[ Moving this to observability-discuss. ] On 10/7/07, Mike Gerdts <mgerdts at gmail.com> wrote: > On 10/6/07, Chad Mynhier <cmynhier at gmail.com> wrote: > > If anyone's curious about the details of my proposed solution, I've > > written up a fast-track which can be seen here: > > http://interstel.net/~mynhier/6518130.txt. > > Very interesting extension to the type of arguments normally given to > p* commands. It looks more like the approach taken by fuser than > pfiles. Or another way, it looks like this might be an extension to > pgrep. One thing missing in the proposed output is the the ability to > see the related IP address(es).
Yes, this command wouldn't really be one of the ptools, as it takes a port as an argument and not a pid. (And given that, I don't know that pport is the best name for the command, but it was a convenient temporary name.) It certainly fits more into the fuser/pgrep approach. I'd argue that fuser isn't the place for this functionality, as we're specifying a port and not a file. (Hmm, maybe "portuser" should be the name?) But I'd also argue that pgrep isn't the place for this, either. Pgrep matches attributes of processes, and even though "listening on port n" could be considered an attribute of a process, it doesn't really fit the other attributes that pgrep matches against. WRT seeing the IP addresses, that could be be part of the verbose output. Chad
