3 ms·
> To sysadmins, a program's output or its lack thereof is a vital part of its behavior Isn't this why there are two output streams, stdout and stderr?
by leereeves 3y ago
> To sysadmins, a program's output or its lack thereof is a vital part of its behavior
Isn't this why there are two output streams, stdout and stderr?
- NoZebra120vClip 3y agoYes, I'm scratching my head about this and wondering how the output method never entered into the discussion of this post. Now, the context is that "fgrep" and "egrep" have been deprecated by GNU, so that eliminates syslogs or GUI messages from the calculus at hand, but it seems like warnings on stderr are pretty normal. Now the further context is that fgrep and egrep have been embedded into larger scripts and systems which call them in turn, so now you've got the stderr of a tool mingling with the stderr of every other tool, and the script itself. But that's why everything on stderr includes argv[0] from time immemorial, too. The comment about rate-limiting makes me imagine someone's got plenty of "find ... -print | xargs fgrep" in there, though! EDIT: scratch the context, since this blog post is from 2007, though cks still had over 15 years of sysadmin experience.