Re: [PATCH 8/6] signals: send_signal: factor out SIGNAL_GROUP_EXIT checks

[Roland McGrath]

> Of course, since I personally don't agree, I'd like to discuss it more
> if possible.

Certainly, but let's do it under separate cover, and after some of
these cleanups settle.  (I'd rather not try to get into it this week.)

> "[PATCH 3/6] signals: use __group_complete_signal() for the specific signals too"
> adds a behaviour change too.

Your log entry was explicit about the semantics change there.  You
explained it up front and justified it.  I also happened to agree
with it, but that's a separate issue.

I am certainly not opposed to semantics changes a priori.
I happen to have reservations about this particular one.  

> Surely, I don't want to change the behaviour quietly, that is why I am
> cc'ing maintainers.

The point is that, whenever possible, a semantics change should be
isolated into a patch separate from any related cleanups.  More
important than that, no semantics change should go unmentioned so
it's only documented as a result of someone's careful review of
the change.  (Of course when a change is inadvertent, then only
review is going to notice it--that's what review is for.)

Your new pair of patches dated 2008-3-12 look like they are doing
exactly this (just the cleanup first).  After those are in, the
semantics change you want is a one-liner and easy to review and
discuss on its own.

> (note also these patches are 8/6, 9/6. Originally I was going to send
>  them in a separate series).

I had noticed that your wholes sometimes go up to 1.5; I just
figured it's because you're 50% more thorough than the rest of us.
;-)


Thanks,
Roland
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/