daemon: add long_opts

[ihor_antonovs.family]

Summary

Long options improve readability of scripts, makes code comprehension easier.
This patch adds long options while preserving existing CLI interface.
Also --help/-h option is added

[ihor_antonovs.family]

ready for review

[pauamma_gundo.com]

usr.sbin/daemon/daemon.8

82–83

Should the defaults be given both here and in the individual flag descriptions?

89–91

I'm not sure whether here and elsewhere in DESCRIPTION where options (other than the one being described, ie -o --output-file here) are mentioned, it's better to use the long form or the short form as diff(1) and others do. Opinions welcome.

158

s/syslog tag/syslog-tag/

[ihor_antonovs.family]

usr.sbin/daemon/daemon.8

89–91

IMHO long form improves readability everywhere, I don't have to look up what -H means every time. Short options are difficult to remember unless you work with this tool every day.

But that being said I am happy to revert this if other reviewers feel the same.

[kevans]

Meta: please generate phab diffs with full context included (-U999999, just really high); it makes reviewing a bit easier in the average case where additional context in the file might affect the review process (e.g., checking for consistency)

usr.sbin/daemon/daemon.8

60

Prevailing style is to insert a comma between the short option and long option in these listings, e.g.,

.It Fl c , Fl -change-dir

usr.sbin/daemon/daemon.c

86

Might as well const this

134

I think we can drop this line, it's generally assumed that one will seek out the manpage for details.

[ihor_antonovs.family]

address comments by @kevans

[ihor_antonovs.family]

missing closing brace

[ihor_antonovs.family]

minor fix

[ihor_antonovs.family]

@kevans: fixed all comments, added diff with -U9999999

[pauamma_gundo.com]

Manual page changes LGTM. Can't attest to source code or consistency between them.

[freebsd_igalic.co]

how about --help?

usr.sbin/daemon/daemon.c

133

why not add a --help while you're at it?

this way, drop the exit(1) from usage(), and exit as appropriate from main().

[ihor_antonovs.family]

usr.sbin/daemon/daemon.c

133

I was tempted to do this, but I was not sure if reviewers will like it. Not everybody likes the scope creep. I was planning to do this in a separate revision, but I will gladly add it to this diff!

[ihor_antonovs.family]

Add --help as per @freebsd_igalic.co request

[freebsd_igalic.co]

usr.sbin/daemon/daemon.c

110

how we could pass on an option here if we're printing onto stderr or stdout, because often you'd want to foo --help | grep topic

but i'd leave this to more experienced people

[ihor_antonovs.family]

usr.sbin/daemon/daemon.c

110

I like the idea and this is easy to do, but I will keep this feature for the next diff, I want to keep this change scoped

[allanjude]

usr.sbin/daemon/daemon.c

85

l: is specified twice

233

without the braces around this if, this exit(1) runs unconditionally

[imp]

usr.sbin/daemon/daemon.c

222

usage(0) would be better, and adjust usage to take a parameter and exit that value. Avoids errors like no {} from creeping in.

[allanjude]

usr.sbin/daemon/daemon.c

222

and could double for the other requested change, 'if usage() is not an error, write to stdout instead of stderr'

[ihor_antonovs.family]

usr.sbin/daemon/daemon.c

85

It is not me, it was like that before, I promise :)

Will fix.

222

ack, will do

233

ah, my bad , that was late change after adding --help...

[imp]

usr.sbin/daemon/daemon.c

222

I'm not aware of another program does that in the tree, so I'd suggest that we don't do that here.

[ihor_antonovs.family]

Address @allanjude and @imp comments

[ihor_antonovs.family]

minor fix

[allanjude]

Reviewed by: allanjude

[ihor_antonovs.family]

add __builtin_unreachable(); for --help case
remove unnecessary usage() prototype