Capsicumize traceroute

[oshogbo]

Summary

PR: 193973 (+ some of my changes)

[emaste]

Please upload with context, e.g. git diff -U9999 or the svn command from https://wiki.freebsd.org/Phabricator

contrib/traceroute/traceroute.c

561

Is this message correct in the current libcasper world order? Should it be an error?

[oshogbo]

Thank you emaste@.

[bapt]

contrib/traceroute/traceroute.c

563

Given the if block above will exit, no need for a else block here

589

casper cannot be null at this point given there checks above that exits

1013

This will always be true there is a check above for that.

1848–1853

maybe if would be more readable to have a bunch of defined earlier:

#if HAVE_LIBCAPSER
#define tcrt_gethostbyname(domain) cpa_gethostbyname(capdns, domain)
#else 
#define tcrt_gethostbyname(domain) cpa_gethostbyname(capdns, domain)
#endif

with all functions like that

[emaste]

contrib/traceroute/traceroute.c

230

we compile with -DHAVE_LIBCASPER; does this #if work?

370–374

maybe lose the extra space here, like you do with HAVE_LIBCASPER stack variables in fns?

561–562

errx instead of perror/exit?

1024

given that we already have cansandbox variable I'd prefer if we set if false if capsicum is not available, and drop the ENOSYS case

usr.sbin/traceroute/Makefile

3–4

it seems odd to have both of these; afaik we need only src.opts.mk?

[oshogbo]

Fixes proposed by emaste and bapt.

[oshogbo]

contrib/traceroute/traceroute.c

230

Not sure if I get this.

1848–1853

When we will have libcaspermock this will not be needed any more.

[oshogbo]

One more fix.

[oshogbo]

contrib/traceroute/traceroute.c

230

After one more look I get this ;)

[emaste]

contrib/traceroute/traceroute.c

565–566

what about initializing these with const char *types[] = { "NAME", "ADDR" };?

[allanjude]

tested fine here