* add bug repro test
* add test and start fix implementation
* adjust for useAlias preference
* fix existing renaming test
* refactor to get rid of options
* fix named bindings & other imports cases
* fix eslint error
* address cr comments
* hopefully actually fix eslint
* clean up stale baseline
* make API change non-breaking
* add/fix comments
* Merge 'decorators' into 'modifiers' on various Nodes
* Drop RESERVED argument in favor of removing parameter
* Ignore grammar error nodes when asserting invariants
* Revert 'illegalX' property renames
* PR Feedback
* Use fixed time for vfs so baselining is consistent
* Baseline buildinfos
* Write new file text in baseline even if the file wasnt read on the shadow
* Remove unnecessary debugger statement
* Make sure that incremental correctness is checked with correct writeFile so we know buildInfo was written
Also baseline these so its easy to verify the changes
* More baselines for the tsbuildinfo
* If we are writing dts file and have used file text as version, we can update the signature when doing actual emit
* Make WriteFileCallback Api ready for future
* Assert that there is only single source file when emitting d.ts file
* Add test
* Renames
* More refactoring
* If we are updating dts of any of the file and it affects global scope, everything needs update in signature and dts emit
Fixes#42769
* Stacktrace optimization for getModified time in anticipation of using it more than fileExists wherever possible
* Baseline getModifiedTime, setModifiedTime, fileExits and directoryExits for experiment
* Remove unnecessary write file finger print code since its not used at all
* Use modified time instead of file existence check
* Remove unnecessary getModifiedTime
* No need to check for file existence before reading the d.ts file
* Do project reference errors before doing input/output file checks
* Dont call getModifiedTimes if dts change
* Passdown modified time if queried
* Use modified time passed through the file watching in tsbuild
* Handle force build as separate upto date status
* uptodate status worker to read buildinfo and use it to determine upto date ness
* No need to update output timestamps if buildinfo will determine uptodateness
* Store change file set instead of hasPendingChange to be able to reuse the information
* Add test that shows input file is not present
* No need to check input time stamp before buildinfo
* Keep buildinfos for lifetime of the solution builder and project
* Store modified time along with text of buildinfo
* Non composite projects dont need to track declaration change time
* Pass through buildInfo so we dont have to parse it back
* Save dts change time in buildinfo itself
* Store dts time for --out in the buildInfo
* Store hash of text in the bundle info so it can be verified before manipulating text for fast updates during prepend
This helps when text changes during incremental build toggling and we determine we can just manipulate text
* Since buildinfo is cached no need to maintain version check state
* Store output time stamps for non incremental builds
* Revert "Baseline getModifiedTime, setModifiedTime, fileExits and directoryExits for experiment"
This reverts commit 7e65cd3315ccf387f6d7e7c40263e85bcc3c961c.
* Change verbose messages for upto date status
* Reconcile reusable builder state and builder state so there are not two different types that are almost similar looking
* Cleanup impliedFormat
* Cleanup
* Cleanup noEmit option
* BuildInfo options emit as a flag
* Factor out types for program written in buildinfo with and without bundle emit
* No need to store output file stamps if not in watch mode
* Cleanup
* Test for single watch per file
* Fix emit and error update baselines that were duplicate
* More refactoring
* Only copy emit state fields when backing up to restore if emit fails
* Instead of maintaining delta of changes, maintain old state for those changes
* Add test to verify build when input file does not change
* If version of the input file does not change, dont mark as out of date
* Disable lint warning as build fails without the assert
* Report aggregate statistics for solution as well as some solution perf numbers
* Options solutionDiagnostics instead so that its not too verbose when printing diagnostics
* When tsc --build --clean, only remove tsbuildinfo if its incremental build
* Revert "Options solutionDiagnostics instead so that its not too verbose when printing diagnostics"
This reverts commit 0cf9e301038267bc1d9439c15d279940809b0c00.
* Revert "Report aggregate statistics for solution as well as some solution perf numbers"
This reverts commit 405d8e91978708a613fe8cb631a83b13c430a808.
* Revert "When tsc --build --clean, only remove tsbuildinfo if its incremental build"
This reverts commit e4e66726098201f38f5e8908a538e7a506f15efd.
