Specifications developed in PR #3206 require that groundcover content files must not be allowed to corrupt normal content files. With this PR we simply isolate our existing loading logic by instantiating a separate `ESMStore` for `Groundcover`. In addition, we remove some outdated workarounds.
In file included from /<<BUILDDIR>>/openmw-0.47.0+git202111080927~ubuntu18.04.1/components/detournavigator/findsmoothpath.cpp:1:0:
/<<BUILDDIR>>/openmw-0.47.0+git202111080927~ubuntu18.04.1/components/detournavigator/findsmoothpath.hpp:93:48: error: field ‘mSettings’ has incomplete type ‘std::reference_wrapper<const DetourNavigator::Settings>’
std::reference_wrapper<const Settings> mSettings;
^~~~~~~~~
In file included from /usr/include/c++/7/bits/move.h:54:0,
from /usr/include/c++/7/bits/nested_exception.h:40,
from /usr/include/c++/7/exception:143,
from /usr/include/c++/7/ios:39,
from /usr/include/c++/7/istream:38,
from /usr/include/c++/7/sstream:38,
from /<<BUILDDIR>>/openmw-0.47.0+git202111080927~ubuntu18.04.1/components/detournavigator/dtstatus.hpp:6,
from /<<BUILDDIR>>/openmw-0.47.0+git202111080927~ubuntu18.04.1/components/detournavigator/findsmoothpath.hpp:4,
from /<<BUILDDIR>>/openmw-0.47.0+git202111080927~ubuntu18.04.1/components/detournavigator/findsmoothpath.cpp:1:
/usr/include/c++/7/type_traits:2125:11: note: declaration of ‘class std::reference_wrapper<const DetourNavigator::Settings>’
class reference_wrapper;
This PR aims to solve all issues with `Groundcover` view distance handling in a satisfying way while preserving other optimisations that benefit other features. The main idea here is not to rely on `ViewData` updates for distance culling calculations because we can not accurately determine distance against lazily updated views. Instead, we perform an accurate measurement in a cull callback we can run every frame in `Groundcover` itself. While we do have to add some code to handle this feature, it is quite loosely coupled code that could be useful elsewhere in the future. These changes should address a performance regression @akortunov experienced.
To minimize overhead for calls when level is less than current. For example
Log(Debug::Debug) should not lock mutex when current logging level is Verbose.
This PR aims to solve `uniform block LightBufferBinding has no binding` messages @glassmancody has reportedly encountered since PR #3110 due to an apparent bug in OSG. While we do have to add a workaround here that adds a bit of clunkiness, #3216 should allow us to clean up these interactions a bit in the future.
OscillatingRecastMeshObject::update should be called with tile bounds in real
coordinates not in navmesh. But proper scaling was done only in
RecastMeshManager::getMesh and RecastMeshManager::updateObject used tile bounds
in navmesh coordinates.
Add a new function to create tile bounds with proper scaling and pass correct
value into RecastMeshManager constructor through CachedRecastMeshManager
constuctor from TileCachedRecastMeshManager member functions.
With this PR we refactor `StringUtils::replaceAll` to accept `string_view` as suggested in a code comment. In addition, while we are touching this rebuild happy file, we slim it down a bit by moving a few sparingly used functions elsewhere.
This PR removes dummy serialisers for `StateSetUpdater`, `NodeCallback` and the respective `META` macros that trigger serialisation requirement here.
`StateSetUpdater` and `NodeCallback` are just base classes that can not be used on their own, so there is no need to incorporate them into serialisation. These changes might have minor effects on derived classes that forget to override `className()`, `libraryName()` through `META`, but it makes hardly a difference to now serialise such classes as a dysfunctional `osg::Callback` instead of a dysfunctional `SceneUtil::NodeCallback`.
This PR removes unneeded `lowerCaseInPlace` calls in in a hot path of `objectpaging.cpp` that are no longer necessary after PR #3197. In addition, I have been informed that these changes should by coincidence address a compiler specific compilation error we currently experience.
We currently apply a strange algorithm to `LightManager::mStateSetCache`. For some reason this algorithm inserts hashed keys into `std::map` in a way that fails to handle hash collisions and exhibits worse lookup complexity than `std::unordered_map`. With this PR we just use `std::unordered_map` here.
This PR aims to start addressing `ESM` design issues that have silenced errors we incorporated into groundcover `ESM` loading approaches.
- We move the resolution of `parentFileIndices` from `ESMStore` to `ESMReader` as suggested in a `TODO` comment.
- We improve a highly misleading comment which downplayed the significance of `parentFileIndices`.
- We document important preconditions.
- We move a user facing error message to the highest level and improve its context.
- We remove an inappropriate `setGlobalReaderList` method. We now pass this reader list into the method that requires it.
- We remove a thoroughly pointless optimisation of `Store<ESM::LandTexture>`'s construction that has unnecessarily depended on `getGlobalReaderList`.
There should be no functional changes for `master`, but this PR should remove an issue blocking PR #3208.
With this PR we apply `lightMask` to a `Transform` node we create specifically for a light. This mask will allow us to stop traversing such nodes sooner and avoid costly processing associated with `Transform` nodes in the cull visitor.
I have been informed by @akortunov that my addition of `Groundcover::getViewDistance` is not working in some cases. Investigations revealed that some `ViewData` code interacting with my additions had been quite thoroughly optimised in a way that was not sufficiently documented and interfered with the new feature. With this PR we repair `getViewDistance` while preserving such optimisations and add a necessary comment to avoid issues in the future. In addition, we now rebuild views when the global `mViewDistance` changes.
With this PR we refactor `SceneUtil::KeyframeController` not to require `virtual osg::Callback` inheritance. I suppose such `virtual` overhead is not justified here because it negatively impacts many other classes we derive from `osg::Callback`.
