[procps] Re: spindle clearing
- From: Jim Warner <james.warner@xxxxxxxxxxx>
- To: Procps-ng Mail list <procps@xxxxxxxxxxxxx>
- Date: Thu, 8 Dec 2016 14:04:41 -0600
On 12/07/2016 05:23 AM, Craig Small wrote:
Yes, you could say they are coming along very very slowly.
I also adjusted my git hooks :)
Hi Craig,
I'm not sure that hook will catch trailing whitespace from a merge, but
it definitely stops local commits. If ever you wish to override that
pre-commit hook, man githooks explains how.
As for documentation, if I was in charge here's what I'd try to sell you on.
--------------------------------------------------
Unlike the approach in doc/libproc.3, I would not document what most of
our enumerators represent. Instead, I would refer users to kernel
documentation under Documentation/filesystems/proc.txt. That way we're
blameless for existing/future gaps plus we dodge a huge maintenance
nightmare.
However, any enumerators representing our own unique derived values
should be shown with a brief explanation.
--------------------------------------------------
I would strive for just two documents.
One would deal with the concept of stacks (variable length records) and
the basic interface mechanics. Additionally, any specifics for a
particular API could be covered like those unique derived items.
The other would be devoted to the <pids> API since its 'new' and
'select' operate a little differently and there are the extra functions
'reset' & 'fatal_proc_unmounted'. I see no harm in duplicating much of
the first document in this one too. And it's here that all enumerators
and their source directory would be mentioned along with any unique
derived items.
Regards,
Jim
Other related posts: