Rusty wrappers are already marked `unsafe`, and by definition the raw
Vulkan function pointers used under the hood are `unsafe` too. Not
needing `unsafe` when going directly through `self.fp_v1_x()` is odd at
best.
* generator: Simplify contains+insert pairs to just insert()
Insert returns true when the value was inserted, false when it was
already present in the set.
* generator: Strictify "Vk"/"vk" prefix stripping with strip_prefix+unwrap
This way we are sure only the expected prefix ("Vk" or "vk") is
stripped, anything else panics the generator.
* generator: `format_ident` can format strings directly
As per [1] no explicit length field is available for `pSampleMask`, this
is based on `ceil(rasterizationSamples/32)` instead. It is valid to not
set an array for `pSampleMask` (under normal circumestances this is
signaled by setting the length to zero, and the array pointer is
ignored) but this has to happen by setting the pointer to `NULL`.
[1]: https://github.com/MaikKlein/ash/issues/256
Following [1] a stable release including the fixed commit has been made;
we can now depend on the released crate version again. This also
includes support for all new fields up to Vulkan 1.2.176.
[1]: https://github.com/krolli/vk-parse/issues/18#issuecomment-837815599
`()` is an empty tuple which can be constructed, while `ffi::c_void`
cannot. This is only ever used as an opaque pointer anyway and not used
by Ash; those wishing to access members directly should cast it to a
struct representation of choice (`winapi`, `windows-rs` or something
custom).
Fixes#156
* 1.2.171 new types
* generator: Keep platform_types checked in to git only
This static content is better kept in the platform_types.rs file only,
where it can be edited directly.
(Perhaps we should be more clear about generated versus manual files
within the ash crate, by storing generated outputs in an auto/
module/subdirectory that is reexported, like glib and friends?)
* Update Vulkan-Headers to 12.171
* Update Vulkan-Headers to 12.172
* Update Vulkan-Headers to 12.173
* Update Vulkan-Headers to 12.174
* generator: Generalize C number parsing
Co-authored-by: caradhras11@gmail.com <caradhras11@gmail.com>
* GetPhysicalDeviceProperties2: Make API functions public
Fixes: 5eb39fe ("Add VK_KHR_get_physical_device_properties2 extension (#400)")
* BufferDeviceAddress: Make API functions public
Fixes: 98d66c6 ("Add VK_KHR/EXT_buffer_device_address extension (#405)")
* GetMemoryRequirements2: Make API functions public
Fixes: d8d7423 ("Add VK_KHR_get_memory_requirements2 extension (#401)")
* Maintenance1/Maintenance3: Make API functions public
Fixes: a0a1f5d ("Add VK_KHR_maintenance extensions (#406)")
* Globally remove all `allow(dead_code)` exceptions
This is hiding the fact that some extension functions are inadvertently
not public and hence unusable by crate users, nor does it enforce clean
coding practices.
In addition remove `ash/src/allocator.rs` which does not appear to be
used (anymore?) and isn't in a working state anyway.
* Update Vulkan-Headers to 1.2.169
* generator: Add support for vkFlags64
Since Vulkan-Headers v1.2.170 with VK_KHR_synchronization2 there are now
bitmasks/enums using 64-bits instead of the default 32. vk-parse has
been updated to convey this info though the typedefs for these
enumerations could be parsed as well.
* generator: Insert underscores before trailing type number
Enum types like `VkAccessFlags2KHR` are turned into `VK_ACCESS2` after
demangling and `SHOUTY_SNAKE_CASE` conversion by Heck, but the enum
variants for the type start with `VK_ACCESS_2` (or similar) which fails
the `strip_suffix` and leads to long names to show up.
Inserting an underscore here makes sure the match succeeds.
* Update Vulkan-Headers to 12.170
* generator: Add Intel to the list of vendors for enum-postfix-stripping
* generator: Fortify enum-variant type prefix stripping
As brought up in [1] a new enum variant "pattern" not dealth with by our
type-prefix stripper went silently unnoticed. Explicitly listing the
few misnamed enum variants and panicking otherwise makes sure this won't
happen again.
[1]: https://github.com/MaikKlein/ash/pull/411#issuecomment-826013950
The docs clearly state:
#[rustc_deprecated(since = "1.42.0", reason = "use the Display impl or to_string()")]
fn description(&self) -> &str {
"description() is deprecated; use Display"
}
We already have a `Display` implementation containing an identical
`match` block and has further improvements on the way in [1].
[1]: https://github.com/MaikKlein/ash/pull/424
Previously, the `Display` impl for `vk::Result` did not include handling
for extension values. For example, `VK_ERROR_OUT_OF_DATE_KHR` would
display simply as `"-1000001004"`. Now, we fall back to the `Debug` impl
of the `Result` if the value is unknown (i.e. from an extension). This
preserves the current nice messages for non-extension values, and the
numeric output for truly unknown values, but displays the enum variant
name (e.g. `"ERROR_OUT_OF_DATE"`) for extension values.
* Fix clippy::use_self warnings in extension modules
* Deny clippy::use_self in extension modules
Note that clippy::use_self generates some false positives that must be ignored
* generator: Generate enums from vk_parse representation
This change prepares for future additions in vk_parse fields ([1]) by
converting over the enum generation path from vkxml. Most of the
conversion is easy by repurposing the existing `EnumSpec` parsing logic
from extension constants for all enumeration variants, with slight
modifications to not bail when `extends` is not set which is specific to
extension constants.
