From @andk

git bisect

13b0f67 is the first bad commit
commit 13b0f67
Author​: David Mitchell <davem@​iabyn.com>
Date​: Wed May 22 16​:38​:29 2013 +0100

  re-enable Copy-on-Write by default.

sample fail report

http​://www.cpantesters.org/cpan/report/2c57d796-c6d6-11e2-954c-7bff979cd64d

perl -V

Summary of my perl5 (revision 5 version 19 subversion 1) configuration​: |
  Commit id​: 13b0f67 |
  Platform​: |
  osname=linux, osvers=3.9-1-amd64, archname=x86_64-linux-ld |
  uname='linux k83 3.9-1-amd64 #1 smp debian 3.9.8-1 x86_64 gnulinux ' |
  config_args='-Dprefix=/home/sand/src/perl/repoperls/installed-perls/perl/v5.|
19.0-246-g13b0f67/127e -Dmyhostname=k83 -Dinstallusrbinperl=n -Uversiononly -Dus|
edevel -des -Ui_db -Uuseithreads -Duselongdouble -DDEBUGGING=-g' |
  hint=recommended, useposix=true, d_sigaction=define |
  useithreads=undef, usemultiplicity=undef |
  useperlio=define, d_sfio=undef, uselargefiles=define, usesocks=undef |
  use64bitint=define, use64bitall=define, uselongdouble=define |
  usemymalloc=n, bincompat5005=undef
  Compiler​:
  cc='cc', ccflags ='-fno-strict-aliasing -pipe -fstack-protector -I/usr/local
/include -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64',
  optimize='-O2 -g',
  cppflags='-fno-strict-aliasing -pipe -fstack-protector -I/usr/local/include'
  ccversion='', gccversion='4.8.1', gccosandvers=''
  intsize=4, longsize=8, ptrsize=8, doublesize=8, byteorder=12345678
  d_longlong=define, longlongsize=8, d_longdbl=define, longdblsize=16
  ivtype='long', ivsize=8, nvtype='long double', nvsize=16, Off_t='off_t', lse
eksize=8
  alignbytes=16, prototype=define
  Linker and Libraries​:
  ld='cc', ldflags =' -fstack-protector -L/usr/local/lib'
  libpth=/usr/local/lib /lib/x86_64-linux-gnu /lib/../lib /usr/lib/x86_64-linu
x-gnu /usr/lib/../lib /lib /usr/lib
  libs=-lnsl -lgdbm -ldb -ldl -lm -lcrypt -lutil -lc -lgdbm_compat
  perllibs=-lnsl -ldl -lm -lcrypt -lutil -lc
  libc=, so=so, useshrplib=false, libperl=libperl.a
  gnulibc_version='2.17'
  Dynamic Linking​:
  dlsrc=dl_dlopen.xs, dlext=so, d_dlsymun=undef, ccdlflags='-Wl,-E'
  cccdlflags='-fPIC', lddlflags='-shared -O2 -g -L/usr/local/lib -fstack-prote
ctor'
 
 
Characteristics of this binary (from libperl)​:
  Compile-time options​: HAS_TIMES PERLIO_LAYERS PERL_DONT_CREATE_GVSV
  PERL_HASH_FUNC_ONE_AT_A_TIME_HARD PERL_MALLOC_WRAP
  PERL_NEW_COPY_ON_WRITE PERL_NEW_COPY_ON_WRITE
  PERL_PRESERVE_IVUV PERL_USE_DEVEL USE_64_BIT_ALL
  USE_64_BIT_INT USE_LARGE_FILES USE_LOCALE
  USE_LOCALE_COLLATE USE_LOCALE_CTYPE
  USE_LOCALE_NUMERIC USE_LONG_DOUBLE USE_PERLIO
  USE_PERL_ATOF
  Built under linux
  Compiled at Jul 4 2013 08​:11​:52
  @​INC​:
  /home/sand/src/perl/repoperls/installed-perls/perl/v5.19.0-246-g13b0f67/127e
/lib/site_perl/5.19.1/x86_64-linux-ld
  /home/sand/src/perl/repoperls/installed-perls/perl/v5.19.0-246-g13b0f67/127e
/lib/site_perl/5.19.1
  /home/sand/src/perl/repoperls/installed-perls/perl/v5.19.0-246-g13b0f67/127e
/lib/5.19.1/x86_64-linux-ld
  /home/sand/src/perl/repoperls/installed-perls/perl/v5.19.0-246-g13b0f67/127e
/lib/5.19.1
  .

--
andreas

From @Leont

On Thu, Jul 4, 2013 at 8​:28 AM, Andreas J. Koenig via RT
<perlbug-followup@​perl.org> wrote​:

# New Ticket Created by (Andreas J. Koenig)
# Please include the string​: [perl #118747]
# in the subject line of all future correspondence about this issue.
# <URL​: https://rt-archive.perl.org/perl5/Ticket/Display.html?id=118747 >

git bisect
----------
13b0f67 is the first bad commit
commit 13b0f67
Author​: David Mitchell <davem@​iabyn.com>
Date​: Wed May 22 16​:38​:29 2013 +0100

re\-enable Copy\-on\-Write by default\.

sample fail report
------------------
http​://www.cpantesters.org/cpan/report/2c57d796-c6d6-11e2-954c-7bff979cd64d

Yeah, or better said blead re-breaks it, for previously discussed
reasons (s/foo/bar/ should not reallocate the buffer to make it COW).

