Notices by Wolf480pl (wolf480pl@mstdn.io), page 2

@lanodan @ayo
ok but like

we're not waiting for pictures to load these days anymore, right?

Now, on all these pro websites, we're waiting for text to load.

@lanodan @ayo idk, I've only seen a few pictures load like that before 2010, and IIRC they were on a forum / imageshack / etc so probably whoever uploaded them didn't know any better...

Idk how often one would see that on professionally made websites...

Though that could be anoter axis around which tables flipped - nowadays only hobby websites load fast xD

@ayo these placeholders show how much we've progressed since the 90s...

@ignaloidas @lanodan @ayo
they have better abstraction than C, and you could pull that off with C as long as you only pass it by pointer...

also, IIRC Zig does have did make an abstraction for Allocators? So I think that proves it's possible

@ignaloidas @lanodan @ayo
You mean if you never have to write an allocator implementation?

What's wrong with manually written out VTables anyway? All abstractions in Zig work that way AFAIU, and that's also how they work in many C projects, such as Linux kernel...

what makes it not an abstraction?

@lanodan @ayo
actually

You don't need to know the size.
The "size" could only encompass a vtable and a "private data" pointer pointing to the actual data structure.

@lanodan @ayo that's only an issue if you don't require all dependencies to be available at build time

@lanodan @ayo
though it would suck if there was eg. an abstract list, but getting an actual implementation of a list would require implementing the rest of the list yourself, or using a 3rd party library

@lanodan @ayo
arguably, it's enough for the stdlib to standardize datastructures' API, not necessarily implementation :P

@lanodan @ayo
it's implied: no longer used by the compiler

@ayo hey, at least they remember when a thing was done and undone and done again.

That's better than 80%* software projects...

*made up number

@sun
is the description full of affiliate links and other jump that makes it hard to find anything n it?

@BrodieOnLinux
btw there are also partuuids in the GPT partition table if you're using that, you may want to check those as well.

@tk @piggo
Or all the interior plastics. Gotta love the extra harmonics.

@quad
like if the game has a stupid shader

and the driver replaces that with a shader that is written in a non-stupid way

why not send the non-stupid shader to the game developers so that they can include it in the game?

@quad I'd understand this dynamic if the optimizations were added to the driver after the fact

but like

if they're available on the day the game is released

that means the GPU vendor was working with the game developers

so why not put the code in the game?

I'm still perplexed by the idea that GPU drivers contain per-game optimizations and this is considered normal and desirable.

Linux Mint Troubleshooting Guide doesn't seem to know the difference between bugs reports and tasks

> Bug reports are only valuable when they lead to a bug fix or an improvement.
> The reason we report bugs is not to document or catalog anything that is wrong with the software

@gvlx just checked out the vanilla icecream story and it's a great example!

And yeah, I agree, bugs should be reported.
1/

@gvlx see also, @simontatham's blog post about separating facts from plans in a bug tracker https://www.chiark.greenend.org.uk/~sgtatham/quasiblog/bugtracker-separate/