perl_eval_{sv,pv} does not trap syntax errors

[p5pRT]

Migrated from rt.perl.org#3719 (status was 'resolved')

Searchable as RT3719$

[p5pRT]

From @gisle

Created by gisle@aas.no

I have this application (actually Python) that embeds perl and calls
perl_eval_pv() to pass code to it. One problem with this is that
perl_eval_pv(code, FALSE) does not set ERRSV (aka $@​) on syntax
errors.

This is a little example program showing this behaviour​:

---------------------------------------------------------
#include <EXTERN.h>
#include <perl.h>

int main(int argc, char**argv, char**env)
{
  int i;
  char *embedding[] = {"", "-e", "0"};
  PerlInterpreter *my_perl = perl_alloc();
  perl_construct(my_perl);
  perl_parse(my_perl, NULL, 3, embedding, NULL);
  perl_run(my_perl);

  perl_eval_pv("sub { ", TRUE);
  /*** the program ought to abort here ***/

  if (SvTRUE(ERRSV))
  PerlIO_printf(Perl_debug_log, "XXX %s", SvPV_nolen(ERRSV));
  perl_eval_pv("1/0", TRUE);
  if (SvTRUE(ERRSV))
  PerlIO_printf(Perl_debug_log, "YYY %s", SvPV_nolen(ERRSV));
  /*** it aborts here ***/

  printf("done\n");
}

/*
  Compile with​: cc check-eval-syntax-error.c `perl -MExtUtils​::Embed -e ccopts -e ldopts`
*/
---------------------------------------------------------

This program prints​:

  $ ./a.out
  Illegal division by zero at (eval 2) line 2.

instead of something with "syntax error in (eval 1)".

Perl Info

Flags:
    category=core
    severity=low

Site configuration information for perl v5.6.0:

Configured by gisle at Mon May 22 17:07:41 CEST 2000.

Summary of my perl5 (revision 5.0 version 6 subversion 0) configuration:
  Platform:
    osname=linux, osvers=2.2.5, archname=i686-linux
    uname='linux eik 2.2.5 #1 mon sep 13 14:53:01 cest 1999 i686 unknown '
    config_args='-Dprefix=/local/perl/5.6.0_priv -Doptimize=-g -ders'
    hint=recommended, useposix=true, d_sigaction=define
    usethreads=undef use5005threads=undef useithreads=undef usemultiplicity=undef
    useperlio=undef d_sfio=undef uselargefiles=define 
    use64bitint=undef use64bitall=undef uselongdouble=undef usesocks=undef
  Compiler:
    cc='cc', optimize='-g', gccversion=egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)
    cppflags='-DDEBUGGING -fno-strict-aliasing -I/usr/local/include'
    ccflags ='-DDEBUGGING -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 -ldbm -ldb -ldl -lm -lc -lposix -lcrypt
    libc=, 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:
    /local/perl/5.6.0_priv/lib/5.6.0/i686-linux
    /local/perl/5.6.0_priv/lib/5.6.0
    /local/perl/5.6.0_priv/lib/site_perl/5.6.0/i686-linux
    /local/perl/5.6.0_priv/lib/site_perl/5.6.0
    /local/perl/5.6.0_priv/lib/site_perl
    .


Environment for perl v5.6.0:
    HOME=/home/gisle
    LANG=POSIX
    LANGUAGE (unset)
    LD_LIBRARY_PATH (unset)
    LOGDIR (unset)
    PATH=...
    PERL_BADLANG (unset)
    SHELL=/bin/bash

[p5pRT]

From @gannett-ggreer

On Wed Aug 16 05​:18​:16 2000, gaas wrote​:

I have this application (actually Python) that embeds perl and calls
perl_eval_pv() to pass code to it. One problem with this is that
perl_eval_pv(code, FALSE) does not set ERRSV (aka $@​) on syntax
errors.

This is a little example program showing this behaviour​:

---------------------------------------------------------
#include <EXTERN.h>
#include <perl.h>

int main(int argc, char**argv, char**env)
{
int i;
char *embedding[] = {"", "-e", "0"};
PerlInterpreter *my_perl = perl_alloc();
perl_construct(my_perl);
perl_parse(my_perl, NULL, 3, embedding, NULL);
perl_run(my_perl);

perl_eval_pv("sub { ", TRUE);
/*** the program ought to abort here ***/

if (SvTRUE(ERRSV))
PerlIO_printf(Perl_debug_log, "XXX %s", SvPV_nolen(ERRSV));
perl_eval_pv("1/0", TRUE);
if (SvTRUE(ERRSV))
PerlIO_printf(Perl_debug_log, "YYY %s", SvPV_nolen(ERRSV));
/*** it aborts here ***/

printf("done\n");
}

/*
Compile with​: cc check-eval-syntax-error.c `perl -MExtUtils​::Embed
-e ccopts -e ldopts`
*/
---------------------------------------------------------

This program prints​:

$ ./a.out
Illegal division by zero at (eval 2) line 2.

instead of something with "syntax error in (eval 1)".

