Discard statements no longer affect the behavior or uniformity
analysis. Update the resolver, validator, and several tests to reflect
this.
Some E2E tests were removed as they had loops that are now considered
to be infinite.
Use the DemoteToHelper transform to emulate the correct semantics on
platforms where discard is (or may) terminate the invocation in a
manner that would affect derivative operations.
We no longer need the UnwindDiscardFunctions transform for HLSL, which
already implements the correct semantics. However, we still run the
DemoteToHelper transform for the HLSL backend due to issues with FXC's
handling of discard statements (see crbug.com/tint/1118).
Fixed: tint:1723
Change-Id: Ib49ff187919ae81c4af8675e1b66acd57e2ff7d2
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/109003
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: James Price <jrprice@google.com>
The updated WGSL validation rule now requires that the memory view of
the argument matches its root identifier.
This allows for code like this:
let p = &v;
foo(p);
Fixed: tint:1754, tint:1734
Change-Id: I3239ec84e1c06398a6ce5bebb1e0b28986764bc6
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/109221
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: James Price <jrprice@google.com>
This is now a well-defined term in the WGSL spec, so we should use it.
Change-Id: Icc46a77f0a465afbfd39cdaec84e506b143c8c0c
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/109220
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Auto-Submit: James Price <jrprice@google.com>
If count is highest and offset is non-zero, or vice-versa, we'd overflow
the count + offset > bit-width check. This CL fixes this case.
Bug: tint:1581
Bug: chromium:1381810
Change-Id: I6ee60ec1a13230fca6f4bb6407cd33bcc6730eb7
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/109162
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
This Cl changes the default acos test value so it equals `cos(.25)`
instead of `cos(.5)`. The `.5` result was slightly different on
various machines.
Bug: tint:1581
Change-Id: I768141ae87a63b2c35ec4d61eb030edad77e11e4
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/108863
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Also:
Print unrepresentable numbers with higher precision - otherwise values can round, and diagnostics can be very confusing.
Improve diagnostic distinction between `( )` `[ ]` interval ranges.
Change-Id: I9269fbf1738f0bce5f2ddb5a387687543fd5d0bb
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/108700
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
This Cl adds const-eval for the `floor` builtin.
Bug: tint:1581
Change-Id: I992eba3aa6c66707e923907a4bb912c2f6f8d290
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/108343
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
This reverts commit c22b8b9dc8c058fef45c72a3c2cdd70d90fc911d.
Reason for revert: Cost not worth benefit.
Original change's description:
> Add CMake build option to generate intrinsic files.
>
> This Cl adds an option to cmake to generate the intrinsic files
> as part of the buid.
>
> Change-Id: I785fbaa57b6d3b3ecd16e36b2f2baaa5da3cc9d9
> Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/108000
> Kokoro: Kokoro <noreply+kokoro@google.com>
> Reviewed-by: Ben Clayton <bclayton@chromium.org>
> Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
# Not skipping CQ checks because original CL landed > 1 day ago.
Change-Id: Idc57ea21d9d9152703e0a6d99fa518e8da079092
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/108460
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
This CL adds const-eval for the `ceil` builtin
Bug: tint:1581
Change-Id: I32957d17000df1a8a59ce37f49b6177d1087406c
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/108342
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
This CL adds const-eval for the `acos` builtin.
Bug: tint:1581
Change-Id: I01c0f48e73eedf87cf9c912715487f8eea44f64e
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/108341
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
This CL adds const-eval for the `abs` builtin.
Bug: tint:1581
Change-Id: I6ee25c07620990f72a6962441aec62ae7665653e
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/108340
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
This CL adds const-eval for the `all` builtin.
Bug: tint:1581
Change-Id: Ib1d857ba24114c61e1a90b30dd66d063edfa136d
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/107464
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Involves expanding the source range of a variable declaration so we can point at something that can include the 'const'.
Fixed: tint:1740
Change-Id: Ie8f784de34a1792002aaa708c1b77053be54f1b5
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/108120
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Now that expressions can only ever have the `Next` behavior,
short-circuiting operators will always reconverge.
Change-Id: Ib9fe4c774191b8a304fa7f7a6eafdfa6c6e6f18a
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/108202
Kokoro: Kokoro <noreply+kokoro@google.com>
Auto-Submit: James Price <jrprice@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: James Price <jrprice@google.com>
Invalid values to @binding(), @group() and @location() would fail resolving without an error diagnostic. This later triggers and ICE.
