Basically it writes a small trampoline function in the stack that sets a "hidden param" to the nested function. Since it is stack allocated, once you exit the containing function it is deallocated. In your example, it worked by pure luck as you have yet to overwrite it
11
Replying to @lzsthw
This is UB, quoting the docs for the gcc extension: gcc.gnu.org/onlinedocs/gcc/N… "If you try to call the nested function through its address after the containing function exits, all hell breaks loose" eg look at what happens here: godbolt.org/z/EGh8hde5T & the result changes with -O3!
1
190
Replying to @nicovaoe
68
According to itch, they still allow downloads even after content notices if you bought the game before it was taken down. Contact support: itch.io/support (from itch.io/updates/update-on-ns… )
1
83
Replying to @dervish_candela
not a native either, as evidenced by my earlier typo on "appropriate". Maybe "tone" was the wrong word? I was thinking of statements like "Who does this provincialism serve?" or the others in this pic
1
1
30
a GBA emulator
11
Replying to @DominicTarason
They also have US patents for the same iirc patents.justia.com/assignee/…
107
Replying to @nice_byte
Nice! The docs linked in the readme seem to be dead
1
116
Lo unico que veo es esto
13
Replying to @nikc_12 @Zetta_330
Another thing it claims is that yuzu's telemetry will likely show yuzu devs knew that a lot of people were pirating games, because it would show people playing games before release (suck as TOTK)
1
64
Replying to @nikc_12 @Zetta_330
It also claims they published a patreon post showing Xenoblade running on yuzu the day before it released, but I don't know if this is just a timezone/region trick since they explicitly mention "release in the US"
1
1
69
Replying to @Zetta_330
The documents do mention this tho, that the devs shared ROMs amongst themselves. And iirc there are screenshots of public discord msgs in their discord from mods talking about a ROM "stash"
1
3
114
Replying to @migueldeicaza
Your posts show as unavailable to me when looking at responses
1
25
Replying to @kindasortamax
would you happen to have watched this vid already? Curious about your thoughts on it as someone who saw it all start to unravel youtube.com/watch?v=yDp3cB5f…
1
414
Replying to @FreyaHolmer
I can take another one if this one's not good enough
5
Looks like they killed it already
6
Replying to @GJSP_Kdash
It wasn't evolving for me even after walking manually, and I figured out why thanks to this info on a discord server. The count resets if it faints or you move it. The pokemon center roof trick still works.
1
20
Being judged as a person is far scarier than being accused of bad taste
34
TIL "buttresses" is a word
"10 minutes"

ALT Surprised Pikachu GIF

Replying to @fasterthanlime
The next time someone asks me why I dislike C++ I'm going to say "It's like the matrix" and refuse to elaborate
1
13
Do you have ClearType on? (cttune.exe)
Replying to @awr_hey
I once accidentally made this mistake in a rust github issue and was corrected by the author of ralfj.de/blog/2020/12/14/pro… Wanted to die of shame :P I'm now a firm believer of "Pointers proven nuanced"
1
Here, I fixed it
9
Replying to @Foone
The version mismatch was coming from inside the package manager
2
2
Replying to @psobot
Very simple explanations of hard things, thanks for sharing. Is it possible "source" and "destination" are backwards here?
Replying to @molecularmusing
10 Home 20H2 19042.1288 No console SDKs
I just saw that the docker docs for windows suggest using alpine-pkg-glibc. I opened an issue about it github.com/docker/docker.git…
1
I just ran into this C++ paper "No Sane Compiler Would Optimize Atomics" open-std.org/jtc1/sc22/wg21/… and I had to make this image
1
11
Replying to @badamczewski01
You're right, thanks! It still does 4 separate adds instead of 1 even when not using a try catch when using the parameter instead of a local, but the try catch code does look better. It seems to still spill even with the fix for x64, but still less so after the catch
1
It's really odd, even without the try catch using a local is better than a parameter: sharplab.io/#v2:C4LghgzgtgPg… And locally I'm seeing even worse codegen for the parameter version
1
Replying to @badamczewski01
From what I can see here sharplab.io/#v2:C4LghgzgtgPg… Both of them spill x before the try catch, the second one just gets better at codegen after the try catch (I'm assuming because it's another stack slot) I'm honestly surprised, I would have expected better codegen than this
1