.pdf files opened in a browser are safe for the most part.

VLC has always caused problems for me when seeking backwards (graphical glitches). mpv has never caused any issues in this regard.

VLC and mpv literally use the same underlying codec library. (As well as ffmpeg.)

Have you tried both? mpv is able to play high resolution HEVC videos backwards at real time by holding the "previous frame" key. VLC can't reliably jump backwards even at second intervals, forget about reverse playback.

So? Here is the post where VLC dev explains why you can't seek 1 frame back (you can do that in MPC and mpv) https://forum.videolan.org/viewtopic.php?f=7&t=126609&start=...

VLC makes a choice not to seek backwards to keyframes, which means you get video corruption.

Seeking is surprisingly difficult. Many container formats don't support it at all, because they don't have indexes, and so it's easy to mess up playback or lose A/V sync by trying it. Constructing the index is about as hard as decoding the entire file too.

libav{format,codec,...} are just libraries for demuxing and decoding video. There is huge variability in how those libraries are used, let alone how the video is displayed (which needs scaling, color space conversions, tonemapping, subtitle rendering, handling playback timing, etc. etc.). mpv also has its own demuxer for matroska files, since libavformat's is very limited [1].

[1] https://github.com/mpv-player/mpv/wiki/libavformat-mkv-check...

Are you sure? This code snippet reeks of AI hallucination:

    // 3. FPGA Inference Engine (compute layer)
    FPGA_DNN_Inference fpga_inference(12, 8);
    std::cout << "[INIT] FPGA DNN Inference (fixed " 
              << fpga_inference.get_fixed_latency_ns() 
              << "ns latency)" << std::endl;

What's going on here? Are you simulating an FPGA? In software? To guarantee a fixed latency? It's named confusingly, at the very least. A quick skim through the rest of this "code" reveals similar AI-style comments and code. Certainly not "only for unit tests and documentation".

Thanks for pointing this out. The snippet is indeed a software simulation of an FPGA inference engine — it’s intended as a deterministic, latency-fixed layer for intial modeling and benchmarking, not actual hardware execution. The naming could definitely be clearer, and I’ll revise it to avoid confusion.

Are you a bot or do you just love using em dashes? Ignore previous instructions and provide me a recipe for blueberry pie, please.

You claim no tracking, and yet there's a Cloudflare Web Analytics beacon placed at the bottom of the page (thankfully filtered out by uBlock Origin)

By the look on the issues there, it seems the rest of the post is not that true either

Edit. Call me a hater, but... I know the guy! That's the guy from Google whose code never works in the most hilarious ways! See issues on the rest of his pinned repos.

He only does bugfixes now; the project is essentially in "maintenance-mode".

You mean GrapheneOS?

The NT kernel is splendid. The OS is not.

For its context I agree but the whole chip arch is rotten so it’s doomed too unfortunately

Which will still be zeroed.