Fix issue 19829 - __traits(isSame) returns true for some non-local de…

[Biotronic]

…legate lambdas even when they are different

As shown in the issue,

pragma(msg, __traits(isSame, i => x[i], a => a));

prints true. This is because the presence of x is not indicated in any way in the lambda serialization. This PR fixes that by making lambdas incomparable when they use identifiers that are not in the parameter list.

[dlang-bot]

Thanks for your pull request and interest in making D better, @Biotronic! We are looking forward to reviewing it, and you should be hearing from a maintainer soon.
Please verify that your PR follows this checklist:

• My PR is fully covered with tests (you can see the annotated coverage diff directly on GitHub with CodeCov's browser extension
• My PR is as minimal as possible (smaller, focused PRs are easier to review than big ones)
• I have provided a detailed rationale explaining my changes
• New or modified functions have Ddoc comments (with Params: and Returns:)

Please see CONTRIBUTING.md for more information.

---

If you have addressed all reviews or aren't sure how to proceed, don't hesitate to ping us with a simple comment.

Bugzilla references

Auto-close	Bugzilla	Severity	Description
✓	19829	major	__traits(isSame) returns true for some non-local delegate lambdas even when they are different

Testing this PR locally

If you don't have a local development environment setup, you can use Digger to test this PR:

dub fetch digger
dub run digger -- build "master + dmd#9710"

[aG0aep6G]

This is because the presence of x is not indicated in any way in the lambda serialization.

So i => x[i] was being treated like i => [i], right? But that means i => [i] is considered the "same" as a => a. Indeed:

pragma(msg, __traits(isSame, i => [i], a => a)); // Prints "true".

That's another bug, isn't it?

[Biotronic]

You're right. In fact, there's a whole bunch of different things that could happen inside the lambda that should mark it as incomparable. Working on that now.

[RazvanN7]

Maybe a better fix would be to use the mangled name of the variable so that the comparison is rightly done.

[Biotronic]

That would mean i => i wouldn't compare equal to a => a, if I'm understanding you correctly.

[RazvanN7]

That would mean i => i wouldn't compare equal to a => a, if I'm understanding you correctly.

No. I forgot that lambda comparison cannot be done for lambdas that contain non-parameter variables and I was suggesting that the functionality is extended so that lambda comparison works properly with local/global variables by using in the serialization the mangled name of the variable. I just figured that that would need a bit more work because it would require additional semantic passes to be performed.

For now, the fix should be to add a serialization mechanism for arrays.

[thewilsonator]

Please retarget this to stable. Should this be merged soon and then the other issues fixed in another PR?

[RazvanN7]

Yes.

[thewilsonator]

Alright this is good to go as soon as it targets stable.

[Biotronic]

Why stable? If I'm reading CONTRIBUTING.md correctly, that's for regressions only.

Also, the version you approved does not, as aG0aep6G pointed out, correctly handle __traits(isSame, i => [i], a => a) and various other possibilities.

Supporting global variables and the like would be an enhancement, and so should be handled elsewhere.

[Biotronic]

Bah. Or maybe not, since the tests in testlambdacomp assume that many things work that aren't implemented.

[thewilsonator]

Why stable? If I'm reading CONTRIBUTING.md correctly, that's for regressions only.

Its also for serious bug fixes and for bug fixes that will trivially result in no regressions.

[thewilsonator]

compilable/testlambdacomp.d(108): Error: static assert:  `__traits(isSame, (a) => a + r1.a, (b) => b + r2.a)` is false

[RazvanN7]

Wouldn't simply overwriting the Expression node account for all those cases?

override void visit(Expression) { buf.reset(); }

[Biotronic]

That seems reasonable, and was the first thing I tried, but I can't seem to get it to work.

[Biotronic]

It's actually worse than I thought. Try this:

enum i = 2;
static assert(!__traits(isSame, a => i, get!())); // Fails.

template get() {
    enum i = "foo";
    alias get = b => i;
}

The identifier lookup in lambdacomp is very simplistic, and will match identifiers in the local scope when it shouldn't.

[RazvanN7]

The lambda comparison works only for things that can be known right after parsing and with minimal semantic information. The examples that you have shown all require that semantic3 is ran before fetching that kind of information.

[thewilsonator]

No point waiting any longer for this.