[JSC] ModuleNamespaceObject's property caching has different mechanism than IC
Created attachment 408116 [details] Patch
<rdar://problem/68415681>
Comment on attachment 408116 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=408116&action=review > JSTests/stress/not-cache-over-uncacheable-dictionary.js:14 > +function shouldBe(actual, expected) { > + if (actual !== expected) > + throw new Error('bad value: ' + actual); > +} > + > +var object = {}; > +$vm.toUncacheableDictionary(object); > +shouldBe(String(object), `[object Object]`); > +shouldBe(String(object), `[object Object]`); > +shouldBe(String(object), `[object Object]`); > +object[Symbol.toStringTag] = "OK"; > +shouldBe(String(object), `[object OK]`); > +shouldBe(String(object), `[object OK]`); > +shouldBe(String(object), `[object OK]`); This test does not fail without this patch while we were caching. This is because we are relying on watchpoints for properties instead of structure ID. But in this patch, we changed the code not to cache properties in uncacheable-dictionary. So testing it here is nice.
Created attachment 408117 [details] Patch
Created attachment 408118 [details] Patch
*** Bug 216262 has been marked as a duplicate of this bug. ***
Comment on attachment 408118 [details] Patch r=me
Committed r266715: <https://trac.webkit.org/changeset/266715>