I do not know what linenoise is, in this context, but control-z is
basic functionality which predates libedit by decades.

My inclination would be to remove that entire section of code (from
the case statement to the continue statement) from the feature-reduced
implementation of jconsole.

Thanks,

-- 
Raul

On Mon, May 20, 2019 at 11:55 AM bill lam <[email protected]> wrote:
>
> Right, the original linenoise ignore ctrl-z. There is another
> fork that seemed has a patch for it.
>
>         case ctrl('Z'):     /* ctrl-z */
> #ifdef SIGTSTP
>             /* send ourselves SIGSUSP */
>             disableRawMode(current);
>             raise(SIGTSTP);
>             /* and resume */
>             enableRawMode(current);
>             refreshLine(current);
> #endif
>             continue;
>
> If users want the same level of functions of libedit, the
> linenoise version can never satisfy them no matter how hard
> we put efforts to improve linenoise. It is better to use the
> libedit in that case.
>
> Mon, 20 May 2019, Raul Miller написал(а):
> > On Mon, May 20, 2019 at 11:04 AM bill lam <[email protected]> wrote:
> > > jconsole can not run on rhel if users do not have sudo right to symlink
> > > some shared library. I hope this issue can be addressed properly
> > > in j901.
> >
> > This is an example of a bigger problem that can occur in a variety of 
> > contexts.
> >
> > I think the "jconsole without libedit" mechanism is a good thing. If
> > we want J to be usable in docker instances, ec2 instances, docker
> > instances in ec2 instances, etc. etc. then we want it to be simple and
> > robust. (We also would want relevant documentation and some way of
> > coping with feature drift and stale documentation in the hosting
> > environments, but that just gives you an idea of the scope of the
> > problems.)
> >
> > That said:
> >
> > (1) It's odd that this beta jconsole instance is ignoring control-Z.
> > This issue suggests that the SIGTSTP signal is being caught and then
> > jconsole is ignoring it. (It would probably be a good idea to not
> > catch signals if you do not have a specific reason for doing so. If
> > there's a reason for this one, it might need a foreign?)
> >
> > (2) If symlinks are the showstopper for the feature-enhanced jconsole,
> > it would make sense to put the symlinks in the same directory as J and
> > link against those. If there are other issues (like different versions
> > of libedit having different interfaces for needed mechanisms) it would
> > make sense to let the user compile the code that uses the libraries.
> > If the user needs privileges to compile and deploy which they do not
> > have, they can fall back on the feature reduced jconsole.
> >
> > Thanks,
> >
> > --
> > Raul
> > ----------------------------------------------------------------------
> > For information about J forums see http://www.jsoftware.com/forums.htm
>
> --
> regards,
> ====================================================
> GPG key 1024D/4434BAB3 2008-08-24
> gpg --keyserver subkeys.pgp.net --armor --export 4434BAB3
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to