Re: Devuan
Rick Moen <rick-IyCrq+X4Fdq2oZ/[email protected]>
| Newsgroups | gmane.org.user-groups.linux.svlug |
|---|---|
| Organization | If you lived here, you'd be $HOME already. |
| Message-ID | <[email protected]> |
Quoting Ivan Sergio Borgonovo ([email protected]): > If there is any chance that the effort put into making Debian support > more than one init as smoothly as possible the prerequisite is > developers start to respect each other work. There are some subtleties, here. First, anyone who's been anywhere near IT over the last fifty years and speaks the English language knows that the word 'support' is a political football. Inside the institutional politics of the Debian Project in recent context it was a squabble over efforts to bludgeon developers with any of several alternative sets of proposed bureaucratic decrees that would (variously) either require integration for particular init systems or not require it. (I will elaborate.) Debian _does_ package several init systems, they all work fine, and there's no rational reason to suppose they won't continue working fine. If occasionally a system daemon package ships without glue for some init system that someone's in love with, that's about the easiest system glitch to solve that I can think of. That's not a senior sysadmin problem, and I'm not sure it requires a PFY, either: I think it could be offloaded to the NOC tapemonkey. It's like: Oh, look on the Internet, get a likely-looking init-script fragment, test it, add it to configuration management and publish it in a public repo pro bono publico. The political fighting within the Project, most particularly Ian Jackson's much-resented General Resolution, concerned efforts to decree from the top that everyone march to one tune or another. The majority of Debian developers looked at the proposed four alternatives, and opted for the one that equated to 'No, go away, Ian, we're busy getting actual work done.' IMO, completely wrong battle, and they made an omnishambles out of it. IMO, the lot of them should have said 'Well, GNOME is certainly setting off a lot of time-wasting sturm und drang by decreeing that they cannot do without "multiseat" API support and finding that ConsoleKit being EOLed meant that they "had to" start requiring systemd-logind, which means everyone on GNOME with "multiseat" would need substantively the entire systemd suite and flamewars then followed tout suite. I know, here's an idea: Let's dump GNOME.' IMO, that would have been the most parsimonious and intelligent solution to the problem GNOME dumped in their laps -- but the political mechanism once set in motion by Paul Tagliamonte's tech-ctte bug and Ian Jackson's General Resolution had a momentum of their own, and nobody had the time and leisure to stop and think 'Wait, this is not even the right issue.' So, a lot of people got angry over a dumb discussion, people resigned from various positions en-masse, and pretty much the only intelligent action that occurred was the voters telling Ian Jackson 'No, go away.' Disclaimer: I have no connection to the Debian Project; I just use some of the software. Above is just how things look from here.