[OpenAFS-devel] openafs warnings

Derrick Brashear shadow@gmail.com
Mon, 6 Jul 2009 13:22:11 -0400


On Mon, Jul 6, 2009 at 1:19 PM, Marcus Watts<mdw@umich.edu> wrote:
> At the last workshop, Simon identified "reducing warnings" as a desirable
> goal of new patches for openafs. =A0This is an interesting problem for
> large patches, because the # of warnings in openafs is large enough that
> it's very difficult to sort out were inherited, and which ones are new.
>
> Also, it looks bad when the mainline code is not up to the standard
> expected of new code.

What, so, targets aren't desirable?

>
> Without any more ado, here are some interesting statistics:
>
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0grep -c =A0 =A0 =A0 =A0 un=
ique source
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0warning: =A0 =A0 =A0 =A0wa=
rning count
> openafs 1.5.60(*) =A0 =A0 =A0 2900 =A0 =A0 =A0 =A0 =A0 =A01955
> hk28 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A05101 =A0 =A0 =A0 =A0 =A0 =A03=
053
> hk32 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0419 =A0 =A0 =A0 =A0 =A0 =A0 2=
16
>
> stock openafs 1.5.60 needs 2 patches to build in my env.
> hk28 and hk32 are "rxk5 against cvs head".
> Some warnings occur more than once as various source files
> are built different ways.
>
> I expect at some point soonish to have a somewhat simplier
> patch that just does this warnings cleanup.

Good. Much easier to review. Ideally, better yet, one patch per topic
or per subsystem but we'll take what we can get.

Also, remember some of it will no doubt be reverted when we discover
it breaks (pick a sad platform).