I just experimented with putting a deliberate error into a file with
multiple function definitions.  The error message I get upon loading it
begins with "[-613]" for the error on line 613 (origin 0).

On Mon, Oct 12, 2020 at 2:45 PM 'Michael Day' via General <
[email protected]> wrote:

> I expect the answer depends on your setup,  but, at least in mine, in
> JQt under Windows 10,  I have
> the line numbers option set to ON.   The sequence ALT-v-n toggles that
> option,  ie VIEW/Toggle Line Numbers
>
> If it's still difficult to spot what's going wrong,  you can try
> reloading your script using
>     0!:011 < <script-filename>
> which displays every line.  In case of an error,  you see something like
> this:
> <<
>     rec =: - %
>
>     nonz =:
> |syntax error
> |   nonz=:
> |[-34] c:\users\mike_\j902-user\p729.ijs
>
>
>     f   =: 3 : 0
>  >>
> ... which actually happened a few minutes ago.  (I knew where the error
> was,  but used 0!:011 for
> your benefit.)
>
> But perhaps you have a somewhat different problem,  as I don't find that
> the line-numbers are in error,
> or counter-intuitive;  they even start from 1 .
>
> Cheers,
>
> Mike
>
>
>
> On 12/10/2020 19:29, Arthur Anger wrote:
> > I recently spent 20-30 minutes tracking down the source of an error
> condition caused by an ill-formed number.  The error reported during load
> began with '[-726]', and named the offending verb, but did not display the
> line (or even just the "number").
> >
> > The verb contained only about half that number of lines.  The script had
> a few more than 700, so scrolled back from its end about that far, to an
> area in the offending verb, but did not easily spot the error.
> >
> > I copied the verb definition to a separate file, and reloaded, getting
> the identical errror message.  Only after two more splits did I get a small
> enough file to get the negative-line count close enough to the error to
> recognize it on careful reading
> >
> > My conclusion is that the error reports a count of statements, perhaps
> including empties, but also including controls such as do. and else.
> >
> > It is certainly useful to have some indicator of the error's location,
> but I find the current indicator difficult to use.
> > --A number based on physical lines in the file would be more useful.
> > --A number counted from the top would be more conventional.
> > --A display of the offending character string would allow an immediate
> search for it, which should nearly always be a unique target.
> >
> > Thanks in advance to the team that considers an improvement in error
> reporting and--perhaps--chooses and manages to install one.
> > --Art
> > ----------------------------------------------------------------------
> > For information about J forums see http://www.jsoftware.com/forums.htm
>
>
> --
> This email has been checked for viruses by Avast antivirus software.
> https://www.avast.com/antivirus
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>


-- 

Devon McCormick, CFA

Quantitative Consultant
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to