Cleaning up unused code

Bug: b/153177839
Change-Id: I42479778f0f6ec9a955ba7f863766429b68c73a3
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43468
Presubmit-Ready: Alexis Hétu <sugoi@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Alexis Hétu <sugoi@google.com>

These were added for the .gn build, but were never enabled for the CMake build.

Test these as part of the Kokoro presubmits.

Bug: b/153193374
Change-Id: I36e7428bfe1a4867bea39b2fbb302aed6c3dff7b
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43488
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

Bug: b/153168751
Bug: b/153206490
Change-Id: I2822f60aa380ae3885bc674d9f7b1f283c087e95
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43490
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

Bug: none
Change-Id: I9f58cc3bf6b457fc8b916a98242138865f89c577
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43469Reviewed-by: Ben Clayton <bclayton@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>

I broke this when I removed the relative path to the third_party astc
include file in this change:
https://swiftshader-review.googlesource.com/c/SwiftShader/+/43269

Bug: none
Change-Id: I45de8a6159c69c893df4449f6c6e10b09a9e6fe4
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43448Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>

When `ENABLE_VK_DEBUGGER` is defined, VkDevice needs to be declared in order to build.

Bug: b/148401179
Change-Id: If35f3a1425ec0d739750d632305a625c7e12b721
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43432Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

`VirtualFile`s are typically used for files with no physical backing.
Use `PhysicalFile` if there's no content provided in the debug info.

Bug: b/148401179
Change-Id: I1e1ed75ce9a9fb0965420267aba7f6d8de958012
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43430Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

This is common for virtual and physical files.

Bug: b/148401179
Change-Id: I266097382098c1daf776ae64d878f08febee5d0e
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43429Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

https://github.com/KhronosGroup/SPIRV-Headers/issues/137 is still ongoing.

Fixes: b/15093697
Change-Id: I2498f35a8049fb2a1a86aa0efe6fbd6e59ad6493
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43428Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

Bug: b/145758253
Bug: b/152606052
Change-Id: I826225ed393ed324301555ebaea14947997632cf
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43289Reviewed-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>

Also remove dependency of third_party/astc-encoder on
src/System/Debug.hpp.

Bug: b/145758253
Change-Id: I329928d9221c99a416fcdb1b0dd44d2578a22cb8
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43269Reviewed-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

Remove unused and useless variables. Also fix the build when
SWIFTSHADER_ENABLE_VULKAN_DEBUGGER is enabled.

Bug: b/145758253
Change-Id: Ie49dd88d6716b24fe31baa8f8d162e7dacfa5a3c
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43268Reviewed-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

Bug: b/145758253
Change-Id: I36f1bbf599389c27e084fea9af3d732225567076
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43230Reviewed-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>

This adds a third_party/boost folder with a CMakeLists.txt. If this
folder is added via add_subdirectory, it downloads boost to the
binary directory and exposes the Boost::boost target.

Bug: b/145758253
Change-Id: I7d4e383630cb84da928cf0622a641e9e0efc2ad8
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43191Reviewed-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>

This change also fixes compilation on Linux when REACTOR_EMIT_DEBUG_INFO
is enabled, which has been broken since "-fno-exceptions" was added to
SWIFTSHADER_COMPILE_OPTIONS. Also, at some point,
LIBBACKTRACE_CONFIG_DIR stopped being used, which further broke
compiling libbacktrace.

Bug: b/145758253
Change-Id: I50f0e2c2c1dabc357dfc255a471eaf1e9fed1c2e
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43189Reviewed-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>

Bug: b/145758253
Change-Id: Ife32c322cf87c9b1b98b4b50f591b75db6ea51e2
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43148Reviewed-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>

Bug: None - Fix Kokoro presubmits.
Change-Id: I16074e733edd257647840d1bf51ebdc14dd4fb89
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43431Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

This code has no remaining known active uses. It has been archived in
the 'legacy-d3d9' branch.

Bug: b/139189696
Change-Id: I64bf27ff6c02be560963d7eab36c9c328bb280c5
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43232Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

CMake should be used instead.