Refactor duplicate @location resolving in 4 places to a single method.
Canonicalize the diagnostic messages for attributes.
Remove a bunch of TODOs
Bug: chromium:1380212
Bug: tint:1633
Change-Id: Id2cc6ba4b807f12f350a2a31ef87fa0f185b64c3
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/108144
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
This Cl adds an option to cmake to generate the intrinsic files
as part of the buid.
Change-Id: I785fbaa57b6d3b3ecd16e36b2f2baaa5da3cc9d9
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/108000
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@chromium.org>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
I got the rules around this wrong. This should be allowed, but the array types cannot compare equal if they are unnamed override-expressions.
Fixed tint:1737
Change-Id: I83dc49703eed015e9c183e804474886da5dad7b9
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/107685
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
We now use the MergeReturn transform for SPIR-V, which is the only
backend that requires special handling to meet these function call
reconvergence requirements.
Fixed: tint:1627, tint:1726
Change-Id: I25f848f4b9ff0fd301b8a27a220bb09cdb2867ca
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/107364
Commit-Queue: James Price <jrprice@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
This CL adds const-eval for the `any` builtin.
Bug: tint:1581
Change-Id: I32d8946b3cd5c6d210b75104fa37c4d1ef6a6f84
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/107542
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
For "vec3(vec3<T>) -> vec3<T>", this CL makes T also include abstract
int and float (all scalars).
Bug: tint:1731
Change-Id: I776c7ba6872c3d680982dbec6b0970389e720611
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/107540
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
This CL adds const-eval for the `sign` builtin.
Bug: tint:1581
Change-Id: I5d9bfd3f3f742bcba69fbb0d7f47dc57ce18e134
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/107460
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
The value of 1 for the `asin` is providing a slightly different result on OSX
compared to the Windows and Linux results. This CL updates the test case to use
asin(0.479425538604) = 0.5
Bug: tint:1581
Change-Id: Ibf7a921d1476c0d24da696eb6cc2b491c1ca1516
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/107541
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
The incorrect dispatch call was used for step and saturate. It
included the int and uint aspects when not needed.
Bug: tint:1581
Change-Id: Ibb80a7869eaa674d38bdae69929c9e91011918f8
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/107463
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
This CL adds const-eval for the `saturate` builtin.
Bug: tint:1581
Change-Id: I3729ea5b381b04b73bbe1bc8e03e5ce65c27e082
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/107362
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
This CL adds const-eval for the `step` builtin.
Bug: tint:1581
Change-Id: Idbf773fb88892a8a5e620bcfe8b779dee148f746
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/107281
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
This CL adds const-eval for the `asin` and `asinh` operators.
Bug: tint:1581
Change-Id: I18b2eeb4fb85b8979012b48551eefa773d1b980e
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/106980
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
To match the spec.
Also add a bunch of missing texture test cases to
src/tint/ast/builtin_texture_helper_test.cc. Fix all the tests that were
broken because these were not being exercised.
Fixed: tint:1526
Change-Id: I207b51d307bbdc054b595e0e0e0fd3330607e171
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/106681
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Auto-Submit: Ben Clayton <bclayton@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
This CL removes the parsing of a `-` in front of numerics when lexed.
This will cause the number to become a UnaryOperator negation then the
number instead of a negative number.
Bug: tint:1679, tint:1141, tint:1570
Change-Id: I217c0ffcbe5be934c8d56bd83141b47ade83bc60
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/106463
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
This fixes cases where input args of different type are mixed between
constant and runtime evaluated. We initially did this only for builtins
(https://dawn-review.googlesource.com/c/dawn/+/104545) to fix cases like
`select`, but we also need this for binary operators like shift
left/right that also take args of differing types. For consistency, this
change also includes passing in the earliest evaluation stage when
looking up unary ops and initializer/conversions.
Bug: tint:1713
Change-Id: I196c67acfe767aa01aa45298d75cfc4d345d08ea
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/107140
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
This CL adds const-eval for the `atanh` operator.
Bug: tint:1581
Change-Id: I7d8989a348ad1d8ca463dae90233fd82d6faf2d9
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/106849
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
