Comparing changes
Open a pull request
- 19 commits
- 22 files changed
- 2 contributors
Commits on Jun 6, 2025
-
Fix protocol fn DCE issue (#252)
* fixes invoke to cljs.core/str in protocol fns * add another trivial dce test w/ a protocol fn
-
Tiny DCE improvements - just use empty list in IndexedSeq - just invoke toString on StringBuffer - inline toString impl for EmptyList - reify should not emit basis static method - reify should set meta to nil if no actual meta, not empty map
Commits on Jun 9, 2025
-
pr-writer-impl lower level impl for js object printing
* remove pr-writer-impl dependence on lazy seq, MapEntry - use Array.map instead of map - reify IMapEntry instead of concrete MapEntry * use primitive regex method
-
Remove pr-opts calls, backwards compatibility tweaks (#257)
* remove calls to pr-opts - just use dynamic binding - keep it backwards compatible * add long missing infer-tag case for :try
-
Commits on Jun 10, 2025
-
Use primitives in print-map (#258)
* lower print-map - lift pr-map-entry-helper, implement ISeqable - lift-ns uses array of MapEntry instead of actual map
-
Remove the circularity that str has with IndexedSeq (#255)
* remove the circularity that str has with IndexedSeq - add private str_ for cljs.core usage - keep str for now to avoid any potential breakage - add some simple tests for apply + str_ - the other cases already well covered by existing tests around printing
-
More trivial source build tests verifying DCE doesn't regress (#259)
* add two more trivial output tests - keyword should be very small - vector by itself should be reasonable - bump windows version
Commits on Jun 13, 2025
Commits on Jul 6, 2025
-
Refactor has-extern? / js-tag (#246)
* fix up tests so they don't throw if no warnings * resolve-extern, which can be used by both has-extern? and js-tag, returns both resolved prefix and var info * remove various hacks around extern resolution (Number, Window, prototype etc.), resolve-extern handles everything * undefined is a ref cycle, special case * change normalize-js-tag so it marks the ctor prop * add normalize-unresolved-prefix to fix up the cases we can't find * add impl unit tests * more unit test - can finally resolve crypto.subtle, verify type inference as well * add lift-tag-to-js helper * in resolve-var add the ctor to the tag to track later, this also lets the extra information flow * in analyze-dot check to see if the target is a constructor - if it is use that as tag instead, Function is not useful * test assertion that we can figure out the return even if a extern js ctor is bound to a local * add compiler test case for inferring return of Number.isNaN * add non-ctor inference for array, string, boolean and number, will be useful later * add js/isNaN test * add isArray extern test * don't return raised js/Foo types for boolean, number, string * remove ^boolean hint from array? ass test * remove hint for make-array, add test * remove hints for isFinite and isSafeInteger, tests * can infer distinct?, add test * remove various ^boolean cases no longer needed * FIXME comments about dubious ^boolean cases * move ^boolean hint from special-symbol? to contains? where it belongs, test case * goog.object/containsKey type inference doesn't work for reason, leave a trail for later * goog.string/contains does work, add test * remove hint from NaN? * FIXME note about re-matches, another dubious case of ^boolean hints
-
CLJS-3439: REPL doc support for externs (#261)
* break out ->pre * add cljs.analyzer.api/resolve-extern * new doc lookup path in repl for js/foo symbols
Commits on Jul 7, 2025
-
CLJS-3438: Inference for
goog.object/containsKeyreturns any, not b……oolean (#262) - fix cljs.analyzer/desugar-dotted-expr, generated malformed AST in the case of goog.module var - compiler test for goog.object/containsKey - fix parameter parsing in cljs.externs to properly handle var args and optional arguments - fix fn-arity warning so that we use unaliased names if available (goog.module names are aliases) - cljs.core: goog.object/containsKey hint no longer needed
-
This comparison is taking too long to generate.
Unfortunately it looks like we can’t render this comparison for you right now. It might be too big, or there might be something weird with your repository.
You can try running this command locally to see the comparison on your machine:
git diff r1.12.42...master