Mesa Refactors Disk Cache - Working Towards Windows Support

[Brane215]

Is Rust actually going to survive Mozilla’s demise?

I'm certain of that. Firefox ain't bad but Rust is friggin awesome.
I's destined to make waves in many segments, from microcontroller FW upwards.

[Teggs]

No. If you follow the link in to the Merge Request you will see my explanation.

"Note I have no intention of adding Windows support and there are likely opportunities to have better code sharing but I believe this a good start and it made sense to consider Windows support when adding greater abstraction for my own needs."

Sorry to take your time with speculation. I should have RTFM(R).

As a user of some of the software you work on, I appreciate your work... at least, as well as I can understand it. Please have a good day.

[tarceri]

Valve wants Mesa to work on Windows?

No. If you follow the link in to the Merge Request you will see my explanation.

"Note I have no intention of adding Windows support and there are likely opportunities to have better code sharing but I believe this a good start and it made sense to consider Windows support when adding greater abstraction for my own needs."

[illwieckz]

Valve wants Mesa to work on Windows?

I can think of a few reasons right off the bat that aren't crazy […]

I myself distribute a software with Mesa included on Windows (but disabled): see netradiant.gitlab.io/page/download/

The reason I ship mesa is for the software rasteriser. This make this software able to run with GPU with crappy Windows 10 stock driver or within virtual machines without 3D acceleration (user just has to enable the Mesa fallback and there you go). This software is not that hungry so software rendering is OK.

I use those builds of Mesa for windows: https://github.com/pal1000/mesa-dist-win/

[Teggs]

Valve wants Mesa to work on Windows?

I can think of a few reasons right off the bat that aren't crazy. The first one is the age of hardware Steam users employ and the sometimes crappy state of official vendor driver support for those devices on Windows. If Valve could migrate all their users to drivers Valve can read and modify, maybe it would result in less of a support burden?

Or that could be so much hogwash. Trying to intuit corporate strategies based on a few actions can lead to places almost as strange as the people who run corporations.

[skeevy420]

I remember times when the quality of forum posts were better ...

So does Pepperidge Farm

[skeevy420]

It is year 2020 and there is still no word on when they will migrate their stinking C codebase to Rust....

It is year 2020 and I still don't have my flying car they promised back in the 1960s so, I dunno, shit in one hand and migrate their C codebase to Rust with the other and, well, hmmm, see which one fills up faster???

[jrch2k8]

Rewriting code for the sake of rewriting code? You do that.

Especially since there is 0 benefit if your C/C++ is done carefully like mesa's where leaks are weird to find and you have an infrastructure for automated testing in place, that would also add a bunch of dependencies for no benefit and may hinder mesa support/performance on other architectures.

Anyway i'm pretty sure that was a joke from OP but with Rust evangelists logic and reason take a hike, so just in case

[tildearrow]

It is year 2020 and there is still no word on when they will migrate their stinking C codebase to Rust....

Instead of wasting time on this, they should be working on getting VirGL/Virgil/VirtualGL/whatevernameithas on Windows, which would be more, MORE useful as GVT-g does not work at all...

[Kemosabe]

I remember times when the quality of forum posts were better ...