<kyak>
is it about creating a better image of themselves? or are they legally obliged to publish parts of their kernel source?
eintopf has joined #qi-hardware
SopaXorzTaker has joined #qi-hardware
<whitequark>
kyak: neither
<whitequark>
apple routinely releases the core components of their systems under OSS licenses, and has been doing that for a very long time
<whitequark>
that's how we have e.g. a cross-toolchain from linux to macos
<whitequark>
they don't really benefit from it, not directly at least, so sometimes they're being really slow; parts of their systems require approval from legal
<whitequark>
btw that's not parts of the kernel source, that's pretty much the entire thing
sandeepkr has quit [Ping timeout: 240 seconds]
sandeepkr has joined #qi-hardware
<kyak>
whitequark: i see
<kyak>
how do we benefit from their kernel source though?
<kyak>
can user compile it and install instead of shipping kernel for OSX or iOS? I doubt that
<kyak>
even if they could, it still misses essential things: "You don’t get the apps, frameworks, user interface, etc."
<kyak>
apple doesn't really benefit from it, as you said. User don't really benefit from it. But it's right there on Github - and it makes we wonder what is the reason
<whitequark>
kyak: I benefitted from it when writing systems-level software for macOS
<whitequark>
also, security researchers can use it to improve security of iOS
<whitequark>
but there's no real reason other than "OSS software developed at Apple stays OSS"
<whitequark>
y'know sometimes companies have actual policies and not just greed?
SopaXorzTaker has quit [Ping timeout: 248 seconds]
<DocScrutinizer05>
isn't iOS mach based and mach would be FOSS?
<whitequark>
but in practice the published sources only let you build a macos kernel
<whitequark>
because the arm parts were non-public
<whitequark>
well, before a few days ago, that is
<whitequark>
the mach bits are actually relatively small and unused
<whitequark>
most of darwin is the xnu/bsd subsystem and you can no longer even create pure mach tasks in darwin
<whitequark>
it has *got* to have a xnu subsystem task
<whitequark>
bsd subsystem*
<whitequark>
this has really weird ramifications because, for example, inheriting a mach port requires an extremely bizarre dance that happens to work because of something that I think is an accident
<whitequark>
and this is a basic operation in a microkernel...
eintopf has quit [Remote host closed the connection]
zcrc has quit [Quit: AtomicIRC: The nuclear option.]