Gary Winiger wrote:
>> I'm kind of concerned here -- if there is a requirement for software to
>> do something else, then this needs to be fully documented -- including
>> in the man pages.
>>
>
> I've tried to get the folks, if there are any, who own uadmin(2)
> to update the man page to actually describe what it does. The
> man page CR is not getting any attention. Beyond the power management
> stuff, which the Audit team is working with, there doesn't
> appear to be anyone who cares.
>
>
>> I don't believe it is reasonable to expect 3rd party callers of
>> documented system calls to know about undocumented requirements.
>>
>
> Do you have some suggestions? I could file a P2 man page
> bug saying this man page is horribly out of date and I don't
> know how to fix it. Don't use uadmin(1) only call uadmin(1),
> but that seems equally as broken as the uadmin(2) documentation.
> I believe the power management team has a medimum term solution
> to do all the right stuff and maybe even interpose in libc
> with stuff that does all the right stuff.
> sys-suspend(1) proposed here is to help OpenSolaris until
> that solution is in place.
>
There are other programs that call uadmin. I've written one myself
while I was working at a customer site ages ago.
We need either in-kernel or libc interposer support... or significantly
improved documentation (and possibly a reclassification of uadmin(2)'s
stability level if we think folks shouldn't be calling it directly.)
A P2 man page bug is just a start.
-- Garrett
> Gary..
>
