Notices by mittorn (mittorn@masturbated.one)

#jpeg #dct #compression #lossy #jpg #h263 #h264 #avc #mp3 #mp4

@coaxial @a1ba hdr хуле

@a1ba @coaxial неужели у apple штатные мониторы сидели на аналоге? Или может там просто какой-то поломанный протокол?

@a1ba @coaxial т.е цифра была... А откуда инфа,что не было?

@mischievoustomato @a1ba not very fast because DPI causing drops and slowdowns

@a1ba @Suiseiseki Temporary failure in name resolution error

@a1ba

@jonty i remember soviet cassette deck "Mayak" with two same DINs on back panel. First one is AUX audio input, but second is some remote control connector, which is wired to some 150V circuit.
Of course i used wrong connector and capacitor on portative mp3 audioplayer just exploded. But after replacing capacitor it worked again

@feld so why not ditch gitlab, while gitea/forgejo exist?

@feld so, main issue is account migration?

@SuperDicq just change your user-agent to something without Mozilla

@a1ba @phnt @SuperDicq @piggo @sun @birdulon @foxido @mischievoustomato @lina we need egl and vulkan wsi over dmabuf implementation.
Now almost every EGL driver implements device platform, which is enough to fully implement wayland and x11 platforms over it. But instead we are using internal x11/wayland platform implentations. Same for vulkan wsi. Even vulkan-wsi-layer requires internal VkSurfaceKHR implementation, that hides dmabuf or nvidia internals under it.

@a1ba @uis @lanodan @thesamesam anyway, safe fetching should always include checksums and fetch should not be done in configuration stage, it must be separate stage. And FetchContent is worse than calling wget because it's standart feature (so cmake at some point advertise using it)

@lanodan @uis @thesamesam @a1ba or only one yet.
autotools at least does not have FetchContent, while cmake may SILENTLY download something. This might be good for separate task, not during configure

@a1ba what about microsoft?

@a1ba @prettygood is it still not fixed?
Anyway, it seems, you lost actor's private keys, so it possibly will not work correctly with old instances

@a1ba @prettygood blockers gonna block

@a1ba @th3rdsergeevich @ZySoua only thing Nintendo should is ИСЧЕЗНУТЬ ВОТ ОТСЮДА С ЭТОЙ ПЛАНЕТЫ

@lanodan I keeping boost in mask too, only unmasking when building specific packages. This does not keep me from ABI breakage, but helps reduce boost-depended packages count and prevent rebuilds
Now i installed boost to build inkscape, just to test some features and it will be removed after, most time my gentoo installs work without boost and even without gtk3
Also, while boost builds remotely with distcc, it's buildsystem eats 100% CPU. How???

#boost is shit, do not use it!

Portage says:
* Boost.Regex is *extremely* ABI sensitive. If you get errors such as
*
* undefined reference to `boost::re_detail_108500::cpp_regex_traits_implementation
* <char>::transform_primary[abi:cxx11](char const*, char const*) const'
*
* Then you need to recompile Boost and all its reverse dependencies
* using the same toolchain. In general, *every* change of the C++ toolchain
* requires a complete rebuild of the Boost-dependent ecosystem.
*
* See for instance https://bugs.gentoo.org/638138