Optional chaining

[ShortDevelopment]

This PR aims to add support for the stage-4 proposal optional-chaining.
It's inspired by the work of @rhuanjl but uses a more hacky approach to parsing.

Goals

• Minimize amount of changes
• (Hopefully) Performance

ToDo

• Add tests
  • Unused expression result
  • Return expression
  • Root optional-call (e.g. eval?.('a'))
  • Scoped method load (e.g. inside eval)
• the delete operator is currently not working
  We might want to split this into a separat PR
• optional call invoked on eval function should be indirect eval
• Simple function calls
• Preserve this
(a.b)().c should be equivalent to a.b().c
• What about the apply call optimization?
• short circuit embedded expressions like a?.[++x]
++x should not be evaluated if a is null or undefined
• Don't tokenize a?.3:0 (ternary) as tkOptChain (?.)
• Tagged templates are not allowed in optional chain

Out of scope

• Remove EmitInvoke (seems unused)
• Better error message for a call on a non-function

---

Spec: Optional Chains
Fixes #6349

[rhuanjl]

I think this looks neater than my version was, though struggling to follow how it handles every branch. What happens if you do a function call in an optional chain does it crash?

(Not going to merge until we have some significant test coverage)

[ShortDevelopment]

I'm currently working on function calls.
At the moment it would just be handled as if no optional-chaining would be used
⇾ crash in js land

Edit: Function call should work now but this is not always propagated correctly (yet)
Edit: Function calls should now work

[ShortDevelopment]

How can I run async tests?

[rhuanjl]

There are a few ways, have a look at test/es7/async-generator-functionality.js for one example.

Note, I'd recommend putting the async tests in a separate file - I think it's easier/better to manage them separately.

[ShortDevelopment]

The jitted code currently crashes at this assertion (causing the test-failures)

ChakraCore/lib/Backend/GlobOpt.cpp

Line 11384 in 1f6e17c

Assert(block->globOptData.IsTypeSpecialized(varSym));

[rhuanjl]

Minor but I think you can delete this line.

resultSlot should be guaranteed to be the same register as innerNode->location here; I tested offline removing this line and your tests all still pass - except for the existing fail for the Simple method Calls test in the Jit.

May be good to replace with a comment noting this or an: Assert (innerNode->location == resultSlot);

[ShortDevelopment]

There's an edge case:
ByteCodeGenerator::ReturnRegister == resultSlot => innerNode->location != resultSlot

[rhuanjl]

Ugh didn't think of that. I suppose for optimal could do: if (innerNode->location != resultSlot) and add the LdA or even a specific check for resultSlot == ByteCodeGenerator::ReturnRegister

Should include a test of this.

[ShortDevelopment]

I'd just go back to sth like

innerNode->location = funcInfo->AcquireLoc(pnodeOptChain);

[ShortDevelopment]

Question: Is it bad to reserve a register BEFORE emitting the whole chain-expression?

[rhuanjl]

I thought in the version before you did funcInfo->ReleaseLoc(innerNode) but didn't actually do AcquireLoc for it as it happened inside the emitting of the chain.

I'm sorry this has brought in quite a bit of confusion, and I'm sorry I didn't think of the case where the resultSlot was already allocated (e.g. as the return reg) - I think it may be cleanest to return to the logic before but simply include an if(resultSlot != innerNode->location) around the LdA.

[ShortDevelopment]

Ooh, no need to say sorry; I just did not understand 😅
After some testing I actually think we need

innerNode->location = funcInfo->AcquireLoc(pnodeOptChain);

though, because the stack of temp-registers will be asymmetrical released for optional-chains as call-target (another edge-case)

(a?.b)().c

[ShortDevelopment]

During testing, test262 sometimes failed while my own tests passed -> I need more tests!
(See ToDo in PR description)

[rhuanjl]

Thinking of extra tests, do you have a cases for an optional call where the property exists but is not callable?

[ShortDevelopment]

Now I have 😉