[procps] Re: rhbz#737215 - ps does not resolve some user names
- From: Craig Small <csmall-procps@xxxxxxxxxx>
- To: procps@xxxxxxxxxxxxx
- Date: Fri, 14 Sep 2012 09:00:06 +1000
On Wed, Sep 12, 2012 at 05:56:06AM -0400, Jaromir Capik wrote:
> I recently found a forgotten bug in my list, in scope of which
> one of the ps features was discussed and we found it's usefulness
> a bit controversial.
It is the intended behavour of ps and has been that for many years. In
fact I think it has always been that. You can do the magic column width
trick and that shows the UID if you like.
The problem is you will get confusion either way. A good example is exim
with its multiple users; different users will appear as the same users.
Expanding the username column only on those lines would be terrible
idea, it would look bad and be confusing about what column is what.
UID is U32 which is 10 digits long, so there is probably a problem
already.
> Do you find reasonable to modify the code to show just a part
> of the username (followed by star or something) if the username
> doesn't fit? This "auto-magic" apparently leads to cunfusions.
I was really debating about what is the right way to go with this. I
can understand the original concept was to make sure it was always
unique; actually i think the usernames could only be 8 characters
back the the old days.
But we know now that even with the username to uid trick, we still get
caught because UIDs are up to 10 digits long. The switch does confuse
people. I am also of the opinion that anyone trying use the username
for matching is asking for trouble anyhow because it might be the UID,
use the UID!!
I was initially against Jaramir's idea but after thinking through it I
cannot see why the old way is better; matches fail and I get a slow
stream of confused people when they see it. I'm for the change.
We should fix the UID,GID,etc column width while we are at it.
- Craig
--
Craig Small VK2XLZ http://enc.com.au/ csmall at : enc.com.au
Debian GNU/Linux http://www.debian.org/ csmall at : debian.org
GPG fingerprint: 5D2F B320 B825 D939 04D2 0519 3938 F96B DF50 FEA5
Other related posts:
