Path: utzoo!attcan!uunet!cs.utexas.edu!sdd.hp.com!zaphod.mps.ohio-state.edu!mips!prls!pyramid!unify!csusac!ucdavis!sunny!poage
From: poage@sunny.ucdavis.edu (Tom Poage)
Newsgroups: comp.lang.perl
Subject: "Out of memory!"?
Message-ID: <482@sunny.ucdavis.edu>
Date: 13 Jul 90 20:06:01 GMT
Reply-To: poage@sunny.ucdavis.edu (Tom Poage)
Organization: UCDMC Clinical Engineering, Sacto., CA
Lines: 55

I'm tinkering with redirecting stdout and stderr
to files and then looking at the results.  However,
one variant gives me "Out of memory!" (Sun 3/150, 4.0.3).

#START
open(SAVEOUT, ">&STDOUT");	# Save file descriptors.
open(SAVEERR, ">&STDERR");
open(STDOUT, ">$OFILE");	# Output file.
open(STDERR, "+>$EFILE");	# Error file.

#system('/bin/MyProgram', @COMMANDS);
print STDERR "multiple\nline\nSQL:\nstuff\n"; #DEBUG:simulate error.

close(STDOUT);
open(STDOUT, ">&SAVEOUT");
#close(STDERR);
#open(STDERR, ">&SAVEERR");

open(PRGOUT, $OFILE);		# Let's look at the output.
#open(EFILE, $EFILE);		# Let's look for errors.
seek(STDERR, 0, 0);		# We'll just rewind the error file instead.

if (@errs = grep( do{/^SQL:/ ? undef : s/^/\t/o;}, <STDERR>)) {
	print "Error message(s):\n", @errs; #DEBUG
} else {
	print while <PRGOUT>; #DEBUG
}
#END

Now, what I want to do is grub through what was produced to stderr,
remove purely informational stuff (/^SQL/) and prepend a tab to all
the real errors (a format, sorta).

This works properly

	if (@errs = grep(s/^/\t/, grep(!/^SQL:/, <STDERR>))) {
	....

So does this

	@messages = <STDERR>;
	if (@errs = grep( do{/^SQL:/ ? undef : s/^/\t/o;}, @messages)) {
	....

But this

	if (@errs = grep( do{/^SQL:/ ? undef : s/^/\t/o;}, <STDERR>)) {
	....

produces "Out of memory!".  Did I do something to deserve this?
Anyone care to comment?  Tom.
-- 
Tom Poage, Clinical Engineering
Universiy of California, Davis, Medical Center, Sacramento, CA
poage@sunny.ucdavis.edu  {...,ucbvax,uunet}!ucdavis!sunny!poage