Re: [arch-general] [arch-dev-public] [signoff] vc/* -> tty* transition
Holy hell this is out of control. Here's the two sides, boiled down:
* Use an automatic sed to prevent people from complaining
* Post a news item and let people do it manually.
As we can tell from this thread, people are going to bitch either way -
making the 'no bitching' argument a little moot.
Personally, I am against this but I can see it simplifying this process, so
I understand why people would want this. My biggest fear is that it goes
downhill from here. 'We did it for the tty change' will be used to justify
more and more. It's a slippery slope. Sometimes these things happen where
you need to be slightly inconvienanced in order to protect yourself from
falling into a trap like this.
However, I must point out: odds are most people don't touch inittab, so the
upgrade will do things as expected and the sed line will only do work a
small subset of end users.
And to be clear, I definitely do not like the pandering to users thing... if
people whining about stupid shit gets on your nerves, stop visiting the
forums and IRC. It worked for me! ( google 'eternal september' for kicks :).
Pyther, I like your sentiment.
On Jul 18, 2009 10:14 PM, "Daenyth Blank"
Aaron Griffin schrieb:
However, I must point out: odds are most people don't touch inittab, so the upgrade will do things as expected and the sed line will only do work a small subset of end users.
You are wrong here. I would guess virtually any user touched it.
And to be clear, I definitely do not like the pandering to users thing... if people whining about stupid shit gets on your nerves, stop visiting the forums and IRC. It worked for me! ( google 'eternal september' for kicks :).
Yeah, we are proud of our great community which we like to ignore.
On Sun, Jul 19, 2009 at 13:16, Thomas Bächler
Aaron Griffin schrieb:
However, I must point out: odds are most people don't touch inittab, so the upgrade will do things as expected and the sed line will only do work a small subset of end users.
You are wrong here. I would guess virtually any user touched it.
That's questionable. It depends if users configured their X login manager in inittab or just added gdm/kdm/slim/whatever to DAEMONS in rc.conf (as I did). I doubt there is any statistics on it, so it's hard to correctly assume anything. Anyway these are valid points:
That said, we do modify configuration files all the time. We run grpck on a shadow update so users can still log in, some gtk update generate files in /etc so it still finds its plugins and more. We just don't do it ourselves, but hide behind some program provided to us and tell ourselves "It's okay, upstream wanted it this way". And guess what, nobody even notices.
The whole discussion is getting on a way to flamewar IMO. I'm fine with just newsitem in advance and a post_upgrade message, but Thomas' idea about doing sed and saving user's config as .pacsave and posting a message about what was done is reasonable as well: * users who weren't careful will have a working system after reboot, * users who are careful will see the .pacsave and will check\ if sed didn't break their config. -- Roman Kyrylych (Роман Кирилич)
Le Sat, 18 Jul 2009 23:29:28 -0500,
Aaron Griffin
And to be clear, I definitely do not like the pandering to users thing... if people whining about stupid shit gets on your nerves, stop visiting the forums and IRC. It worked for me! ( google 'eternal september' for kicks :). Pyther, I like your sentiment.
Same here. I'm just a user but I like to think that Arch is a distribution that gives control to power users, and that teaches other users how things work. That teaching might require breaking the system of those that don't follow simple rules such as read the output of Pacman. Moreover, I have modified /etc/inittab, and depending on what the sed does, it might break my system. I don't think I'm the only user to have done that. So, even if you go the sed way, you will need to use post_upgrade() to warn the users that you changed something, and probably create a .pacsave... But in fact, if your goal is to make the systems of people who don't know what /etc/inittab is work, why use sed and not just replace the file with a new one using tty? Anyway, I second pyther, phrakture and all those who are against automatic changes to critical configuration files. And I think every post or bug report complaining about that should be closed with a link to the news post about the move from vc to tty. -- catwell
Pierre Chapuis schrieb:
That teaching might require breaking the system of those that don't follow simple rules such as read the output of Pacman.
How can a user distinguish between important pacman output and the crap that is put everywherre?
Moreover, I have modified /etc/inittab, and depending on what the sed does, it might break my system. I don't think I'm the only user to have done that. So, even if you go the sed way, you will need to use post_upgrade() to warn the users that you changed something, and probably create a .pacsave... But in fact, if your goal is to make the systems of people who don't know what /etc/inittab is work, why use sed and not just replace the file with a new one using tty?
There was the time when pacman saved the old file as .pacsave and put the new one in place, effectively restoring the default configuration and leaving it to the user to merge in his custom configuration. I am beginning to understand why someone would do it this way. Nowadays, as soon as you modify the file, the new files are installed as .pacnew which makes sense most of the time. But in this case ... let's admit it, most users don't really care about inittab. Some HOWTO told them to uncomment a line for their login manager and they forgot about it. If you really want to educate users, remove all those HOWTOs that people follow step-by-step without understanding a word and let them figure it out themselves.
i don't get the point of this discussion! isn't it the devs choice to do it how _he_ thinks it's good and reasonable? if thomas thinks putting a sed line in the .install is the best way, then he should do it! On Sun, Jul 19, 2009 at 12:57:30PM +0200, Thomas Bächler wrote:
Pierre Chapuis schrieb:
That teaching might require breaking the system of those that don't follow simple rules such as read the output of Pacman.
How can a user distinguish between important pacman output and the crap that is put everywherre?
Moreover, I have modified /etc/inittab, and depending on what the sed does, it might break my system. I don't think I'm the only user to have done that. So, even if you go the sed way, you will need to use post_upgrade() to warn the users that you changed something, and probably create a .pacsave... But in fact, if your goal is to make the systems of people who don't know what /etc/inittab is work, why use sed and not just replace the file with a new one using tty?
There was the time when pacman saved the old file as .pacsave and put the new one in place, effectively restoring the default configuration and leaving it to the user to merge in his custom configuration. I am beginning to understand why someone would do it this way.
yes, creating a pacsave file is an elegant solution. vlad --
participants (5)
-
Aaron Griffin
-
Pierre Chapuis
-
Roman Kyrylych
-
Thomas Bächler
-
vlad