Re: Subject: DEFINE/LOCAL ERROR/...!!!!

[Klaus Banse <kbanse@eso.org>]

Dear Midas users,
Sender: owner-midas-users
Precedence: bulk
Reply-To: midas-users

   as stated in the Midas doc, local keywords should not
have the same names as "system" keywords, which are e.g. used
to hold names of currently assigned printers, etc.

Unfortunately, the check of these name conflicts never
worked as I found out just early this year. And so the 02FEB
Midas release was the first, where local keywords really were
checked. I had hoped to find all those procedures with
"invalid" local keynames, but unfortunately some procedures
in the long context were still using "bad" local keyword names.

This has been fixed in the upcoming 02SEP release.
However, if you cannot wait for the new release, just go to
the directory /midas/version/stdred/long/proc
and do:


kbanse@dmdsol8b$ grep -i erro *.prg | grep -i loc
lnaver.prg:def/loc error/I/1/1  0
lnextr.prg:DEFINE/LOCAL  ERROR/I/1/1  0
lnhough.prg:define/local  error/I/1/1  1
lnrect.prg:DEFINE/LOCAL ERRSAV/I/1/1 'ERROR(3)'  ! disable errors in
regression...
lnverif.prg:DEFINE/LOCAL ERROR/I/1/1   0

in the procedures
lnaver.prg, lnextr.prg, lnhough.prg, lnverif.prg

replace:
define/local error/...
by:
define/local errors/...
and all references to that keyword as well in these procedures
(nothing has to be done to lnrect.prg).

Then don't forget to do
$ make
in that directory, which links procedures in this directory
to the main stdred/proc directory

After that, all "long" stuff should work again.



Sorry for that hickup with the context long, and good luck,
Klaus Banse