fixes#27625, Object constraints were not tracked in inference
fixes#27933, pin return type from downwards inference
We now prefer to pick the bound (lower or upper) that had some information on, and it also improves inference error messages somewhat (still a ways to go).
The way this works is we now have a type representing an unknown type: ?. We use ? when performing downward inference steps, instead of `dynamic`. This allows more accurate tracking of type constraints.
For example: given:
var x = await Future.wait([a, b]);
Future.wait<T>'s argument type is Iterable<Future<T>>. Since we didn't know T, we previously pushed down Iterable<Future<dynamic>>. The dynamic caused loss of information. Now we push down Iterable<Future<?>>, allowing us to infer the right type there.
R=leafp@google.com, vsm@google.com
Review-Url: https://codereview.chromium.org/2456803004 .
Our internal tooling is complaining about this when building / checking the ddc sdk. It appears to use the analyzer's summary builder tool which seems less forgiving about dangling references / missing files.
R=srawlins@google.com
Review-Url: https://codereview.chromium.org/2740573003 .
There's some changes in strong mode errors we may want to follow up on.
Unforks the following: async, collection, convert, core, developer, internal, io, isolate, math, mirrors.
Does not unfork: html, indexed_db, js, js_util, svg, typed_data, web_audio, web_gl, web_sql.
R=vsm@google.com
Review-Url: https://codereview.chromium.org/2698353003 .
This implements FutureOr<T> in strong mode, otherwise it's ignored (treated as `dynamic`.
Also fixes strong mode's inference subtype function incorrectly treating `void` as a malformed type. This had the consequence of allowing `void` to be inferred as a type argument.
R=leafp@google.com, paulberry@google.com
Review-Url: https://codereview.chromium.org/2647833002 .
EventStreamSubscription.listen does a cast which is only working in
DDC because we're ignoring the failure. This wraps the listener to
check the argument rather than casting the function. Also picks up a
few misc changes to the main SDK that hadn't been merged into DDC.
BUG=
R=jacobr@google.com
Review-Url: https://codereview.chromium.org/2609633002 .
When I'm looking at the dynamic call profiles of an internal app test,
I see a bunch coming from the SDK itself. In fact, the top five most
common dynamic call sites are in the SDK.
This fixes four of them. (The fifth is in runUnary() in Zone, and I
don't think can be made typed yet.)
R=vsm@google.com
Review URL: https://codereview.chromium.org/2542723003 .
Really, this works around them, but the workaround is all round better
code.
Remove the Dart implementations of min() and max() and just forward to
the JS ones. In DDC, all numbers are double, so the type checks to
handle them specifically aren't meaningful. Also, we don't need the
other special case checks in there to help dart2js optimize them.
R=vsm@google.com
Review URL: https://codereview.chromium.org/2386493003 .
If a comparer isn't provided, it was silently assuming the element type
implements Comparable, even though that isn't required. Fixed to add a
manual cast at each comparison.
A stricter faster approach might be to check that the type argument
itself is a type that implements Comparable, but I don't know if that's
too pessimistic, or even possible to do.
R=vsm@google.com
Review URL: https://codereview.chromium.org/2311513002 .
Jennifer Messerly