3 ms·
There have been several failed attempts at launchd. There was also a NeXTBSD. * http://jdebp.eu./FGA/launchd-on-bsd.html http://jdebp.eu./FGA/launchd-on-bsd.h
by JdeBP 8y ago
There have been several failed attempts at launchd. There was also a NeXTBSD.
* http://jdebp.eu./FGA/launchd-on-bsd.html http://jdebp.eu./FGA/launchd-on-bsd.html
* https://news.ycombinator.com/item?id=13899766 https://news.ycombinator.com/item?id=13899766
* https://news.ycombinator.com/item?id=14983970 https://news.ycombinator.com/item?id=14983970
* https://news.ycombinator.com/item?id=15161361 https://news.ycombinator.com/item?id=15161361
I mention nosh there. I am going to release version 1.38 soon.
- peatmoss 8y agoWow, looks like others have had the same idea. Wonder what stalled things out? Maybe there’d be something worth picking back up.
- loeg 8y agoThe big sticking point the last go-around was adding Mach IPC to FreeBSD. Mach IPC affects the security of the system in potentially subtle ways (it seems Apple struggled with Mach IPC security for a while in the early days of OS X). In addition to security concerns, there was a question of whether FreeBSD should have Mach IPC, as a sort of project identity question. Some suggested using plain Unix sockets instead of Mach IPC, but the effort required for that seemed to be very non-trivial. There was some concern about XML plists, which no one is excited about, but I view those as somewhat trivial. (FreeBSD launchd could easily use a JSON or "libucl" format plist.) Finally, the developer behind the Mach IPC and launchd port didn't really want to do the politics required to land a big change like this. I can't really fault him that; it's tedious and thankless.