perl -MO=BACKEND file.pl appears to not function in 5.005_62

[p5pRT]

Migrated from rt.perl.org#1714 (status was 'resolved')

Searchable as RT1714$

[p5pRT]

From bkuhn@ebb.org

My humblest apologies if I am reporting something that is already known. I
didn't see anything go by on p5p about this, but I must admit that I do not
read some threads very closely.

I have been having a problem using Backends via the O module in 5.005_62. I
have reconfigured the source and attempted to get everything working from
the build directory. However, right after compiling, and making all tests
(successfully), the following occurs​:

$ ./perl -I lib -MO=Bytecode t/base/cond.t > x
t/base/cond.t syntax OK
$ ls -l x
-rw------- 1 bkuhn bkuhn 0 Oct 26 02​:39 x
$ cat x
$ ./perl -I lib -MO=Terse t/base/cond.t
t/base/cond.t syntax OK
$

On 5.005_61, configured the same way on the same system, both perl commands
produce copious output.

I discovered the problem while working on my own backend (which works fine
with 5.005_61, but fails in 5.005_62).

I have confirmed that the O module is calling compile() from the appropriate
Backend module; however, it appears that the anonymous sub returned by said
Backend module *never* gets called.

This simple example exhibits it​:

place this file in foo/B/TestProblem.pm
###############################################################################
package B​::TestProblem;

sub compile {
  print "Got here\n";

  return sub { print "Never gets here\n"; }
}
1;
###############################################################################

Then try​:

$ perl5.00562 -Ifoo -MO=TestProblem some_script.pl
Got here
some_script.pl syntax OK
$ perl5.00561 -Ifoo -MO=TestProblem some_script.pl
Got here
some_script.pl syntax OK
Never gets here
$

I am sorry that I cannot give any more information on the specifics of the
problem, but I am still a bit unclear on the details of how the O module
interfaces back the B modules.

Thanks, and again, sorry if this is already a known bug.

Perl Info

Site configuration information for perl 5.00562:

Configured by bkuhn at Tue Oct 26 02:09:53 EDT 1999.

Summary of my perl5 (revision 5.0 version 5 subversion 62) configuration:
  Platform:
    osname=linux, osvers=2.2.2, archname=i586-linux
    uname='linux atheist 2.2.2 #5 thu feb 25 00:45:59 est 1999 i586 unknown '
    config_args='-Dcc=/usr/local/bin/gcc -Dprefix=/usr/experimental -Doptimize=-DDEBUGGING -g -Dinstallprefix=/usr/experimental/Stow/perl-5.005_62 -Dldflags=-L/usr/lib -L/lib -L/usr/experimental/lib -L/usr/local/lib -Dlibperl=libperl.so -Duseshrplib=true -Dperladmin=bkuhn@ebb.org -Dcf_email=bkuhn@ebb.org -Ods'
    hint=recommended, useposix=true, d_sigaction=define
    usethreads=undef useperlio=undef d_sfio=undef
    use64bits=undef usemultiplicity=undef
  Compiler:
    cc='/usr/local/bin/gcc', optimize='-DDEBUGGING -g', gccversion=egcs-2.91.60 19981201 (egcs-1.1.1 release)
    cppflags='-Dbool=char -DHAS_BOOL -DDEBUGGING -fno-strict-aliasing -I/usr/local/include'
    ccflags ='-Dbool=char -DHAS_BOOL -DDEBUGGING -fno-strict-aliasing -I/usr/local/include'
    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
    alignbytes=4, usemymalloc=n, prototype=define
  Linker and Libraries:
    ld='/usr/local/bin/gcc', ldflags ='-L/usr/lib -L/lib -L/usr/experimental/lib -L/usr/local/lib'
    libpth=/usr/local/lib /lib /usr/lib
    libs=-lnsl -lndbm -lgdbm -ldb -ldl -lm -lc -lposix -lcrypt
    libc=/lib/libc-2.0.7.so, so=so, useshrplib=true, libperl=libperl.so
  Dynamic Linking:
    dlsrc=dl_dlopen.xs, dlext=so, d_dlsymun=undef, ccdlflags='-rdynamic -Wl,-rpath,/usr/experimental/lib/perl5/5.00562/i586-linux/CORE'
    cccdlflags='-fpic', lddlflags='-shared -L/usr/lib -L/lib -L/usr/experimental/lib -L/usr/local/lib'

Locally applied patches:
    


@INC for perl 5.00562:
    lib
    /usr/experimental/lib/perl5/5.00562/i586-linux
    /usr/experimental/lib/perl5/5.00562
    /usr/experimental/lib/site_perl/5.00562/i586-linux
    /usr/experimental/lib/site_perl
    .


Environment for perl 5.00562:
    HOME=/home/bkuhn
    LANG (unset)
    LANGUAGE (unset)
    LD_LIBRARY_PATH=/home/bkuhn/Files/Graduate/University-of-Cincinnati/Thesis/Perl-Source/perl5.005_62
    LOGDIR (unset)
    PATH=/usr/experimental/bin:/home/bkuhn/bin:/usr/local/bin:/sbin:/usr/bin:/bin:/usr/bin/X11:/usr/TeX/bin:/usr/games:/usr/openwin/bin:.
    PERL_BADLANG (unset)
    SHELL=/bin/tcsh

[p5pRT]

From @gsar

On Tue, 26 Oct 1999 02​:54​:36 EDT, "Bradley M. Kuhn" wrote​:

My humblest apologies if I am reporting something that is already known. I
didn't see anything go by on p5p about this, but I must admit that I do not
read some threads very closely.

When it doubt, look up the p5p archives.

  http​://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/1999-10/msg00652.html
  http​://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/1999-10/msg00586.html

Sarathy
gsar@​ActiveState.com

[p5pRT]

From [Unknown Contact. See original ticket]

Gurusamy Sarathy wrote​:

On Tue, 26 Oct 1999 02​:54​:36 EDT, "Bradley M. Kuhn" wrote​:

My humblest apologies if I am reporting something that is already known. I
didn't see anything go by on p5p about this, but I must admit that I do not
read some threads very closely.

http​://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/1999-10/msg00652.html
http​://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/1999-10/msg00586.html

Sarathy, my apologies. I did a search, but it clearly wasn't extensive
enough. :)

I read the threads in question. It appears that your STOP {} block solution
will make it possible to fix O.pm given the new -c behavior. Is this the
path that will be pursued?

(It wasn't clear from the thread in question if this would be the pursued
answer. Is somone working on this?)
--
  - bkuhn@​ebb.org - Bradley M. Kuhn - bkuhn@​gnu.org -
  http​://www.ebb.org/bkuhn

[p5pRT]

From @gsar

On Thu, 04 Nov 1999 01​:32​:25 EST, "Bradley M. Kuhn" wrote​:

I read the threads in question. It appears that your STOP {} block solution
will make it possible to fix O.pm given the new -c behavior. Is this the
path that will be pursued?

I think so, unless there are objections from others.

(It wasn't clear from the thread in question if this would be the pursued
answer. Is somone working on this?)

Since I proposed the idea, I think it's on my plate, but I wouldn't
mind getting a patch that implemented it. :-)

Doing it should be very simple--add a PL_stopav, push CVs onto it
just like it's done for INIT{}, and finally call_list(...,PL_stopav)
when returning from perl_parse(). The B​::* stuff will need
s/\bEND\b/STOP/g.

Sarathy
gsar@​ActiveState.com