00:33:08 <jbk> i remember having to do something like that with hp-ux 10.20
00:33:19 <jbk> where it relinked the kernel or such
00:33:38 <jbk> into the 00s
01:04:07 <sommerfeld> richlowe: I have this vague notion that it generated one or maybe a couple *.c files that defined global variables containing config parameters -- the sort of things we set in /etc/system today.  
02:47:14 <jbk> i found it amusing that they made such a big deal about when all of that went away
02:47:34 <jbk> and when you could set the sysv limits and just reboot
02:48:36 <jbk> in 11.x
02:48:45 <jbk> in like 2002
14:17:15 <gitomat> [illumos-gate] 18197 RFC1918 addresses get site-local scope defeating RFC 6724 destination selection -- Andy Fiddaman <illumos⊙fn>
14:22:20 <gitomat> [illumos-gate] 18212 mdb: nv_var_alloc() error: strcpy() 'name' too large for 'v->v_lname' -- Toomas Soome <tsoome⊙mc>
14:27:58 <gitomat> [illumos-gate] 17173 mdb: check sign expansion -- Toomas Soome <tsoome⊙mc>
16:27:16 <alanc> yeah, as far as I remember, all it really compiled was a list of global variables and their values into a .o to be linked with the rest, before you could set values via /etc/system
16:42:04 <jbk> yeah that sounds a lot like what HP-UX did (at least through 10.20 until they finally moderinized a bit)
16:42:22 <jbk> though they also unhelpfully didn't document the valid values you could set
16:46:17 <jbk> or at least well enough.. I think it was two sysv semaphores whose product had to be < 16386 (i.e. almost certainly fit in a int16_t)
16:46:28 <jbk> err two sysv ipc semphore parameters
16:46:39 <jbk> and if they didn't... you just crashed