Hi,
Upstream of the procps-ng adopted removing -v from pkill, while
keeping it for pgrep. The change allows one to reach the inversion
logic with long option e.g. 'pkill --invert' but that is not
documented (while it perhaps should be).
https://gitorious.org/procps/procps/commit/1af18c260a87dc38f
Hi Jason,
On Thu, Feb 04, 2010 at 04:01:33AM -0500, Jason A. Spiro wrote:
> Thanks for writing NetBSD pgrep/pkill. When you have some spare time
> to carefully read ten messages and think, you please read through
> http://bugs.debian.org/558044 and suggest what you think would be a
> good algori
Hi Andrew,
Thanks for writing NetBSD pgrep/pkill. When you have some spare time
to carefully read ten messages and think, you please read through
http://bugs.debian.org/558044 and suggest what you think would be a
good algorithm for making -v safer?
Thanks in advance,
-Jason
--
To UNSUBSCRIB
as the original author of Linux' pgrep, I support the idea of making the
use of -v an error unless a search criterium other than pattern is
specified. (-f is not a search criterium in itself.)
there is a sort of precedent for this, since -o and -n has always been
disallowed with -v since the beha
reopen 558044 =
thankyou
The current way of using pgrep -v can often lead to a large number of
processes killed and often this is not the intention of the user. A way
of limiting the damage needs to be added to pkill without changing the
flag.
Some ideas are to disable -v if no process (before t
cc added: Albert Cahalan . Albert, welcome
to the thread, and thanks for maintaining procps.
On Wed, Dec 2, 2009 at 5:14 AM, Craig Small wrote:
>
> On Wed, Dec 02, 2009 at 04:09:30AM -0500, Jason A. Spiro wrote:
>>
>> When? What were the complaints? :)
>>
> It was quite a few a years ago and
-- Forwarded message --
From: Craig Small
Date: Wed, Dec 2, 2009 at 5:14 AM
Subject: Re: Bug#558044: pkill, pgrep: don't use "-v" for negation;
it's dangerous & some think -v means verbose
To: "Jason A. Spiro"
On Wed, Dec 02, 2009 at 04:0
On Wed, Dec 2, 2009 at 3:06 AM, Craig Small wrote:
> That's been done before, there were lots of complaints.
When? What were the complaints? :)
> They copied the FAQ from upstream, I saw nothing about it being a Debian
> specific patched version.
But Craig, aren't you the upstream maintainer
-- Forwarded message --
From: Craig Small
Date: Wed, Dec 2, 2009 at 3:06 AM
Subject: Re: Bug#558044:
To: "Jason A. Spiro"
On Wed, Dec 02, 2009 at 02:59:38AM -0500, Jason A. Spiro wrote:
> Maybe you could gradually transition it in, by making both options
> wo
On Thu, Nov 26, 2009 at 6:41 AM, Craig Small wrote:
> I disagree. pgrep/pkill (they're the same binary) are like grep and use
> the same flags. I think that is important.
I think it is more important to protect users. But you are the
maintainer and so the decision is up to you.
Now I will cont
I see from
http://www.cyberciti.biz/tips/my-10-unix-command-line-mistakes.html#comment-150923
that pkill's options are derived from pgrep's which are derived from
grep's. But still, the benefits of preserving grep-like -v behavior
don't outweigh the costs.
The benefits of preserving Solaris compa
Package: procps
Version: 1:3.2.8-2
Severity: wishlist
X-Debbugs-CC: mo...@mozai.com;volker.maib...@eberspaecher.com
Thank you for maintaining procps.
A suggestion: "killall -v Foo" means "kill process Foo, and print
verbose output". Therefore many users think that "pkill -v Foo" means
the same
12 matches
Mail list logo
