develooper Front page | perl.perl5.porters | Postings from October 2003

Re: [perlguts] s/USE_5005THREADS/USE_ITHREADS/?

Thread Previous | Thread Next
H.Merijn Brand
October 24, 2003 03:22
Re: [perlguts] s/USE_5005THREADS/USE_ITHREADS/?
Message ID:
On Thu 23 Oct 2003 23:37, Stas Bekman <> wrote:
> Should perlguts.pod be updated to use USE_ITHREADS instead of/in addition to 
> USE_5005THREADS? Or does MULTIPLICITY implies USE_ITHREADS. From config.h:
> #  if !defined(MULTIPLICITY)
> #    define MULTIPLICITY
> #  endif
> #endif

There are more points of attention here as posted on 03-10-2002 in

Removal in Configure will be a seperate task (maybe take 2)

- ext/B/B.xs now leaves an empty sub threadsv_names. Remove?

  625 void
  626 threadsv_names()
  627     PPCODE:
- ext/B/B/ should be updated seperately for the docs?

- ext/Thread/Thread.xs leaves a lot of empty subs. There might be a more
  insightful cleanup needed over there

- I /think/ the comment on line 87 of intrp.h can just go, but I'm not sure

  /* this one needs to be moved to thrdvar.h and accessed via
   * find_threadsv() when USE_5005THREADS */
  PERLVAR(Ireplgv,        GV *)

- perl.c #124 dTHX; is in between USE_5005THREADS. is this correct?
- perl.c #308 same

- perl.h #1954 How valid is the rest of this comment
- perl.h #2524 leaves empty define for DEBUG_S, but it is still used

  ext/Thread/Thread.xs:    DEBUG_S(PerlIO_printf(Perl_debug_log,
  ext/Thread/Thread.xs:   DEBUG_S(PerlIO_printf(Perl_debug_log,
  ext/Thread/typemap:         DEBUG_S(PerlIO_printf(Perl_debug_log,
  perl.c: DEBUG_S(PerlIO_printf(Perl_debug_log, "main thread is 0x%"UVxf"\n",
  perl.c:    DEBUG_S(PerlIO_printf(Perl_debug_log, "my_exit: thread %p, status %lu\n",
  perl.h:#define DEBUG_S_FLAG             0x00010000 /*  65536 */
  perl.h:#  define DEBUG_S_TEST_ (PL_debug & DEBUG_S_FLAG)
  perl.h:#  define DEBUG_S_TEST DEBUG_S_TEST_
  perl.h:#  define DEBUG_S(a)
  perl.h:#  define DEBUG_S_TEST (0)
  perl.h:#  define DEBUG_S(a)
  pp_hot.c:           DEBUG_S(PerlIO_printf(Perl_debug_log,
  pp_hot.c:       DEBUG_S(PerlIO_printf(Perl_debug_log,
  scope.c:            DEBUG_S(PerlIO_printf(Perl_debug_log,
  util.c:    DEBUG_S(PerlIO_printf(Perl_debug_log,
  util.c:    DEBUG_S(PerlIO_printf(Perl_debug_log,
  util.c:    DEBUG_S(PerlIO_printf(Perl_debug_log,
  util.c:    DEBUG_S(PerlIO_printf(Perl_debug_log, "croak: 0x%"UVxf" %s",

- Someone with better understanding should modify pod/perlguts.pod. If I just
  strip the 5005 issues, I might introduce ambiguities
- Same for INSTALL
- Same for pod/perlthrtut.pod

- pp_ctl.c #769 Comment should go?

  /* SAVE_DEFSV does *not* suffice here for USE_5005THREADS */
  ENTER;                                      /* enter inner scope */

- pp_ctl.c #2631 Any implications of these comments?

- pp_hot.c #1795 Comment still applicable?

- regexec.c #2057 What about this comment?

- thrdvar.h #21 Comment confuses me. please look into this

- vms/ #160 Have a second look
- same for

- t/op/nothr5005.t should IMHO now be merged into a test that does not refer
  to 5005 threads

- README.irix should be verified if all comments are still valid if 5.005
  threads are gone
- same for README.tru64

- ext/B/t/debug.t #52 defines is_thread only based on use5005threads. I don't
  dare touch this
- ext/B/t/showlex.t #30 same

- ext/Safe/safe2.t #66 Safe to remove the condition?

- lib/ needs a shakedown. Removal ain't that hard, but it's the docs

Side notes:

- win32/Makefile still takes perlio in the archname. Now that perlio is
  default and non-perlio is very likely to be unsupported/depricated, this is
  not the most logical way
- win32/ same story

H.Merijn Brand        Amsterdam Perl Mongers (
using perl-5.6.1, 5.8.0, & 5.9.x, and 806 on  HP-UX 10.20 & 11.00, 11i,
   AIX 4.3, SuSE 8.2, and Win2k. 
send smoke reports to:, QA:

Thread Previous | Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About