Note that the CMake build does not support the D3D9 and D3D8 targets.
They can still be built with Visual Studio 2019 from the 'legacy-d3d9'
branch.

Bug: b/139189696
Change-Id: I7b5006fa98e479d6d1a3bb75d379a5818a560c20
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43231
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

cf. https://swiftshader-review.googlesource.com/c/SwiftShader/+/43388

Bug: build fix
Change-Id: I8d2acc0bbb19fa71748aafeebd6c0d500673ed93
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43408Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>

Spotted by Chris Forbes:
* `std::remove_if` wasn't introduced in C++20, it's been there since 98. I just can't read docs.
* Also fix the lack of `list.erase()`.

Bug: None. drive-by-code-reviewing-fixes.
Change-Id: I456d540ba973640747d6e4cb23faec65a4a9c83d
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43348
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

It was only still being used to store framebuffer blend constants.
Use sw::float4 instead.

Bug: b/146224130
Change-Id: Ie5660e237f4b675564d5a492f9d0a38505d07953
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43388Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

This CL optimizes C++11 range-based for loops where the variable is copied in each iteration but it would suffice to obtain it by const reference.  This is only applied to loop variables of types that are expensive to copy which means they are not trivially copyable or have a non-trivial copy constructor or destructor.

To ensure that it is safe to replace the copy with a const reference the following heuristic is employed:
  The loop variable is const qualified.
  The loop variable is not const, but only const methods or operators are invoked on it, or it is used as const reference or value argument in constructors or function calls.

See cl/304001946 for more information.

Bug: None.
Change-Id: Idbb5a0dbf19c6dccb16a0d525900aa99419a8527
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43368
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

Erroneous ImageView size computation was causing out of bounds
memory accesses whenever an ImageView only represents a subsection
of an Image. The getSizeInBytes() function now only returns the
size in bytes within the Image object from the beginning to the
end of the bytes used by the ImageView.

Bug: b/150464740
Change-Id: I0985af9cdfbb85b10336a7691f76009496bb2ae5
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43329
Presubmit-Ready: Alexis Hétu <sugoi@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Alexis Hétu <sugoi@google.com>

If layer.device is null, [layer nextDrawable] returns nil, so on surface
initialization we set layer.device to a dummy MTLDevice by getting a
reference of the system default device.

MTLCreateSystemDefaultDevice() requires linking against
Metal.framework but Chromium needs to ship on platforms that don't have
Metal. This means we can't directly link against the framework, instead
we "weakly" link against it so that if it is not present, the function
pointers are just null (instead of failing to launch).

Bug: dawn:269
Change-Id: I5984613f576849c6f219c253da7eb277fa75ef52
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43209
Kokoro-Presubmit: Corentin Wallez <cwallez@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Corentin Wallez <cwallez@google.com>

invertCommon looks for tree nodes with the majority of the child nodes with the same spans.
This span is promoted up to the parent, and the children have the
span inverted.

This causes the span IDs to flip (inclusive / exlusive) each time they're seen while decending the tree. This adds a bit of complexity to the parsing of the data, but the file size reduction is significant.

Bug: b/152192800
Change-Id: I5ec294d42004d936004c27ef15bf94a7143372ba
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43311
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

There's more to come.

Bug: b/152192800
Change-Id: Icc0928fa0051269f9ca14e9177a9a8ef56d05a34
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43310
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

Update tests.

Bug: b/152192800
Change-Id: I4e57f0d2b049f0bce7459b29f7ffc3a2ca03534a
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43309
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

This was getting unruly.

Bug: b/152192800
Change-Id: I41a393e0548367ea4d69c3982b89a6ffe198e9f7
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43308
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

Compress common spans into groups and index these.
Reduces the full coverage json from ~500MB to just under 100MB.

Use zlib compression on the final file instead of a zip. The compression algorithm is the same, and produces roughly the same sized output, but a raw zlib file is much quicker to decompress with pako instead of jszip.

Bug: b/152192800
Change-Id: Id444ddf4027b6ace03570719f159ae14e481cd37
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43249
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

And add test to check it behaves as expected.

