This seems to be a regression in between ESR52 and ESR60. The attached example program compiles and runs with ESR52 (barring the JSClassOps layout change) but does not compile with ESR60: In file included from bug.cpp:5: In file included from .../include/mozjs-60/jsapi.h:29: In file included from .../include/mozjs-60/js/CallArgs.h:73: .../include/mozjs-60/js/RootingAPI.h:274:32: error: no member named 'exposeToJS' in 'js::BarrierMethods<JSString *>' js::BarrierMethods<T>::exposeToJS(ptr); ^ .../include/mozjs-60/js/RootingAPI.h:277:9: note: in instantiation of member function 'JS::Heap<JSString *>::exposeToActiveJS' requested here exposeToActiveJS(); ^ .../include/mozjs-60/js/RootingAPI.h:268:5: note: in instantiation of member function 'JS::Heap<JSString *>::get' requested here DECLARE_POINTER_CONSTREF_OPS(T); ^ .../include/mozjs-60/js/RootingAPI.h:160:40: note: expanded from macro 'DECLARE_POINTER_CONSTREF_OPS' operator const T&() const { return get(); ... ^ ...l/include/mozjs-60/js/HashTable.h:1378:34: note: in instantiation of member function 'JS::Heap<JSString *>::operator JSString *const &' requested here return HashPolicy::match(HashPolicy::getKey(e.get()), l); ^ .../include/mozjs-60/js/HashTable.h:1406:42: note: in instantiation of member function 'js::detail::HashTable<js::HashMapEntry<JS::Heap<JSString *>, void *>, js::HashMap<JS::Heap<JSString *>, void *, js::DefaultHasher<JSString *>, js::SystemAllocPolicy>::MapHashPolicy, js::SystemAllocPolicy>::match' requested here if (entry->matchHash(keyHash) && match(*entry, l)) { ^ .../include/mozjs-60/js/HashTable.h:1782:23: note: in instantiation of member function 'js::detail::HashTable<js::HashMapEntry<JS::Heap<JSString *>, void *>, js::HashMap<JS::Heap<JSString *>, void *, js::DefaultHasher<JSString *>, js::SystemAllocPolicy>::MapHashPolicy, js::SystemAllocPolicy>::lookup' requested here return AddPtr(lookup(l, keyHash, sCollisionBit), *this, keyHash); ^ .../include/mozjs-60/js/HashTable.h:152:21: note: in instantiation of member function 'js::detail::HashTable<js::HashMapEntry<JS::Heap<JSString *>, void *>, js::HashMap<JS::Heap<JSString *>, void *, js::DefaultHasher<JSString *>, js::SystemAllocPolicy>::MapHashPolicy, js::SystemAllocPolicy>::lookupForAdd' requested here return impl.lookupForAdd(l); ^ bug.cpp:46:15: note: in instantiation of member function 'js::HashMap<JS::Heap<JSString *>, void *, js::DefaultHasher<JSString *>, js::SystemAllocPolicy>::lookupForAdd' requested here hashMap.lookupForAdd(str); ^ Adding the requested template specialization makes the program compile and run as expected.

Comment on attachment 8981252 [details] [diff] [review] Add exposeToActiveJS specialization for JSString Review of attachment 8981252 [details] [diff] [review]: ----------------------------------------------------------------- Great, thanks for fixing this.

Thanks for the review! I wasn't sure if it was a bug or I was using the API incorrectly.

Pushed by ryanvm@gmail.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/d44967058ffa Add exposeToActiveJS specialization for JSString. r=sfink

Comment on attachment 8981252 [details] [diff] [review] Add exposeToActiveJS specialization for JSString [Approval Request Comment] If this is not a sec:{high,crit} bug, please state case for ESR consideration: Allow a SpiderMonkey API to work as intended User impact if declined: None to Firefox as the API isn't used within the codebase, but it's needed by embedders Fix Landed on Version: 62 Risk to taking this patch (and alternatives if risky): Unlikely to cause any risk to Firefox, as any use of this API would previously not have compiled. String or UUID changes made by this patch: None See https://wiki.mozilla.org/Release_Management/ESR_Landing_Process for more info.

Comment on attachment 8981252 [details] [diff] [review] Add exposeToActiveJS specialization for JSString Spidermonkey API fix for embedders. Approved for ESR 60.1. I'm not going to bother approving this for Beta61 since it seems unlikely to be particularly important there. Feel free to nominate the patch if I'm missing something, though.