-
Notifications
You must be signed in to change notification settings - Fork 12.6k
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 string literal completions when a partially-typed string fixes inference to a type parameter #48410
Conversation
@@ -22662,7 +22675,7 @@ namespace ts { | |||
const properties = getPropertiesOfObjectType(target); | |||
for (const targetProp of properties) { | |||
const sourceProp = getPropertyOfType(source, targetProp.escapedName); | |||
if (sourceProp) { | |||
if (sourceProp && !some(sourceProp.declarations, hasSkipDirectInferenceFlag)) { |
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.
Maybe it would be nice to skip this at least in non-services calls, but it would be a lot of plumbing to get checkMode
through. I can make a local in createTypeChecker
if it seems worthwhile.
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 pretty sure you could translate the CheckMode
to an InferenceFlag
if you'd like.
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.
inferTypes
doesn’t have access to the inference flags, so this doesn’t actually improve the plumbing situation 😕
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.
A dummy InferencePriority
that's unset by the start of inference then? 😅
@@ -570,6 +555,8 @@ namespace ts { | |||
getFullyQualifiedName, | |||
getResolvedSignature: (node, candidatesOutArray, argumentCount) => | |||
getResolvedSignatureWorker(node, candidatesOutArray, argumentCount, CheckMode.Normal), | |||
getResolvedSignatureForStringLiteralCompletions: (call, editingArgument, candidatesOutArray) => |
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.
Can't we just add an @internal
overload for getResolvedSignature
that allows CheckMode
to be passed in?
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.
editingArgument
is also needed, and argumentCount
isn’t needed... and I like that the name implies the signature you’re going to get is in fact not at all the same signature you would get for normal checking purposes.
@typescript-bot perf test this |
Heya @andrewbranch, I've started to run the perf test suite on this PR at 0d266cd. You can monitor the build here. Update: The results are in! |
@andrewbranch Here they are:Comparison Report - main..48410
System
Hosts
Scenarios
Developer Information: |
@weswigham the perf impact seems to be lost in the noise, so I’d rather not do something hacky to avoid the check. |
|
||
verify.completions({ marker: ["ts", "tsx", "js"], exact: ["", "drag", "dragenter"] }); | ||
edit.insert("drag"); | ||
verify.completions({ marker: ["ts", "tsx", "js"], exact: ["", "drag", "dragenter"] }); |
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.
Should we also have a test where drag
is already entered and we check that we don't output ""
? I mean, I guess we should just so we know what our current behavior is, I suppose.
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.
We do output ""
, but editors filter it out. This is intentional, because it means if the user backspaces, the editor doesn’t have to re-request completions given a different prefix.
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.
(The test you’re suggesting is already what’s happening on this line)
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.
Ahh, ok, I misread - thanks!
* upstream/main: (473 commits) Correct node used for isDefinition calculation (microsoft#48499) fix(48405): emit dummy members from a mapped type (microsoft#48481) CFA for dependent parameters typed by generic constraints (microsoft#48411) No contextual typing from return types for boolean literals (microsoft#48380) fix(47733): omit JSDoc comment template suggestion on node with existing JSDoc (microsoft#47748) Ensure that we copy empty NodeArrays during transform (microsoft#48490) Add a new compiler option `moduleSuffixes` to expand the node module resolver's search algorithm (microsoft#48189) feat(27615): Add missing member fix should work for type literals (microsoft#47212) Add label details to completion entry (microsoft#48429) Enable method signature completion for object literals (microsoft#48168) Fix string literal completions when a partially-typed string fixes inference to a type parameter (microsoft#48410) fix(48445): show errors on type-only import/export specifiers in JavaScript files (microsoft#48449) Fix newline inserted in empty block at end of formatting range (microsoft#48463) Prevent looking up symbol for as const from triggering an error (microsoft#48464) Revise accessor resolution logic and error reporting (microsoft#48459) fix(48166): skip checking module.exports in a truthiness call expression (microsoft#48337) LEGO: Merge pull request 48450 LEGO: Merge pull request 48436 fix(48031): show circularity error for self referential get accessor annotations (microsoft#48050) Revert "Fix contextual discrimination for omitted members (microsoft#43937)" (microsoft#48426) ...
Fixes #43646
Very involved with #36556