* Comments in the code
* Feedback
* Improve reduction of intersection types
* Accept new baselines
* Improve CFA for truthy, equality, and typeof checks
* Accept new baselines
* Remove special case for Function type
* Don't reduce intersections of form {...} & object
* Accept new baselines
* Anything is assignable to unknown-like union
* Accept new baselines
* Tweak subtype check
* Recombine unknown type from unknown-like union in more cases
* Display union origin only if it is shorter than union itself
* Accept new baselines
* Add tests
* Only attach origin type when it is shorter than union itself
* Specially preserve string & {}, number & {}, bigint & {}
* Accept new baselines
* Add additional tests
* Fix getNormalizedType and getNarrowableTypeForReference for intersections
* Switch NonNullable<T> to use T & {}
* Accept new baselines
* Use NonNullable<T> in place of anonymous T & {}
* Accept new baselines
* Add fourslash test
* More fourslash tests
* Fix getFalsyFlags handling of intersections
* Accept new baselines
* Add constraint to compareProperties type parameter
* Unconstrained type parameter not assignable to {} with strictNullChecks
* Accept new baselines
* More specific inference for constrained 'infer' types in template literal types
* PR feedback
* Add inference priority for template type placeholders
* Infer to a preferred constraint instead of a union
* Add reduceType
* Switch tests to use infer..extends
* Add missing primitive constraint cases
* Update .types tests
* Remove TemplateTypePlaceholderPriority
* Remove reduceType
* fix(47562): Add option to suppress type hint if variable name matches type
* Remove the unnecessary debug code
* Re-run gulp runtests
* Use equateStringsCaseInsensitive to compare strings
* Allow export map entries to remap back to input files for a program
* Fix file casing issues on windows
* Implement abiguity error, doesnt quite work
* Refine selection logic in error case to use getCommonSourceDirectory, add more tests
* Remove Node12, add Node16.
* Accepted baselines.
* Refactor checking for top-level await, give a better error message in CJS files.
* Accepted baselines.
* Stop erroring on JSON module imports in node ESM since they're no longer experimental.
* Accepted baselines.
* More refactoring, do the same checks for for-await loops.
* Accepted baselines.
* Adjust phrasing to permit for-await on CJS error.
* Accepted baselines.
* Accepted baselines.
* Fix lints.
* Add failing test
* Block auto-import module specifiers including node_modules path
* Eagerly resolve module specifiers in completions in nodenext so failures can be filtered
* Add completion info flags for telemetry
* Update API baseline
* Update completions baselines
* Fix missed boolean flip
* Fix remaining tests
* If we are writing dts file and have used file text as version, we can update the signature when doing actual emit
* Make WriteFileCallback Api ready for future
* Assert that there is only single source file when emitting d.ts file
* Prototype resolving to JS when go-to-def aliases all resolve to ambient declarations
* Add test infrastructure
* Start fleshing out test coverage
* Fix some go-to-def stuff
* Finish lodash test case
* Make go-to-implementation never return ambient results
* Build new functionality into go-to-implementation
* Update baselines
* Two more test cases
* Refine definition searches for unresolved imports
* Revert "Build new functionality into go-to-implementation"
This reverts commit 381799d0f104000c676334c8e08fcdff9ccac896.
* Fix tests
* Revert go-to-implementation changes
* Wow a bunch of code was unnecessary
* Update baselines and go-to-def test
* Fix navigation on symbols that are not aliases but resolve through aliases in chain
* Temporarily replace go-to-def with new command implementation
* Revert "Temporarily replace go-to-def with new command implementation"
This reverts commit 34c6cfdebbb9e2506ea5e6d528d2cb57b2cdeb2e.
* Revert "Wow a bunch of code was unnecessary"
This reverts commit 1cb2ba646c9e80d1ca08b6fc3f6a2219aac44f67.
* Bring back some deleted code needed for a new test case
* Clean up a little
* Rename more stuff
* Update test
* Update API baseline
* Temporarily replace go-to-def with new command implementation
* PR review fixes
* Fix getTopMostDeclarationNamesInFile
* Rename local
* Use hash set
* Remove option from commandLineParser
* Keep noDtsResolution project around
* Handle AuxiliaryProject kind in ScriptInfo getDefaultProject etc.
* Do not run updateGraph in the background for AuxiliaryProject
* Don’t create auxiliary project outside of semantic mode
* No-op on scheduled invalidation
* Add comments to unit test
* Sync compiler options to auxiliary project
* Fix case sensitivity
* Update extensionIsOk with new file extensions
* PR feedback
* Update API baseline
* Mark scheduleInvalidateResolutionsOfFailedLookupLocations internal
* Use same heuristics on property accesses of loosely-resolvable aliases as unresolvable named imports
* Rename command, and no need to return the bound span
* Update API baseline
* Reuse type nodes from optional parameters and properties even when not written as a union with `undefined`
* Remove newly unneeded NodeBuilderFlag
* Update public API
* Update baselines from main
* Add 'extends' clause to 'infer' type
* Revise parse for infer..extends, improve parenthesizer
* More aggressive parens to match existing DT tests
* tests 'infer' constraint using outer type parameter
* Adds a test showing 'infer' cannot reference other 'infer' in same 'extends'
* Emit extends clause for synthetic infer typesduring declaration emit
* Add moduleSuffixes compiler option and related tests. Update baselines for compiler options tests.
