1 |
On Mon, 17 Feb 2014 21:52:55 +0400 |
2 |
Andrew Savchenko <bircoph@×××××.com> wrote: |
3 |
|
4 |
> On Sun, 16 Feb 2014 15:16:36 -0600 Canek Peláez Valdés wrote: |
5 |
> > On Sun, Feb 16, 2014 at 2:58 PM, Volker Armin Hemmann |
6 |
> > <volkerarmin@××××××××××.com> wrote: |
7 |
> > > Am 16.02.2014 21:08, schrieb Canek Peláez Valdés: |
8 |
> > >> On Sun, Feb 16, 2014 at 12:59 PM, Volker Armin Hemmann |
9 |
> > >> <volkerarmin@××××××××××.com> wrote: |
10 |
> > >> [ snip ] |
11 |
> > >>> or it is an idiotic decision. Because features means complexity. |
12 |
> > >> Yeah, like the kernel. |
13 |
> > >> |
14 |
> > >>> Complexity means bugs. |
15 |
> > >> Bugs get reported, bugs get fixes. Life goes on. |
16 |
> > |
17 |
> > You didn't answered this, did you? |
18 |
> |
19 |
> Bugs are different. Bugs in the critical system components are |
20 |
> critical to the whole system. If Libreoffice or browser |
21 |
> segfaults, some data may be lost and inconvenience created, but the |
22 |
> system will continue to run. If PID 1 segfaults — everything is |
23 |
> lost, you have a kernel panic. That's why critical components should |
24 |
> be as simple and clean as possible. |
25 |
|
26 |
If it does, but does it? We have run it for ages without a segfault. |
27 |
|
28 |
> SysVinit code size is about 10 000 lines of code, OpenRC contains |
29 |
> about 13 000 lines, systemd — about 200 000 lines. |
30 |
|
31 |
That is an unfair comparison, be fair and consider PID 1's code size. |
32 |
|
33 |
> Even assuming systemd code is as mature as sysvinit or openrc (though |
34 |
> I doubt this) you can calculate probabilities of segfaults yourself |
35 |
> easily. |
36 |
|
37 |
Practical statistics are more reliable than theoretical probabilities. |
38 |
|
39 |
> > >> All of them are different tools providing one capability to |
40 |
> > >> systemd as a whole. So systemd is a collection of tools, where |
41 |
> > >> each one does one thing, and it does it well. |
42 |
> > >> |
43 |
> > >> By your definition, systemd perfectly follows "the unix way". |
44 |
> > >> |
45 |
> > > |
46 |
> > > no, it isn't. |
47 |
> > > |
48 |
> > > How are those binaries talk to each other? |
49 |
> > |
50 |
> > dbus, which is about to be integrated into the kernel with kdbus. |
51 |
> |
52 |
> And this is a very, very bad idea. Looks like you don't know matter at |
53 |
> all: to begin with kdbus protocol is NOT compatible dbus and special |
54 |
> converter daemon will be needed to enable dbus to talk to kdbus. |
55 |
|
56 |
That claims it to be a bad idea, but doesn't tell why; furthermore, no |
57 |
technical reasoning as to why it is incompatible is given. Do you know? |
58 |
|
59 |
> The whole kdbus technology is very questionable itself (and was |
60 |
> forcefully pushed by RH devs), anyway it is possible to disable this |
61 |
> stuff in kernel and guess what will be done on my systems. |
62 |
|
63 |
Similar claims again, without any weight; that is subjective opinion. |
64 |
|
65 |
> > > Looks broken. Broken by design. The worst form of broken. |
66 |
> > |
67 |
> > By your opinion, not others. |
68 |
> |
69 |
> That is not just an opinion. |
70 |
|
71 |
It is due to the lack of science and experience in your response. |
72 |
|
73 |
> And all that science was ignored during systemd architecture process |
74 |
> if there was any at all. |
75 |
|
76 |
For it to be claimed as "ignored", you need to know about the process; |
77 |
given that you don't even know its presence, such claim can't be made. |
78 |
|
79 |
-- |
80 |
With kind regards, |
81 |
|
82 |
Tom Wijsman (TomWij) |
83 |
Gentoo Developer |
84 |
|
85 |
E-mail address : TomWij@g.o |
86 |
GPG Public Key : 6D34E57D |
87 |
GPG Fingerprint : C165 AF18 AB4C 400B C3D2 ABF0 95B2 1FCD 6D34 E57D |