As an (unintended) added bonus this unification of the `EnumSpec`
codepath allows aliases (for backwards-compatible names) to be generated
as discussed earlier in [2].
[1]: https://github.com/krolli/vk-parse/pull/17
[2]: https://github.com/MaikKlein/ash/pull/384#discussion_r588693967
* generator: Turn "backwards"-mentioning docs into deprecation notices
All constant aliases for misspelled items (missing `_BIT` andsoforth)
contain the text "backwards compatibility" or "Backwards-compatible
alias".
* generator: Drop aliases whose name becomes identical after de-mangling
* generator: Remove aliases from const_debugs
These already have a match against the name they're aliasing, and some
of the aliases are "deprecated" (since they're just typo fixups for
backwards compatibility).
* Check for entry point in EntryCustom::new and return error
* Split off into load_checked
* Remove string from MissingEntryPoint
* Change MissingEntryPoint error message
* extensions/khr: Add ExternalSemaphoreFd safe wrapper
Implement the two small import and export and functions, part of the
VK_KHR_external_semaphore_fd extension.
* device: Pass ImageSubresource by immutable reference
It is only used as pointer by the native function, and should not (need
to) be moved.
* ci: Test docs in addition to `cargo t --all-targets`
Unfortunately docs are not explicitly (build-)tested as part of
`--all-targets` allowing broken code to slip in as shown by #390.
Also remove the `rust` listing type which is the default, leaving only
`no_run` (until the CI has a loadable Vulkan library).
* ash: Fix errors and warnings in (now tested) documentation comments
MemoryFdPropertiesKHR is a struct that, despite being used as output,
hsa to be initialized properly with a valid sType set, otherwise:
Validation Error: [ VUID-VkMemoryFdPropertiesKHR-sType-sType ] Object 0: handle = 0x7f035464d648, type = VK_OBJECT_TYPE_DEVICE; | MessageID = 0xdc6b3a0d | vkGetMemoryFdPropertiesKHR: parameter pMemoryFdProperties->sType must be VK_STRUCTURE_TYPE_MEMORY_FD_PROPERTIES_KHR. The Vulkan spec states: sType must be VK_STRUCTURE_TYPE_MEMORY_FD_PROPERTIES_KHR (https://www.khronos.org/registry/vulkan/specs/1.2-extensions/html/vkspec.html#VUID-VkMemoryFdPropertiesKHR-sType-sType)
All type parameters are implicitly bound by `Sized`; matching the `&`
out of `&str` implies that `str` must be bound by `Sized` which is not
true, resulting in:
18 | let entry = ash::Entry::with_library("foo")?;
| ^^^^^ doesn't have a size known at compile-time
|
::: /home/marijn/Code/TraverseResearch/ash/ash/src/entry_libloading.rs:73:39
|
73 | pub unsafe fn with_library(path: &impl AsRef<OsStr>) -> Result<Entry, LoadingError> {
| ----------------- required by this bound in `ash::entry_libloading::<impl EntryCustom<Arc<libloading::Library>>>::with_library`
After all `Library::new` accepts an `impl AsRef<OsStr>` without borrow,
which is what this function is modeled after to pass the same parameter
on directly.
Fixes: c6d5d66 ("entry_libloading: Provide Vulkan library loader from custom path (#319)")
* generator: Only replace "VK_" at the start of enum variant names
Replacing the occurence of "VK" is dangerous as shown by this diff; MVK
shorthands for MoltenVK are accidentally mangled. Using
strip_prefix/suffix is not only more correct and performs more runtime
checks, it is also more performant by returning a slice into the string.
* generator: Correct duplicate suffix stripping
The original `.replace(vendor, "")` operation and `.trim_end_matches`
replace every (trailing) occurence of the vendor, causing
DEBUG_REPORT_CALLBACK_EXT_EXT to end up without `_EXT` at all. This name
is also provided by vk.xml as fallback alias because it is used but
technically incorrect; the current generator can however not generate it
(limitation in vkxml representation).
Constants (enumeration bitflags) containing the word `external` were not
stripped of their common typename, and quick debugging showed the word
`VKTERNAL_...` in `struct_name`. `_EXT` would be the first extension to
match in `VK_EXTERNAL_...` messing up `struct_name` with a global
`.replace`, in turn not removing it from the variant name.
It isn't too uncommon to pass a single builder element into another
builder, and `slice::from_ref` fits perfectly while preserving lifetime
of the builder without calling `.build()`. We have used this
successfully in our codebase to uncover and prevent numerous
use-after-free bugs, and thought it's worth to share on the front page.
* khr/acceleration_structure: Fix fn signature of indirect build cmd
max_primitive_counts is an array of arrays, not an array of references
to single u32's.
* khr/acceleration_structure: Validate array length against geometry_count
This is a safe wrapper that already iterates over all the slice
references (fat pointers) to collect just the pointers for use in the
Vulkan API. It can and should at the same time make sure these slices
are of the specified length to prevent accidental out-of-bounds reads.
CString::new needs to perform an allocation to own the `&'static str`
and append a NULL byte to it. We can instead perform this "allocation"
in the generator and emit a byte-string literal with trailing NULL byte
that can immediately be used by the function loader.
