integrate resolver's lambda unraveling into semantic translator pass #6295
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This commit integrates the lambda unraveling technique used in the resolver into the first-pass of the semantic translator. This is possible by indexing the lambda permutation using decl IDs instead of call tags based on the intuition that the only thing that can give rise to variation in unraveled combinations is the declared function, i.e., at the outermost call site, each variation of lamba-parameterized function depends only on the function passed in, which are uniquely defined by their ID not their unraveled instances. This means the evaluator can simply traverse an expression to determine if it's constant and then dagen it. There is no need anymore to re-enter the resolver. This is also the key step needed to integrate type checking into the translator, which will then allow us to utilize dataflow-analyzed fused types from pipe queries in schema-column resolution for SQL queries that include pipe subqueries. This will be forthcoming in a subsequent PR.
nwt
approved these changes
Oct 13, 2025
Member
nwt
left a comment
nwt
left a comment
There was a problem hiding this comment.
I'm still working on understanding the nuance in here but it looks good at a high level and it clearly works so 
.
compiler/semantic/expr.go
Outdated
Comment on lines
100
to
101
| id := e.Name | ||
| if boundID, _ := t.scope.lookupFuncDeclOrParam(e.Name); boundID != "" { |
Member
There was a problem hiding this comment.
Nit: Using ID here would make the data flow a little clearer.
Suggested change
| id := e.Name | |
| if boundID, _ := t.scope.lookupFuncDeclOrParam(e.Name); boundID != "" { | |
| id := e.Name | |
| if boundID, _ := t.scope.lookupFuncDeclOrParam(id); boundID != "" { |
compiler/semantic/resolver.go
Outdated
Comment on lines
38
to
39
| // at a built-in // or function declaration and they can't be modified | ||
| // *(only passed by reference), the variants are determined by decl ID |
Member
There was a problem hiding this comment.
Nit:
Suggested change
| // at a built-in // or function declaration and they can't be modified | |
| // *(only passed by reference), the variants are determined by decl ID | |
| // at a built-in or function declaration and they can't be modified | |
| // (only passed by reference), the variants are determined by decl ID |
compiler/semantic/resolver.go
Outdated
| } | ||
|
|
||
| type lambda struct { | ||
| param string // parameter name this lambda arg appeared asa |
Member
There was a problem hiding this comment.
Nit:
Suggested change
| param string // parameter name this lambda arg appeared asa | |
| param string // parameter name this lambda arg appeared as |
compiler/semantic/resolver.go
Outdated
Comment on lines
134
to
135
| params := idsToStrings(d.lambda.Params) | ||
| for _, param := range idsToStrings(d.lambda.Params) { |
Member
There was a problem hiding this comment.
Nit:
Suggested change
| params := idsToStrings(d.lambda.Params) | |
| for _, param := range idsToStrings(d.lambda.Params) { | |
| params := idsToStrings(d.lambda.Params) | |
| for _, param := range params { |
compiler/semantic/expr.go
Outdated
Comment on lines
799
to
810
| func (t *translator) resolveCall(n ast.Node, id string, args []sem.Expr) sem.Expr { | ||
| if isBuiltin(id) { | ||
| // Check argument count here for builtin functions. | ||
| if _, err := function.New(super.NewContext(), id, len(args)); err != nil { | ||
| t.error(n, err) | ||
| return badExpr() | ||
| } | ||
| return sem.NewCall(n, id, args) | ||
| } | ||
| return t.resolver.mustResolveCall(n, id, args) | ||
| } | ||
|
|
Member
There was a problem hiding this comment.
Nit: Putting this immediately before mustResolveCall would improve readability a bit.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This commit integrates the lambda unraveling technique used in the resolver into the first-pass of the semantic translator. This is possible by indexing the lambda permutation using decl IDs instead of call tags based on the intuition that the only thing that can give rise to variation in unraveled combinations is the declared function, i.e., at the outermost call site, each variation of lamba-parameterized function depends only on the function passed in, which are uniquely defined by their ID not their unraveled instances.
This means the evaluator can simply traverse an expression to determine if it's constant and then dagen it. There is no need anymore to re-enter the resolver.
This is also the key step needed to integrate type checking into the translator, which will then allow us to utilize dataflow-analyzed fused types from pipe queries in schema-column resolution for SQL queries that include pipe subqueries. This will be forthcoming in a subsequent PR.