My embed-fu is weak, but it appears to be a simple fix.

http​://m-l.org/~perl/misc/0001-Fix-for-perl-3719-Preserve-for-perl_eval_pv-while-em.patch

http​://github.com/greerga/perl/commit/8ba06c38ccf1582a01143760851f9396c16b78f5

or attached.

--
George Greer

[p5pRT]

From @gannett-ggreer

0001-Fix-for-perl-3719-Preserve-for-perl_eval_pv-while-em.patch

From 8ba06c38ccf1582a01143760851f9396c16b78f5 Mon Sep 17 00:00:00 2001
From: George Greer <perl@greerga.m-l.org>
Date: Sun, 25 Jul 2010 21:08:40 -0400
Subject: [PATCH] Fix for [perl #3719]: Preserve $@ for perl_eval_pv() while embedding.

Perl_eval_pv() wraps around Perl_eval_sv() but Perl_eval_sv() throws
away the error before Perl_eval_pv() can see it.
---
 perl.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/perl.c b/perl.c
index d52d79f..13076b1 100644
--- a/perl.c
+++ b/perl.c
@@ -2765,7 +2765,7 @@ Perl_eval_pv(pTHX_ const char *p, I32 croak_on_error)
 
     PERL_ARGS_ASSERT_EVAL_PV;
 
-    eval_sv(sv, G_SCALAR);
+    eval_sv(sv, G_SCALAR | G_KEEPERR);
     SvREFCNT_dec(sv);
 
     SPAGAIN;
-- 
1.7.0.4

[p5pRT]

From @cpansprout

On Sun Jul 25 18​:18​:26 2010, greerga wrote​:

My embed-fu is weak,

So is mine.

but it appears to be a simple fix.

http​://m-l.org/~perl/misc/0001-Fix-for-perl-3719-Preserve-for-
perl_eval_pv-while-em.patch

http​://github.com/greerga/perl/commit/8ba06c38ccf1582a01143760851f9396c16b78f5

or attached.

Is anyone else able to review this?

[p5pRT]

From @Leont

On Mon, Jul 26, 2010 at 3​:18 AM, George Greer via RT
<perlbug-followup@​perl.org> wrote​:

My embed-fu is weak, but it appears to be a simple fix.

http​://m-l.org/~perl/misc/0001-Fix-for-perl-3719-Preserve-for-perl_eval_pv-while-em.patch

http​://github.com/greerga/perl/commit/8ba06c38ccf1582a01143760851f9396c16b78f5

or attached.

Disclaimer​: I have embed-fu, but not so much internals-fu.

The G_KEEPERR flag seems to affect eval in at least 3 places
1) in doeval, right before compilation is tried.
2) in pp_leave, as the very last thing is does. I think this is only run on
3) in eval_sv, right after a successful eval.

I suspect this patch has a wider impact than intended. I don't really
know if that's really a problem though.

Also, the fact that this G_KEEPERR behavior of eval_sv is not
documented bothers me.

Leon

[p5pRT]

From @iabyn

On Mon, Sep 27, 2010 at 12​:44​:16AM +0200, Leon Timmermans wrote​:

On Mon, Jul 26, 2010 at 3​:18 AM, George Greer via RT
<perlbug-followup@​perl.org> wrote​:

My embed-fu is weak, but it appears to be a simple fix.

http​://m-l.org/~perl/misc/0001-Fix-for-perl-3719-Preserve-for-perl_eval_pv-while-em.patch

http​://github.com/greerga/perl/commit/8ba06c38ccf1582a01143760851f9396c16b78f5

or attached.

Disclaimer​: I have embed-fu, but not so much internals-fu.

The G_KEEPERR flag seems to affect eval in at least 3 places
1) in doeval, right before compilation is tried.
2) in pp_leave, as the very last thing is does. I think this is only run on
3) in eval_sv, right after a successful eval.

I suspect this patch has a wider impact than intended. I don't really
know if that's really a problem though.

The patch is wrong, unfortunately. It only "works" in some circumstances
because it exploits another bug whereby G_KEEPERR still sets $@​ sometimes,

In particular​: under the original fix, if $@​ wasn't null when calling
eval_pv(pv,TRUE), it would croak, even if pv compiled and ran okay.

I've fixed the original bug with commit

  4aca2f6

and fixed the G_KEEPERR issue with

  6b2fb38

Also, the fact that this G_KEEPERR behavior of eval_sv is not
documented bothers me.

I've fixed that with

  be064c4

--
Thank God I'm an atheist.....

[p5pRT]

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