Thanks, Raul.

These are impossible errors to trigger just by using J code. On reflection
i guess that's going to be the story with most if not all of them.


On Tue, May 20, 2014 at 11:58 PM, Raul Miller <[email protected]> wrote:

> File access error can happen when permission is denied by the host
> operating system.
>
> For example:
>
> $ date >/tmp/example; chmod 0 /tmp/example; ijconsole
>    1!:1<'/tmp/example'
> |file access error
> |       1!:1<'/tmp/example'
>
> I think allocation error is similar - I expect that it would happen when J
> asks for memory and none is available (it's analogous to ws full - but it's
> not happening inside J's accounting system).
>
> Thanks,
>
> --
> Raul
>
>
> On Tue, May 20, 2014 at 6:08 PM, Roger Hui <[email protected]
> >wrote:
>
> >    s=: 1 $. 100;0;2
> >    t=: 1 $. 200;0;3
> >    s,"0/t
> > |non-unique sparse elements
> > |   s    ,"0/t
> >
> > For that matter, I'd expected that s,t would signal the same error, but
> it
> > doesn't.  I guess if I looked at the source I can figure out why that is.
> >
> >    s,t
> > |nonce error
> > |   s    ,t
> >
> > In contrast:
> >
> >    s ,"0/ s
> >    $ s ,"0/ s
> > 100 100 2
> >    3 $. s ,"0/ s  NB. the sparse element
> > 2
> >
> > I expect "read only data" to be signaled if you try to amend data which
> is
> > mapped on a read-only file (e.g. if the mapped file is on a CDROM).
> >
> > "stop" would be signaled when execution reaches a line in an explicit
> defn
> > which has stop set.
> >
> > "break" would be signaled if (on Windows) you hit the yellow J icon while
> > the J session is executing.  e.g. execute 6!:3]100 and then hit the
> yellow
> > icon 2 or 3 times.
> >
> > "input interrupt" would be signaled if you asked for "terminal input",
> > 1!:1]1, hit the yellow icon, and then enter.  It dates from a time when
> it
> > was possible to enter o backspace u backspace t (remember that?) in
> > response to 1!:1]1.
> >
> > "system error" is not supposed to happen.
> >
> > Not sure about "file access error".
> >
> >
> >
> >
> >
> > On Tue, May 20, 2014 at 2:32 PM, Ian Clark <[email protected]>
> wrote:
> >
> > > For NuVoc purposes I'm documenting J's error messages...
> > >
> > > Have you ever seen any of the following error messages in your J
> session
> > > (j602 onwards)?
> > >
> > >     allocation error
> > >     break
> > >     file access error
> > >     input interrupt
> > >     non-unique sparse elements
> > >     read-only data
> > >     stop
> > >     system error
> > >
> > > If so, can you furnish a simple sample of "normal" J code that causes
> the
> > > said error?
> > >
> > > I specify "normal" because I don't want samples which rely on
> > constructing
> > > a pathological noun as a core image in hex and shoe-horning it into J
> by
> > > use of 3!:2, or use of the "darker" Foreigns, especially undocumented
> > ones.
> > > I want something which could conceivable arise by bona-fide J usage by
> an
> > > innocent newbie, ill-conceived though it might seem to someone more
> > > experienced.
> > >
> > > Having said that, it probably takes deep knowledge of the J source to
> > know
> > > what might trigger some of these errors.
> > >
> > > It would also be nice to know which of the above error messages can't
> > > possibly arise in normal usage. (But consider it "cheating" to use
> 13!:8
> > to
> > > signal any given message: assume I know how to do that.)
> > > ----------------------------------------------------------------------
> > > For information about J forums see http://www.jsoftware.com/forums.htm
> > >
> > ----------------------------------------------------------------------
> > For information about J forums see http://www.jsoftware.com/forums.htm
> >
> ----------------------------------------------------------------------
> 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