# Gitlab availability part 2

**URL:** <https://discourse.vtk.org/t/gitlab-availability-part-2/16535>\
**Category:** Support\
**Created:** [September 16, 2026, 4:26pm UTC](https://discourse.vtk.org/t/gitlab-availability-part-2/16535 "2026-09-16T16:26:12Z")\
**Posts on this page:** 1\
**Showing post:** 39

<div class="post-metadata">

**Author:** ![jaswantp](https://discourse.vtk.org/user_avatar/discourse.vtk.org/jaswantp/32/10046_2.png) [@jaswantp](https://discourse.vtk.org/u/jaswantp)\
**Post date:** [October 23, 2026, 9:38pm UTC](https://discourse.vtk.org/t/gitlab-availability-part-2/16535/39 "2026-10-23T21:38:02Z")

</div>

> [@banesullivan](#):
>
> The new backend changes the GPU API but keeps VTK’s mappers, modification-time tracking, text-spliced shaders and object model, and those are the layers that impact the kind of work I’m trying to do.

yes, the text-spliced shaders are annoying. Atleast in the webgpu backend, string-substitution is being replaced with [slang shading language](https://shader-slang.org/) that allows polymorphism in shaders. The work is tracked at [https://gitlab.kitware.com/vtk/vtk/-/work\_items/19731](https://gitlab.kitware.com/vtk/vtk/-/work_items/19731). What are the problems with the mappers, MTime tracking and object model that affect PyVista? Yes, I understand, they get in your way sometimes as middleware developer, but that is generally not a great pain once you’ve built the right abstractions.

> [@banesullivan](#):
>
> I see VTK as an excellent reference implementation (correct, high fidelity), kind of like how there are many CPU-based ray tracing implementations used as a reference when developing new GPU-accelerated methods.

There are cases where even VTK produces incorrect results, and we’re improving them. For example: [Say goodbye to lines bleeding through thin walls with polygon offset - #3 by michaelhogg](https://discourse.vtk.org/t/say-goodbye-to-lines-bleeding-through-thin-walls-with-polygon-offset/16561/3).

---

_[View the full topic](https://discourse.vtk.org/t/gitlab-availability-part-2/16535)._