Re: ftrace

[Tom Tromey <tromey at redhat dot com>]

>>>>> "Andrew" == Andrew Cagney <cagney@redhat.com> writes:

Andrew> If you like it better than jargs, then I'm guessing it isn't the GNU
Andrew> getopt-long JNI wrapper based implementation.  We picked up jargs
Andrew> because the JNI getopt-long interface was so horrendous.  Is there
Andrew> javadoc online for this alternative?

This is a new implementation I wrote for Classpath.

I didn't see javadoc online... the source code is fully documented
though.  It is in the GNU Classpath cvs repository.  Here's a URL for
convenient browsing:

  http://cvs.savannah.gnu.org/viewcvs/classpath/tools/gnu/classpath/tools/getopt/?root=classpath

This getopt isn't perfect (there's a couple smaller features I still
want to add) but IMNSHO it is pretty good.  It handles most argument
parsing approaches (joined single-char options, like -Ipath, are on
the to-do list), automatically handles --help and --version, and
allows some user control over grouping of options in --help output.

Andrew> The main issues with adopting another getopt alternative are: that it
Andrew> be sane; that we only have one (we're not in the business of
Andrew> babysitting multiple redundant packages - which would mean eliminating
Andrew> jargs); and its license be workable.

* Sane... well, I think so :-)
* Eliminating jargs... at your discretion of course
* License... GPL+E like the rest of Classpath

Tom