Leon

The RT System itself - Status changed from 'new' to 'open'

From @cpansprout

On Thu Jul 04 00​:05​:46 2013, LeonT wrote​:

On Thu, Jul 4, 2013 at 8​:28 AM, Andreas J. Koenig via RT
<perlbug-followup@​perl.org> wrote​:

# New Ticket Created by (Andreas J. Koenig)
# Please include the string​: [perl #118747]
# in the subject line of all future correspondence about this issue.
# <URL​: https://rt-archive.perl.org/perl5/Ticket/Display.html?id=118747 >

git bisect
----------
13b0f67 is the first bad commit
commit 13b0f67
Author​: David Mitchell <davem@​iabyn.com>
Date​: Wed May 22 16​:38​:29 2013 +0100

re\-enable Copy\-on\-Write by default\.

sample fail report
------------------
http​://www.cpantesters.org/cpan/report/2c57d796-c6d6-11e2-954c-
7bff979cd64d

Yeah, or better said blead re-breaks it, for previously discussed
reasons (s/foo/bar/ should not reallocate the buffer to make it COW).

Do you remember which ticket that was in?

--

Father Chrysostomos

From @Leont

On Tue, Aug 20, 2013 at 5​:41 PM, Father Chrysostomos via RT <
perlbug-followup@​perl.org> wrote​:

Yeah, or better said blead re-breaks it, for previously discussed
reasons (s/foo/bar/ should not reallocate the buffer to make it COW).

Do you remember which ticket that was in?

It was discussed in #115910

Leon

From @cpansprout

On Tue Aug 20 09​:10​:47 2013, LeonT wrote​:

On Tue, Aug 20, 2013 at 5​:41 PM, Father Chrysostomos via RT <
perlbug-followup@​perl.org> wrote​:

Yeah, or better said blead re-breaks it, for previously discussed
reasons (s/foo/bar/ should not reallocate the buffer to make it COW).

Do you remember which ticket that was in?

It was discussed in #115910

Thank you.

’Turns out s/// on SvLEN==0 was already broken​:

$ perl5.18.1 -e 'for(!${\0}) { Internals​::SvREADONLY($_,0); s/1/1/; }'
Bus error​: 10

--

Father Chrysostomos

From @cpansprout

On Wed Aug 21 00​:01​:34 2013, sprout wrote​:

On Tue Aug 20 09​:10​:47 2013, LeonT wrote​:

On Tue, Aug 20, 2013 at 5​:41 PM, Father Chrysostomos via RT <
perlbug-followup@​perl.org> wrote​:

Yeah, or better said blead re-breaks it, for previously discussed
reasons (s/foo/bar/ should not reallocate the buffer to make it
COW).

Do you remember which ticket that was in?

It was discussed in #115910

Thank you.

’Turns out s/// on SvLEN==0 was already broken​:

$ perl5.18.1 -e 'for(!${\0}) { Internals​::SvREADONLY($_,0); s/1/1/; }'
Bus error​: 10

Er, please ignore that.

It has to do with this logic​:

  /* can do inplace substitution? */
  if (c
#ifdef PERL_ANY_COW
  && !is_cow
#endif
  && (I32)clen <= RX_MINLENRET(rx)
  && (once || !(r_flags & REXEC_COPY_STR))
  && !(RX_EXTFLAGS(rx) & RXf_NO_INPLACE_SUBST)
  && (!doutf8 || SvUTF8(TARG))
  && !(rpm->op_pmflags & PMf_NONDESTRUCT))
  {

This line prevents the in-place substitution in File​::Map’s tests​:

  && (once || !(r_flags & REXEC_COPY_STR))

r_flags & REXEC_COPY_STR is always true now. File​::Map fails in earlier
perls if $& is mentioned.

I don’t quite understand the logic here.

--

Father Chrysostomos

From @cpansprout

On Wed Aug 21 00​:36​:09 2013, sprout wrote​:

It has to do with this logic​:

/\* can do inplace substitution? \*/
if \(c

#ifdef PERL_ANY_COW
&& !is_cow
#endif
&& (I32)clen <= RX_MINLENRET(rx)
&& (once || !(r_flags & REXEC_COPY_STR))
&& !(RX_EXTFLAGS(rx) & RXf_NO_INPLACE_SUBST)
&& (!doutf8 || SvUTF8(TARG))
&& !(rpm->op_pmflags & PMf_NONDESTRUCT))
{

This line prevents the in-place substitution in File​::Map’s tests​:

    && \(once || \!\(r\_flags & REXEC\_COPY\_STR\)\)

r_flags & REXEC_COPY_STR is always true now. File​::Map fails in earlier
perls if $& is mentioned.

I don’t quite understand the logic here.

I don’t think that logic is correct or ever was. Match vars can point
into the buffer copied by the initial regexp match that comes before this.

I removed that line in 1555b32 and File​::Map passes its tests again.

--

Father Chrysostomos

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

From @cpansprout

On Wed Aug 21 22​:45​:13 2013, sprout wrote​:

I removed that line in 1555b32 and File​::Map passes its tests again.

That fix turned out to be wrong, so I redid it in 9cefd26.

--

Father Chrysostomos