Re: Really, about udev, not init sytsems

[John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>]

On Sun, Nov 25, 2012 at 10:16:27PM +0800, Thomas Goirand wrote:
 > >> Yes, lots of
> >> udev stuff are moving to /usr, and this is a fact. Yes, lots of
> >> things are annoying in the merge for someone who wishes to use
> >> udev alone, and not systemd. That is a fact as well.
> > 
> > There is tons of stuff that you would need to fix, not just udev
> > stuff. See [1].
> 
> The point is, systemd and udev have recently been patched by upstream so
> that things are going *even more* on the direction of having stuff
> stored in /usr.

Which is still not really a problem when tons of other daemons have
done the same already. We're not "fixing" this by patching systemd and
udev when many other daemons behave the same way. So, it's pointless
anyway.

Besides, can you elaborate what is so important in having /usr
separate? I see that it made sense back on the old Unix workstations
where you could split partitions across different disks, but I don't
see the point nowadays where a cheap harddisk has 1TB of space.

> That is exactly this kind of things (and others, like merging systemd
> and udev), going to the wrong direction, which pushed Gentoo people
> to fork.

Which I don't agree. Just look at FreeBSD, they have the whole
operating system core merged with their kernel source. I don't see
anyone complaining there either.

> >> On 11/25/2012 12:15 AM, John Paul Adrian Glaubitz wrote:
> >>> If both Ubuntu and Gentoo would just go with the rest of the community
> >>> and accept systemd
> >> Yeah, right! Along the line with "why are these idiots so vocal". That is
> >> in fact one of the main concern about udev/systemd people: they don't
> >> care about others, refuse patches, and always claim others are stupid.
> >> You shouldn't go through this dangerous path as well, IMO.
> > 
> > Sorry, but I wouldn't touch code with a ten-feet pole who from someone
> > is so naive claiming that he knows more about writing an open source
> > BIOS than the people at Coreboot who have been doing that since
> > 1999. I started right out laughing when he claiming Coreboot is a
> > Linux distribution and starting a x86 computer just involves setting
> > up a few interrupt vectors.
> > 
> > I know several people in the Coreboot project personal and the
> > claims that Richard makes in [2] are just ridiculous and show the same
> > attitude he has towards working on udev.
> > 
> > Even Greg is ridiculing their work [3]. And he should know, he came up
> > with udev in the first place and he is even a Gentoo developer. Don't
> > you think he understands you a little more than anyone else, so he can
> > assess when people do the right things with the code or not?
> 
> This is the kind of aggressive attitude which happened with systemd
> upstream, claiming that everyone are idiots, while in fact they've been
> pointing rightly at problems. You just made a new occurrence of it. :(

And you are dodging my questions. Don't you think that people like
Richard who make such bold claims that writing a free BIOS replacement
is a matter of a few hundred lines of assembly, ridiculing the almost
15 years of development of Coreboot, cannot be taken seriously?

Do you think that Greg - being a Gentoo developer himself - would make
such statements if he didn't knew what he was talking about?

Sorry, but no. I have seen such claims in the past very often where
people ridiculed existing software stacks and claiming they could do
something better in no time [1]. This guy would even remain on his
bold point of view after one of the Ardour and Jack developers
explained him why he was wrong [2].

I didn't mean to be aggressive in any way, I'm just annoyed. Please
don't get me wrong about forks, they are just fine. As I said before,
I like MATE very much, for example, because these people do great work
and they even collaborate with the original authors from GNOME.

But I don't support forks which are created based on false assumptions
of the original software. Another bad example where people forked and
obviously don't know what they're doing is Trinity [3].

Adrian

> [1] http://klang.eudyptula.org/
> [2] http://ardour.org/pd_on_klang
> [3] http://blog.martin-graesslin.com/blog/2012/10/maintaining-history-done-wrong/

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer - glaubitz@debian.org
`. `'   Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913