@schratze Linux has one big benefit. It supports annoyance driven development (ADD). Basically when you get annoyed enough, you can actually fix things instead of just continuing to be annoyed. Now that might be annoying on its own, but it at least keeps you busy instead of fuming in a corner!
(Sorry for the necro-reply, someone boosted this and I blame them!)
@PCOWandre@schratze Actually, working at WeMoveMobileData and later WeHostMillions I learned to like Solaris 10 a lot. Which was rather unlike the SysV I got to know at the University earlier in its different forms of not being very satisfactory.
@jyrgenn@schratze I think years of earlier Solaris conditioned me a little. When I had a retro trip to Dell UNIX et al, the userland was familiar and I had no trouble building packages.
@jyrgenn@schratze Circa 1990-1992 SVR4 UNIX. Was near unobtainium at the time of release; certainly never found it in Australia.
If you look at some early comparisons of UNIX options on x86, it'll turn up, often with a pretty strong recommendation.
Caveats? The hardware support is pretty narrow. Only a couple of supported NICs; I never managed to get the Adaptec SCSI HBA support working and it doesn't know about large storage.
My plan was to attempt to port some of the NIC drivers from early Solaris x86 but my development box kept shitting itself -- failed power supply then failed motherboard then failed disk then failed NIC. At some point I'll find another reasonable fast-486/early pentium and give it another poke.
@PCOWandre@jyrgenn@schratze Porting drivers from early Solaris to SVR4? Would that even be possible? Wasn't early Solaris basically SunOS (BSD) under the hood?
@resuna@jyrgenn Nope. Solaris -- as in Solaris proper, not the "Solaris 1" backported name for SunOS 4 -- was STREAMS and System V all the way.
With proper API/ABI definitions and compatibility. Look at some of the classic LSI drivers -- one binary covers 2.6 through 10 and the cut at 2.6 is only because of the jump to 64 bit splitting the driver tree.
Add comment