From @aigan

Created by jonas@astral.paranormal.o.se

There is a incinsistent change of behaviour between v5.005_03 and v5.6.0

Only v5.005_03 calls END for this​:

#!/usr/local/bin/perl
END {print "After\n"}
priint "error";
__END__

But both calls END for this compile error​:

#!/usr/local/bin/perl
END {print "After\n"}
compile error now;
__END__

I would prefere the END to be called so that it could clean up all
things done in BEGIN.

Flags:
    category=core
    severity=medium

Site configuration information for perl v5.6.0:

Configured by jonas at Sat Apr  1 02:14:50 CEST 2000.

Summary of my perl5 (revision 5.0 version 6 subversion 0) configuration:
  Platform:
    osname=linux, osvers=2.2.12, archname=i586-linux-thread-multi
    uname='linux astral.paranormal.o.se 2.2.12 #1 thu oct 14 09:29:24 est 1999 i586 unknown '
    config_args='-Uinstallusrbinperl -Dusethreads -des'
    hint=recommended, useposix=true, d_sigaction=define
    usethreads=define use5005threads=undef useithreads=define usemultiplicity=define
    useperlio=undef d_sfio=undef uselargefiles=define 
    use64bitint=undef use64bitall=undef uselongdouble=undef usesocks=undef
  Compiler:
    cc='cc', optimize='-O2', gccversion=2.95.2 20000220 (Debian GNU/Linux)
    cppflags='-D_REENTRANT -fno-strict-aliasing -I/usr/local/include'
    ccflags ='-D_REENTRANT -fno-strict-aliasing -I/usr/local/include -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64'
    stdchar='char', d_stdstdio=define, usevfork=false
    intsize=4, longsize=4, ptrsize=4, doublesize=8
    d_longlong=define, longlongsize=8, d_longdbl=define, longdblsize=12
    ivtype='long', ivsize=4, nvtype='double', nvsize=8, Off_t='off_t', lseeksize=8
    alignbytes=4, usemymalloc=n, prototype=define
  Linker and Libraries:
    ld='cc', ldflags =' -L/usr/local/lib'
    libpth=/usr/local/lib /lib /usr/lib
    libs=-lnsl -lndbm -lgdbm -ldb -ldl -lm -lpthread -lc -lposix -lcrypt
    libc=/lib/libc-2.1.3.so, so=so, useshrplib=false, libperl=libperl.a
  Dynamic Linking:
    dlsrc=dl_dlopen.xs, dlext=so, d_dlsymun=undef, ccdlflags='-rdynamic'
    cccdlflags='-fpic', lddlflags='-shared -L/usr/local/lib'

Locally applied patches:
    


@INC for perl v5.6.0:
    /usr/local/lib/perl5/5.6.0/i586-linux-thread-multi
    /usr/local/lib/perl5/5.6.0
    /usr/local/lib/perl5/site_perl/5.6.0/i586-linux-thread-multi
    /usr/local/lib/perl5/site_perl/5.6.0
    /usr/local/lib/perl5/site_perl
    .


Environment for perl v5.6.0:
    HOME=/home/jonas
    LANG (unset)
    LANGUAGE (unset)
    LD_LIBRARY_PATH (unset)
    LOGDIR (unset)
    PATH=/usr/local/bin:/usr/bin:/bin:/usr/bin/X11:/usr/games:/usr/local/games:/usr/X11R6/bin:/home/jonas/bin:.
    PERL_BADLANG (unset)
    SHELL=/bin/bash

From @tamias

On Mon, Apr 24, 2000 at 05​:36​:34PM +0200, Jonas Liljegren wrote​:

But both calls END for this compile error​:

#!/usr/local/bin/perl
END {print "After\n"}
compile error now;
__END__

That one's a runtime error, not a compile error. It's compiled as
'error'->compile('now');

Ronald

From [Unknown Contact. See original ticket]

jonas@​paranormal.o.se (Jonas Liljegren) wrote

There is a incinsistent change of behaviour between v5.005_03 and v5.6.0

Only v5.005_03 calls END for this​:

#!/usr/local/bin/perl
END {print "After\n"}
priint "error";
__END__

This is a deliberate change. From "perldoc perldelta"​:

  END blocks and the `-c' switch

  Prior versions used to run BEGIN and END blocks when Perl
  was run in compile-only mode. Since this is typically not
  the expected behavior, END blocks are not executed anymore
  when the `-c' switch is used.

Hmmm.. wait a moment - that doesn't actually cover the case you quote.
Seems like the attached doc patch is needed.

Mike Guy

--- ./pod/perldelta.pod.orig	Mon Apr 24 17:28:32 2000
+++ ./pod/perldelta.pod	Mon Apr 24 17:04:39 2000
@@ -1811,7 +1811,7 @@
 Prior versions used to run BEGIN B<and> END blocks when Perl was
 run in compile-only mode.  Since this is typically not the expected
 behavior, END blocks are not executed anymore when the C<-c> switch
-is used.
+is used, or if compilation fails.
 