While currently nothing golang actually parses these files, this allows me to experiment with compression schemes on the full data set without needing to rebuild the coverage data from scratch each time.

Bug: b/152192800
Change-Id: I8070abe2bbebaef3e5889991060f1d1342078eb2
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43248
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

Bug: b/152777669
Change-Id: I271ca0ecb5ff8461530fd6388378524970a2bc2b
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43228
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

We now have an up-to-date third_party/PowerVR_Examples submodule for
these sample applications.

Bug: b/141361933
Change-Id: I216b94d0c3263bf85a6b0a7fca52a2e5b7beefa2
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43190
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

Also add some additional Vulkan samples to the 'good' set.

Bug: b/141361933
Change-Id: I32fe2d691ca2fd176a8c470fdb34b25a55840cd6
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/42389
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

Immutable textures created with glTexStorage*() are always mipmap and
sampler complete. The spec states that

"Array levels k where k < levelbase or k > q arebinsignificant to the
definition ofcompleteness."

Where q is no larger than levelmax, which is in turn limited to
levelimmut - 1 which was specified at glTexStorage.

Change-Id: Ieaa6be856fd35cf2827e753d7c777ebf784027ef
Tests: dEQP-GLES*
Fixes: b/152740217
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43188
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

This reverts commit 81d8c2ad.

Reason for revert: Breaks macOS swiftshader build

Change-Id: I1c54aaf772f651d731d90e843c85549a4f7a38b8
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43208Tested-by: Corentin Wallez <cwallez@google.com>
Reviewed-by: Corentin Wallez <cwallez@google.com>

If layer.device is null, [layer nextDrawable] returns nil, so on surface
initialization we set layer.device to a dummy MTLDevice by getting a
reference of the system default device.

Using MTLCreateSystemDefaultDevice() requires linking against
Metal.framework but Chromium needs to ship on platforms that don't have
Metal. This means we can't directly link against the framework, instead
we "weakly" link against it so that if it is not present, the function
pointers are just null (instead of failing to launch).

Bug: dawn:269
Change-Id: I8719c45a19718ff79ef21f47515fe1c15b99628b
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43112
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Corentin Wallez <cwallez@google.com>

The getOrCreate() method takes care of the mutex locking and unlocking,
and takes the function for creating the routine as a callback argument.

Bug: b/131246679
Change-Id: I0873f6be94c92a11181ce575d1b44ed040177d43
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/42869
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>

If the PowerVR_Examples submodule has been checked out, enable creating
the CMake targets for them. This helps make them available in all
configurations in Visual Studio, without editing the cache for each of
them individually.

Bug: b/141361933
Change-Id: I700407be4c6b89866856b0c4f2d6ae55180c4fcd
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/42488
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>

Push this to https://github.com/swiftshader-regres/swiftshader-coverage so that it can be viewed at:
https://swiftshader-regres.github.io/swiftshader-coverage/

There's quite a lot of new code and fixes in this change. The most notable:
* The regres daily run for the subzero backend now produces combined coverage information for all the test runs. The LLVM backend does not produce coverage information.
* Regres now takes two additional command line arguments: `gh-user` and `gh-pass` for the swiftshader-regres account. If you omit these, then coverage will not be produced.
* test.srcDir has been renamed to checkoutDir, as this was confusing with the `src` directory in the repo.
* The coverage JSON now contains a root field to describe the git revision to which it relates. This prevents the coverage going out of sync with the source.
* `git.CheckoutRemoteBranch()` drops back to a depth of 1 again. This was only increased to 99 to deal with issues checking out from gitlab, which we don't do any more.
* Regres  now builds using `third_party/llvm-10.0`
* Fixed the `--limit` regres command line flag which wasn't actually limiting, as it was using the len() on the number of groups, not the number of tests.

Bug: b/152192800
Bug: b/152339534
Change-Id: I2d25735f485097d4efb080546d989056a3a8aab3
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/43168
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Ben Clayton <bclayton@google.com>