 into master Sep 15, 2019
@rolyp rolyp deleted the move-annotations branch September 15, 2019 17:25
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
@rolyp