 See L<CHECK blocks> for how to run things when the compile phase ends.
 
--- ./pod/perlmod.pod.orig	Mon Mar 13 21:25:37 2000
+++ ./pod/perlmod.pod	Mon Apr 24 17:29:42 2000
@@ -233,7 +233,7 @@
 (if you can).)  You may have multiple C<END> blocks within a file--they
 will execute in reverse order of definition; that is: last in, first
 out (LIFO).  C<END> blocks are not executed when you run perl with the
-C<-c> switch.
+C<-c> switch, or if compilation fails..
 
 Inside an C<END> subroutine, C<$?> contains the value that the program is
 going to pass to C<exit()>.  You can modify C<$?> to change the exit
@@ -252,7 +252,7 @@
 
 When you use the B<-n> and B<-p> switches to Perl, C<BEGIN> and
 C<END> work just as they do in B<awk>, as a degenerate case.  As currently
-implemented (and subject to change, since its inconvenient at best),
+implemented (and subject to change, since it's inconvenient at best),
 both C<BEGIN> and<END> blocks are run when you use the B<-c> switch
 for a compile-only syntax check, although your main code is not.
 

End of patch

From @aigan

"M.J.T. Guy" wrote​:

jonas@​paranormal.o.se (Jonas Liljegren) wrote

There is a incinsistent change of behaviour between v5.005_03 and v5.6.0

Only v5.005_03 calls END for this​:

#!/usr/local/bin/perl
END {print "After\n"}
priint "error";
__END__

This is a deliberate change.

Why should END not be called?

I am using that behaviour in a module that checks STDERR​: CGI​::Debug
It works like this​:

program.cgi​:

#!/usr/bin/perl -w
use CGI​::Debug;
print "Content-type​: text/html\n\n";
priint "<h1>Hi!</h1>\n";
__END__

CGI/Debug.pm (in theory)​:

package CGI​::Debug;
BEGIN
{
  open (STDERR,">/tmp/error-$$");
}
END
{
  close STDERR;
  print "Content-type​: text/plain\n\n";
  open ERR, "/tmp/error-$$";
  print while <ERR>;
}
__END__

(But the whole thing is about 1000 lines...)

Why shouldn't I be able to do this now?

The main idea of the module was that I could just put the
"use CGI​::Debug;" line and get my errors to the browser without having
to change anything else in the program. And it should catch all
errors. That worked fine until now...

/jonas

From [Unknown Contact. See original ticket]

"M.J.T. Guy" wrote​:

jonas@​paranormal.o.se (Jonas Liljegren) wrote

There is a incinsistent change of behaviour between v5.005_03 and v5.6.0

Only v5.005_03 calls END for this​:

#!/usr/local/bin/perl
END {print "After\n"}
priint "error";
__END__

This is a deliberate change.

Why should END not be called?

Because in 5.6.0, then end-of-compilation phase is CHECK{}, not END{},
and the beginning of the runtime phase is INIT{}. Therefore you should
use CHECK{} to clean up non-ephemeral changes done by BEGIN{}, and use
INIT{} to start things that will be cleaned up by END{}.

package CGI​::Debug;
BEGIN
{
open (STDERR,">/tmp/error-$$");
}
END
{
close STDERR;
print "Content-type​: text/plain\n\n";
open ERR, "/tmp/error-$$";
print while <ERR>;
}

Here is the 5.6.0 way of doing that​:

BEGIN { # Start of compilation phase; put this at the top of the file.
  our $error_log = "/tmp/error-$$";
  open (STDERR,">$error_log");
}
CHECK { # End of compilation phase
  $_ = close STDERR ? "" : "close(STDERR)​: $!\n";
  if ($_ ne "" || -s $error_log) {
  local($\); # Read in entire file
  $_ .= open(ERR,$error_log) ? <ERR> : "Could not read error log​: $!";
  print "Content-type​: text/plain\n\n Compilation errors detected​:\n$_\n";
  unlink $error_log ? die("\n") : die("Clean-up failure​: $!\n");
  }
}

INIT { # Start of run time
  open (STDERR,">$error_log");
}
END { # End of run time
  $_ = close STDERR ? "" : "close(STDERR)​: $!\n";
  if ($_ ne "" || -s $error_log) {
  local($\);
  $_ .= open(ERR,$error_log) ? <ERR> : "Could not read error log​: $!";
  print "Content-type​: text/plain\n\n Runtime errors detected​:\n$_\n";
  unlink $error_log ? die("\n") : die("Clean-up failure​: $!\n");
  }
  unlink $error_log or print "&lt;BR&gt;Clean-up failure​: $!\n";
}
# Rest of module goes here.

From @iabyn

Doesn't appear to be a bug.

@iabyn - Status changed from 'open' to 'resolved'