Gentoo Archives: gentoo-dev

From: Rich Freeman <rich0@g.o>
To: gentoo-dev@l.g.o
Cc: Duncan <1i5t5.duncan@×××.net>
Subject: Re: [gentoo-dev] Re: openrc portage news item
Date: Thu, 14 Apr 2011 11:45:04
Message-Id: BANLkTin-BHy1B6ryF7VgPauX00aAeNREUw@mail.gmail.com
In Reply to: [gentoo-dev] Re: openrc portage news item by Duncan <1i5t5.duncan@cox.net>
1 On Thu, Apr 14, 2011 at 4:09 AM, Duncan <1i5t5.duncan@×××.net> wrote:
2 > FWIW/IMHO, I don't believe the news item needs mentioning that it was bash
3 > that made it slow and inflexible.  Most users don't so much care whether
4 > it's C or bash or java that made it so, only that it was.
5
6 If this were Ubuntu I'd be inclined to agree. However, I think that
7 most Gentoo users would be interested. Maybe I have a different
8 perspective because I just gave a talk on booting two nights ago at an
9 LUG, but I wasn't even the one to bring up the shortcomings of bash in
10 the typical linux SysVInit-based service scripts. Various approaches
11 that were discussed included symlinking /bin/sh to dash instead of
12 bash, and C-based solutions (or a combination of both). It was
13 interesting to hear that at least a few other distros struggle with
14 bashisms in their init scripts, but no so much due to licensing/BSD
15 issues but because of a desire to use dash which does not support all
16 bashisms.
17
18 No need to go into gory details, but mentioning that it is C-based
19 instead of bash-based seems reasonable. Granted, we're not really
20 getting rid of one of the problems with bash, which isn't just
21 /sbin/rc but rather it includes the init scripts themselves (every one
22 of which requires spawning a new bash, and many spawn additional
23 processes like sed/awk/etc).
24
25 Rich