typed-geometry issueshttps://gitlab.vci.rwth-aachen.de:9000/ptrettner/typed-geometry/-/issues2019-04-11T13:58:09Zhttps://gitlab.vci.rwth-aachen.de:9000/ptrettner/typed-geometry/-/issues/1Universal constructors2019-04-11T13:58:09ZPhilip TrettnerUniversal constructorsvec/pos/size by anything vec/pos/size-like using templates
triangle by anything array-of-3-pos-likevec/pos/size by anything vec/pos/size-like using templates
triangle by anything array-of-3-pos-likePhilip TrettnerPhilip Trettnerhttps://gitlab.vci.rwth-aachen.de:9000/ptrettner/typed-geometry/-/issues/2Customization point for custom conversions2019-04-11T13:58:18ZPhilip TrettnerCustomization point for custom conversionsPhilip TrettnerPhilip Trettnerhttps://gitlab.vci.rwth-aachen.de:9000/ptrettner/typed-geometry/-/issues/3Improve contains for 3D triangle2019-04-11T14:17:50ZJulian SchakibImprove contains for 3D triangleRight now: assuming point is on plane of triangle, checking for counterclockwise triangle if point lies "inside".Right now: assuming point is on plane of triangle, checking for counterclockwise triangle if point lies "inside".https://gitlab.vci.rwth-aachen.de:9000/ptrettner/typed-geometry/-/issues/5Check default implementations that use "normal" because dir2 normal might mis...2019-04-26T07:12:26ZPhilip TrettnerCheck default implementations that use "normal" because dir2 normal might misbehave!Philip TrettnerPhilip Trettnerhttps://gitlab.vci.rwth-aachen.de:9000/ptrettner/typed-geometry/-/issues/6cylinder3 contains pos3 inaccurate2019-05-04T12:37:49ZJulian Schakibcylinder3 contains pos3 inaccuratecontains function for cylinder3 and pos3 very inaccurate near edges? e.g. a point ON a cap might be evaluated as not in the cylinder?contains function for cylinder3 and pos3 very inaccurate near edges? e.g. a point ON a cap might be evaluated as not in the cylinder?Julian SchakibJulian Schakibhttps://gitlab.vci.rwth-aachen.de:9000/ptrettner/typed-geometry/-/issues/4tg could use..2019-05-21T15:15:22ZJulian Schakibtg could use..maybe:
* closest points function for more objects
* centroid for more objects
* inverse for mat4maybe:
* closest points function for more objects
* centroid for more objects
* inverse for mat4https://gitlab.vci.rwth-aachen.de:9000/ptrettner/typed-geometry/-/issues/7scalar matrix multiplication2019-05-21T15:22:11ZJulian Schakibscalar matrix multiplication```
auto works = tg::mat1::identity * 2.0f; // works
auto noworks = 2.0f * tg::mat1::identity; // wont work
```
scalars can only be multiplied to matrices on the right. looking at the other matrix operations i guess you want it to stay ...```
auto works = tg::mat1::identity * 2.0f; // works
auto noworks = 2.0f * tg::mat1::identity; // wont work
```
scalars can only be multiplied to matrices on the right. looking at the other matrix operations i guess you want it to stay this way?Philip TrettnerPhilip Trettnerhttps://gitlab.vci.rwth-aachen.de:9000/ptrettner/typed-geometry/-/issues/8submatrix2019-05-22T13:05:55ZJulian Schakibsubmatrixunsure whether submatrix is working the way it should. the matrix returned is always of size SCxSR, however those are also used as the "end" row and column. that means you cant access the middle column for example.unsure whether submatrix is working the way it should. the matrix returned is always of size SCxSR, however those are also used as the "end" row and column. that means you cant access the middle column for example.Philip TrettnerPhilip Trettnerhttps://gitlab.vci.rwth-aachen.de:9000/ptrettner/typed-geometry/-/issues/9add reorient_normal_map2019-10-24T03:17:34ZPhilip Trettneradd reorient_normal_mapfor normal map blendingfor normal map blendingPhilip TrettnerPhilip Trettnerhttps://gitlab.vci.rwth-aachen.de:9000/ptrettner/typed-geometry/-/issues/10tg::always_false<D> partially breaks under clang 122021-06-14T12:53:36ZZain Selmantg::always_false<D> partially breaks under clang 12When using clang 12.0.0, all occurences of `tg::always_false<D>` where `D` is not a type cause a compilation error. I.e. D being an int for dimension checking.
This can be observed mainly for the `matrix/eigenvalues.hh` header, where it ...When using clang 12.0.0, all occurences of `tg::always_false<D>` where `D` is not a type cause a compilation error. I.e. D being an int for dimension checking.
This can be observed mainly for the `matrix/eigenvalues.hh` header, where it is frequently used to check the dimension of the matrix.
I've checked the following snipped using [godbolt](https://godbolt.org/z/cfTE1nP7K):
```
template <class...>
constexpr bool always_false = false;
template <int D>
void test()
{
if constexpr (D <= 4)
{
return;
}
else
{
static_assert(always_false<D>, "unsupported dimension");
return;
}
}
```
Starting with clang 12.0.0 the snippet is not valid anymore.https://gitlab.vci.rwth-aachen.de:9000/ptrettner/typed-geometry/-/issues/11`tg::comp<D, ...>::zero` is broken (missing in `special_values.hh`)2021-07-01T14:01:23ZZain Selman`tg::comp<D, ...>::zero` is broken (missing in `special_values.hh`)For `tg::comp<>` there is `static const comp zero;` defined (in `comp.hh`), which does not exist in `special_values.hh` unlike the `vec/dir/pos/etc`-types.
Usage like
```cpp
TEST("comp_zero")
{
LOG("hello: {}", tg::comp2::zero);
}
``...For `tg::comp<>` there is `static const comp zero;` defined (in `comp.hh`), which does not exist in `special_values.hh` unlike the `vec/dir/pos/etc`-types.
Usage like
```cpp
TEST("comp_zero")
{
LOG("hello: {}", tg::comp2::zero);
}
```
results in a `error: undefined reference to 'tg::comp<2, float>::zero'` error.Philip TrettnerPhilip Trettner