* Add a flag to the command-line parser which allows "list" params to preserve "falsy" values such as empty strings. Falsy values are normally stripped out.
* Add tests. Rework resolver logic to only run module-suffix code when needed.
* PR feedback
* Add test
* Remove unnecessary conditional.
* add label details to completion entry
* Use label details for obj literal method completions
* add label details support flag
* add label details support to fourslash
* support both label details and non-label details in object literal method snippets
* CR fixes
* fixes after rebasing
* fix tsserver tests
* skeleton of new feature
* working prototype
* refactor print and format code into its own function
* minor changes; don't support overloads
* have two completion entries
* get rid of accessor support
* add snippet support
* add formatting
* add trailing comma
* add sourcedisplay
* support auto-imports via completion details
* add user preference option and fix ordering of entries
* cleanup
* don't return code actions for no import fixes
* make sortText lower priority for snippets
* get rid of flag
* use optional member sort text
* update baselines
* don't collect method symbols if insert text is not supported
* remove comment
* return undefined if type is not function type
* only slice if needed
* use union reduction; more test cases
* WIP: modify sort text system
* Improve new sort text system
* add signature and union type check
* re-add flag
* fix tests
* rename sort text helper
* fix test and code for union case
* add new flag to protocol type
* fix spaces
* CR: minor fixes
* CR: more fixes
* CR: restructure main flow
* minor fix
* Simplify getVariancesWorker and associated logic
* Accept new API baselines
* Add 'in' and 'out' modififers / add modifiers to type parameters
* Check variance annotations
* Update test runner
* Accept new API baselines
* Allow variance annotations only on certain type parameters
* Add deprecated implementation of createTypeParameterDeclaration
* Accept new API baselines
* Report variance markers as 'sub-XXX' and 'super-XXX'
* Add tests
* Accept new baselines
* Add moduleDetection compiler flag to allow for changing how modules are parsed
The default setting is 'auto', where JSX containing files under react-jsx and react-jsxdev are
always parsed as modules, and esm-format files under module: node12+ are always parsed as modules,
in addition to the 'legacy' detection mode's conditions for other files. (Declaration files are exempt from
these new conditions)
The 'legacy' mode preserves TS's behavior prior to the introduction of this flag - a file is
parsed as a module if it contains an import, export, or import.meta expression.
In addition, there is a 'force' mode that forces all non-declaration files to be parsed as modules.
(Declaration files are still only modules if they contain a top-level import or export.)
This technically breaks the parser API, but it's kinda-sorta backwards compatible so long
as you don't need the functionality associated with more recent compiler flags.
* Fix post-merge lint
* Rename function
* Update default value documentation
* PR feedback
* Fix lint and typo
* Rename SyntaxKind.JSDocComment -> JSDoc
Early on, I made the mistake of using the syntax kind JSDocComment for
the type JSDoc. This arose because
I hadn't decided on the terminology of "jsdoc" for the C-style comment
that is attached to a declaration. "comment"
for the text that follows a tag. By the time I settled on those terms
consistently, a version had already shipped with `interface JSDoc`
having `SyntaxKind.JSDocComment`.
However, there is now a `type JSDocComment` that represents the possible
contents of the comment text (strings and various kinds of `@link`), so
it's doubly confusing that this type has a union of four kinds instead
of SyntaxKind.JSDocComment.
Although this is a public API break, I don't think it's hard to recover
from, and the JSDoc API has accessors for individual tag types, so I
suspect few people refer directly to SyntaxKind.JSDocComment.
* Add deprecate JSDocComment alias to JSDoc
For backward compatibility
* Make JSDoc an alias of JSDocComment
Improves output of JSDocParsing tests without having to make its code
smarter.
* update API baselines
* Add import assertions for type-only imports and import types to change resolver modes
* By popular request, only allow mode assertions on top-level type only imports
* Add specifier options parameter to specifier generation
* Permit type arguments in references to generic functions
* Accept new baselines
* Delete pointless fourslash test
* Fix lint issue
* Finalize implementation
* Add tests
* Accept new baselines
* Properly handle instantiation of instantiation expression types
* Accept new API baselines
* Fix lint error
* Add more tests
* Properly handle unions/intersections of generic types
* Add more tests
* More permissive parsing of type arguments in member expressions
* Update tests
* Accept new baselines
* Triple-slash reference type directives can override the import mode used for their resolution
They now use the file's default mode by default, rather than always using commonjs. The new arguments to the
reference directive look like:
```ts
///<reference types="pkg" resolution-mode="require" />
```
or
```ts
///<reference types="pkg" resolution-mode="import" />
```
* Omit redundant import modes in emitter
* Add test for #47806
* Add server test for triple-slash reference mode overrides
* Move FileReference mode into helper
* Update tests/cases/conformance/node/nodeModulesTripleSlashReferenceModeOverride3.ts
Co-authored-by: Nathan Shively-Sanders <293473+sandersn@users.noreply.github.com>
Co-authored-by: Nathan Shively-Sanders <293473+sandersn@users.noreply.github.com>
* Eliminate redundant or meaningless elaborations in type relations
* Accept new baselines
* Add resetErrorInfo (though, oddly, shouldn't be necessary)
* Less aggressive reduction in second pass union/intersection checks
* Accept new baselines
* Restructure and back off a little bit more
* Only cache union/intersection relations once
* Accept new baselines
* Properly cache identity relations, clean up error reporting
* Move more logic to cached side of relation checks
* Optimize and remove more redundant elaborations
* Accept new baselines
* Remove unnecessary error state capture
* More optimizing
* Cache isWeakType computation
* Revert "Cache isWeakType computation"
This reverts commit 25a71c4de61f6366ffac080d19685dcb200f42b9.
