it seems possible to install GfW into an existing MSYS2:

https://gitforwindows.org/install-inside-msys2-proper

Right, and manually extracting MSYS2 packages into an MSYS2 environment is also possible if very unsupported (used that to get a copy of rsync once). My point, rather, is that GfW isn’t and shouldn’t be your starting point if you want to get a Unix-like environment; use actual MSYS2 instead.

microsoft introduced Dev drives for this reason, they are optimized such that no file system filters are attached

https://learn.microsoft.com/en-us/windows/dev-drive/

I thought by filters they quite literally mean the antivirus filters in Defender? Does it strip down some of the Windows file semantics too?

File system filters are pluggable kernel drivers. For example, ProcMon (Sysinternals tool) monitors file systems via a file system filter -- so fun fact, if you outright disable file system filters on a ReFS volume, you won't get any ProcMon results! This was a 'duh' moment for me.

https://learn.microsoft.com/en-us/windows-hardware/drivers/i...

one might argue that one of the advantages of languages like C is that you only pay for the features you choose to use, no unnecessary overhead like initializing unused variables

You can pay for those features in debug mode or in chaos monkey mode. It's okay to continue to not pay for them in release mode. Heck, Rust has this approach when it comes to handling integer overflow - fully checked in debug mode, silent wraparound in release mode.

In Ada you can pay for integer overflow checks (runtime) if you want to. With Ada SPARK you can prove that your code does not contain integer overflows so that you don't need runtime checks.

And you can disable these checks with a flag when it comes to Ada, and yeah, with SPARK, none of it happens at runtime.

Check the table at https://docs.adacore.com/spark2014-docs/html/ug/en/usage_sce..., look for "SPARK builds on the strengths of Ada to provide even more guarantees statically rather than dynamically.".

More reading:

https://docs.adacore.com/spark2014-docs/html/ug/en/tutorial....

https://learn.adacore.com (many books for learning Ada and SPARK) available in PDF, EPUB, and HTML format.

However, the compiler does not tell you this. We're back to the problem that it's possible to have a "working" C program that relies on UB and will therefore break at some point, but the tools will not yell at you for doing this. Whereas in Java or C# you get warnings or errors for using maybe-uninitialized variables.

Also, scanf should be deprecated. Terrible API. Never use scanf or sscanf etc. We managed to get "gets()" deprecated, time to spread that to other parts of the API.

atoi() or atof() etc. work OK, but really you need a parser.

and here I am still playing Civ II MGE

https://github.com/FoxAhead/Civ2-UI-Additions

I use this handy bookmarklet to kill sticky headers::

https://github.com/t-mart/kill-sticky

While OpenCV was born at Intel, I don't think it's solely developed by Intel anymore, there's now a non-profit organization in charge of the project:

https://opencv.org/opencv-platinum-membership/

the cache being talked about is for recently/frequently accesses things, not stuff pending write

first thing I did before reading the article, using uBO to block JS on the page

This is the weirdest video I ever watched. It's like Black Mirror ... but in real life ... and a somewhat happy ending.

I agree, the visualization is pretty but kinda useless to explain mat-mult.

It would have been more intuitive to show every element in output matrix corresponds to a dot-product of row/column vectors from input matrices, the animation doesn't even highlight those corresponding vectors clearly..