Dan Sugalski ([EMAIL PROTECTED]) wrote:
> At 10:23 AM 4/4/00 -0400, Jordan Henderson wrote:
> >Maybe VMS.C should put out a message to stderr if PERL_ROOT is not defined
> >or if
> >the VMSPIPE.COM can't be found when it is required. That might help.
> Or we could, I suppose, embed a copy of VMSPIPE.COM inside VMS.C and dump
> it out to SYS$SCRATCH if we can't find a real one, though that's probably
> going a little *too* far...
*Sigh* that way lies the "writing tempfiles" solution...
I don't think it's a problem to write out a reasonably informative
error message, since the problem is localized at a single place in the
code. We certainly should do so if we're doing some sort of active
search for VMSPIPE, and point the user in the right direction hopefully.
> When you get right down to it, perl's hosed if it's library's
> dysfunctional, so if folks *do* mess things up we can just tell 'em "Don't
> *do* that!"
%PERL-F-DOPESLAP, logical PERL_ROOT not found, don't *DO* that!
I do wonder what kind of weird situations one might run up against:
perl embedded with libraries not used/libraries on a remote DECNET
node that's unreachable most of the time; etc. etc. But at some
point you have to say 'enough!'
--
Drexel University \V --Chuck Lane
----------------->--------*------------<[EMAIL PROTECTED]
(215) 895-1545 / \ Particle Physics [EMAIL PROTECTED]
FAX: (215) 895-5934 /~~~~~~~~~~~ [EMAIL PROTECTED]