* Address CR feedback
* Accept new baselines
* Import fix
* Wire up completions, add sorting to fix
* Fix overlapping changes when there’s only one import specifier
* Update API baseline
* Add sorting and filtering back to UMD fix
* Use package.json cache in module specifier generation
* Let AutoImportProviderProject reuse module resolution cache of host project
* Add missing module resolution cache access, add logging to getRootFileNames
* Reuse packageJsonImportFilter
* Only log when the project will be created, update API baseline
* Remove override that could mess up watches
* Allowed non-this, non-super code before super call in derived classes
Fixes#8277.
It feels wrong to put a new `forEachChild` loop in the checker, though in the vast majority of user files this will be a very quick one. Is there a better way to check for a reference to `super` or `this`?
* Used new isNewThisScope utility and dummy prop in baselines
* Accounted for parameters and get/set accessor bodies
* Accounted for class extensions
* (node|statement)RefersToSuperOrThis wasn't checking root level scope boundaries
```ts
function () {
return this;
}
```
It was immediately going to `ts.forEachChild` so the statement itself wasn't being counted as a new `this` scope.
* Better 'references' name and comments; accounted for more method edge case
* Limited super calls to root-level statements in constructor bodies
As per discussion in the issue, it would be ideal to consider any block that always ends up calling to super() the equivalent of a root-level super() statement. This would be valid:
```ts
foo = 1;
constructor() {
condition() ? super(1) : super(0);
this.foo;
}
```
...as it would compile to the equivalent of:
```ts
function () {
condition() ? super(1) : super(0);
this.foo = 1;
this.foo;
}
That change would a bit more intense and I'm very timid, so leaving it out of this PR. In the meantime the requirement is that the super() statement must itself be root-level.
* Fixed error number following 'master' merge
* Added decorator test cases
* Again allowed arrow functions
* Accepted new baselines
* Added allowance for (super())
* Reworked emitter transforms for ES this binding semantics
In trying to adjust to rbuckton's PR feedback, this orders derived class constructor bodies into five sections:
1. Pre-`super()` code
2. The `super()` call itself
3. Class properties with initializers
4. Parameter properties
5. The rest of the constructor
I've looked through the updated baselines and it looks like they're generally better than before. Within the existing test cases that result in semantic errors for `this` access before `super`, several previously resulted in a global `_this` being created; now, they correctly defer referring to `_this` until it's assigned to `_super.call(this) || this`.
* Used skipOuterExpressions when diving for super()s; fix prop ordering
* Allow direct var _this = super() when no pre-super() statements; lint fixes
* Always with the TSLint
* One last touchup: skipOuterExpressions in es2015 transformer
* Fixed new lint complaint in utilities.ts
* Again added a falls-through; it'd be swell if I could run linting locally
* This time I think I got it
* Well at least the error is a different one
* Undid irrelevant whitespace changes
* Mostly addressed private/field issues
* Accepted derivedClassSuperProperties baseline
* Lint fix, lovely
* Remove now-unnecesary comment
* First round of feedback
* Moved prologue statements to start of statements
* Added consideration for super statements in loops and the like
* Ordering and a _this_1 test
* Missed the one change I needed...
* First round of feedback corrections
* Feedback round two: statements
* Feedback: used more direct statements
* Fixed classFields emit to not duplicate temp variables
* Refactored es2015 helper to be less overloaded
* Accounted for parentheses
* Simpler feedback: -1, and emptyArray
* Next feedback: superStatementIndex
* Feedback: simplified to no longer create slice arrays
* Adjusted for default and rest parameters
* Added test case for commas
* Corrected comment ranges
* Handled comments after super, with tests
* Fixed Bad/Late super baselines
* Remove unused param and unnecessary baseline comments
Co-authored-by: Orta Therox <orta.therox@gmail.com>
Co-authored-by: Ron Buckton <ron.buckton@microsoft.com>
