Mailing List Archive
tlug.jp Mailing List tlug archive tlug Mailing List Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]Re: [tlug] Kana-Problems
- Date: Tue, 22 Apr 2008 07:06:12 +0900
- From: "Josh Glover" <jmglov@example.com>
- Subject: Re: [tlug] Kana-Problems
- References: <20080417080425.9360bdd2.n.kobschaetzki@googlemail.com> <87y779ctob.fsf@uwakimon.sk.tsukuba.ac.jp> <78d7dd350804201936p582d55d2qaf3b3e9d578c0333@mail.gmail.com> <bf4e1fa10804202205n6a6b1befoc5e437b918d7325c@mail.gmail.com>
On 21/04/2008, Niels Kobschaetzki <n.kobschaetzki@example.com> wrote: > I like my desktop-system up to date and I have not the time to follow > the changelogs of the 1000+ ports I have installed right now to check > if it is necessary to install updates. If you don't have time to follow changelogs, then you almost certainly don't have time to fix a system when a blind update hoses it, right? So stop blindly updating. > I do not really expect an update of ports to break my system. Then adjust your expectations. Remember, you are talking about keeping--in your words--1000+ ports up to date. Would you expect to be able to keep 1000 pieces of software up to date on a MacOS X or Windows system, with no update ever breaking anything? Modern desktop distros (and I include the BSDs here) are complex beasts, with many intertwining tentacles. I'd really advocate that you update only security-related stuff automatically, which you should be able to do easily by either simply trusting the FreeBSD security patches implicitly or by following the proper mailing lists (I think Scott gave you info on this). Other updates should only happen when you *know* the update fixes a bug or provides a feature that you really want. To use myself as an example: as a rule, I update nothing, save kernel security fixes. I don't run any web-facing services on my personal machines, so I don't care if Apache has a crippling security bug--unless an attacker is able to access my loopback interface, I'm safe from it. Only the kernel can bite me: if there is some clever way to exploit the TCP window size scaling code to inject packets into the buffer for say, the loopback interface, I may be screwed. I update software when I need something, e.g. yesterday I updated Amarok because the album cover art fetching feature had broken, and the Amarok site's RSS feed mentioned that bug specifically in the announcement of a new release. And I update software when I want something, e.g. I updated to Firefox 3 beta because I want the vastly improved resource utilisation it offers. Other than that, I don't update. And you know what? My machine keeps working the way I want it to. -- Cheers, Josh
- Follow-Ups:
- Re: [tlug] Kana-Problems
- From: Joe Larabell
- Re: [tlug] Kana-Problems
- From: Scott Robbins
- Re: [tlug] Kana-Problems
- From: Niels Kobschaetzki
- References:
- [tlug] Kana-Problems
- From: Niels Kobschaetzki
- [tlug] Kana-Problems
- From: Stephen J. Turnbull
- Re: [tlug] Kana-Problems
- From: Nguyen Vu Hung
- Re: [tlug] Kana-Problems
- From: Niels Kobschaetzki
Home | Main Index | Thread Index
- Prev by Date: Re: [tlug] Linux distro!: Centos is RHEL
- Next by Date: Re: [tlug] Kana-Problems
- Previous by thread: Re: [tlug] Kana-Problems
- Next by thread: Re: [tlug] Kana-Problems
- Index(es):
Home Page Mailing List Linux and Japan TLUG Members Links