[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Gentoo guys starting a fork of udev

On 11/15/2012 07:57 AM, John Paul Adrian Glaubitz wrote:
> On Nov 14, 2012, at 6:43 PM, lego12239@yandex.ru wrote:
>>> Modern computer systems are much more versatile and complex than they were at the time when System V Init was conceived.
>>  Some things must be as simple as possible even today.
> Care to elaborate why? To save memory on an 8 GB workstation? Even the 25 US$ Raspberry Pi has enough power for systemd.
> Are you also choosing FAT32 over ext4 because it is simpler?
> Why are we having Fibre Channel support in the kernel? Why does the kernel include a virtual machine hypervisor? Why do we support IPv6?
> We could just go back and stick with our good old SunOS 4 boxes.

Would you care reading what's being written to you?
Roger didn't write that the init system should be kept
without features, using old age techniques. He wrote
that he believes that *pid 1* should be kept simple,
and that the complexity should go somewhere else
(like, in a fork of PID 1, for example). That is very
different from what you write above.

>> People are constantly insisting that systemd is too bloated or unreliable, but yet no one has really come up with real examples to prove that.

I'm sorry but the sentence above is just plain wrong.
See Linus post, and all the things that udev broke.

>> Yes, the core binary of System V Init is smaller than systemd's. However, System V Init needs a lot of bloat in form of hacky bash scripts using even more external tools like sed and awk to be actually useful in any regard.

But these hacks / bloats will not ultimately result in a
kernel crash.

>> And I think it makes way more sense to have all the functionality of the init system


>>  integrated into it's core binary

And no! :)

It would have been possible to do the right thing (tm)
and have both feature and reliability. Currently, we
only have the former, which is due to the design.


Reply to: