aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2013-04-12Support accessing value struct and tuple fields and elements via base class ↵Chad Austin
pointers.
2013-04-12Generalize support for struct fieldsChad Austin
2013-04-12Further generalize support for tuple elements, and in the meantime, fix a ↵Chad Austin
bug that made it into one variant of the function but not the other...
2013-04-12Restructure and generalize tuple accessors.Chad Austin
2013-04-12Move context generation/copying out of JavaScript and into C++, where we'll ↵Chad Austin
be able to do some cool things in the future.
2013-04-12Migrate embind tests from IMVU repository to emscripten repository.Jukka Jylanki
2013-04-12Fix equality testing in previous commitJukka Jylanki
2013-04-12Annotate a few places where -O3 level unsafe optimizations could be ↵Jukka Jylanki
performed by registering a faster form of toWireType function that omits type checks. toWireType can potentially be a 'hot' function in C++<->JS interop.
2013-04-12Fix passing of UINT_MAX and ULONG_MAX from C++ to JS function. LLVM doesn't ↵Jukka Jylanki
have u32 type, so UINT_MAX literal comes out as 'i32 -1' literal to JS, which whould treat it as a negative -1.0 double.
2013-04-12Raise a TypeError if user tries to push an integer from JS side to C/C++ ↵Jukka Jylanki
function exported with embind and the integer (char/short/int/long) is out of bounds of the data type expected by the C++ function.
2013-04-12add a helper method for Northstar unit testing (when emval leaks, we want to ↵jinsuck
see what is left there)
2013-04-12Typed array objects require use of operator new. You can't just call them ↵Chad Austin
directly.
2013-04-12can look up module properties like HEAP8 from within C++ :)Chad Austin
2013-04-12Allow multiple arguments to emscripten::val::new_()Chad Austin
2013-04-12Improve the names of functions registered by embind so they show up in the ↵Chad Austin
chrome profiler.
2013-04-12Fix a bug in external class function definitions and allow specifying ↵Chad Austin
functions that take smart pointers for 'this'
2013-04-12Unify 'this' handling in class functions. 'this' is just the first argument ↵Chad Austin
to the method. This removes several special cases and will remove even more soon.
2013-04-12Some tweaks to facilitate making value_struct and value_tuple simpler.Chad Austin
2013-04-12Add sensible error messages when working with classes that depend on unbound ↵Chad Austin
types.
2013-04-12If calling function that uses unbound types, give a sensible error message.Chad Austin
2013-04-12jshintChad Austin
2013-04-12Simplify getTypeNameChad Austin
2013-04-12Fix a bug in embind when using const return valuesChad Austin
2013-04-12yay, out-of-order registration is testedChad Austin
2013-04-12Change EMSCRIPTEN_BINDINGS syntax and allow for out-of-order registration of ↵Chad Austin
value_tuple elements.
2013-04-12baseClasses is not necessaryChad Austin
2013-04-12RegisteredPointer and RegisteredClass are pretty simple now.Chad Austin
2013-04-12Code keeps disappearing :)Chad Austin
2013-04-12Simplify things even further, remove the Handle functions and replace with ↵Chad Austin
an instance prototype and a constructor function.
2013-04-12Remove some dead codeChad Austin
2013-04-12Simplify upcasting/downcasting, allow multiple smart pointer types per ↵Chad Austin
class, and use the same ClassHandle prototype so class handles and smart pointer handles share implementations for .clone() and .delete()
2013-04-12Kill some embind duplicationChad Austin
2013-04-12Simplify RegisteredPointer#fromWireType someChad Austin
2013-04-12Break some cyclic object dependencies.Chad Austin
2013-04-12Kill fromWireTypeAutoDowncastChad Austin
2013-04-12Give each class a typeid functionChad Austin
2013-04-12make a note that it's sometimes okay to pass raw pointers to smart pointersChad Austin
2013-04-12Rename registeredPointer to ptrTypeChad Austin
2013-04-12Kill some dead todosChad Austin
2013-04-12Use typeid name for interface wrappers too. (Also introduce more uses of ↵Chad Austin
makeLegalFunctionName)
2013-04-12Add a static assertion that a classes's smart_ptr is compatibleChad Austin
2013-04-12Implement a sane error message when passing an incompatible smart pointer typeChad Austin
2013-04-12Support the smart_ptr policy BY_EMVAL and pass the original smart pointer if ↵Chad Austin
the parameter has an identical type.
2013-04-12Allow smart pointers to have their actual type names.Chad Austin
2013-04-12:O it works. Use a sharing policy approach to autocasting pointers.Chad Austin
2013-04-12make smart_ptr a method of class_ instead of a globalChad Austin
2013-04-12enable_if is scary and coolChad Austin
2013-04-12Just toodling around with smart pointer binding syntaxChad Austin
2013-04-12tiny refactoringChad Austin
2013-04-12Work towards killing some pointeeType references.Chad Austin