From vanbalen@mc.edu

There appears to be a bug in filesys I/O (under Linux) which breaks up
filenames with spaces in them.

My perl config (both Linux and HP-UX) is attached, along with the
malfunctioning prog (which seems to WORK JUST FINE under HP-UX). They are
also as follows, in case you don't like attachments​:

Summary of my perl5 (5.0 patchlevel 5 subversion 3) configuration​:
  Platform​:
  osname=linux, osvers=2.2.14-1mdksmp, archname=i386-linux
  uname='linux jedi.mandrakesoft.com 2.2.14-1mdksmp #1 smp thu dec 2 01​:02​:03 cet 1999 i686 unknown '
  hint=recommended, useposix=true, d_sigaction=define
  usethreads=undef useperlio=undef d_sfio=undef
  Compiler​:
  cc='cc', optimize='-O3 -fomit-frame-pointer -fno-exceptions -fno-rtti -pipe -s -mpentium -mcpu=pentium -march=pentium -ffast-math -fexpensive-optimizations', gccversion=2.95.2 19991024 (release)
  cppflags='-Dbool=char -DHAS_BOOL -I/usr/local/include'
  ccflags ='-Dbool=char -DHAS_BOOL -I/usr/local/include'
  stdchar='char', d_stdstdio=undef, 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='cc', ldflags =' -L/usr/local/lib'
  libpth=/usr/local/lib /lib /usr/lib
  libs=-lnsl -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'

Characteristics of this binary (from libperl)​:
  Built under linux
  Compiled at Dec 17 1999 17​:16​:45
  @​INC​:
  /usr/lib/perl5/5.00503/i386-linux
  /usr/lib/perl5/5.00503
  /usr/lib/perl5/site_perl/5.005/i386-linux
  /usr/lib/perl5/site_perl/5.005
  .

Malfunc-ing program​:

The behaviour I'm talking about can be obtained by running​:

prompt$ echo fred > "foo bar baz"
prompt$ perl frename.pl ".*baz" "baz" "barney"

which should result in the message "Error renaming baz​: No such file or
directory"

#! /usr/bin/perl
#Takes a min of 3 perl regexps​: 1) pattern to match 2) pattern to
#change in all files matching pattern1 and 3) pattern to replace
#pattern2 with. After the 3rd, all patterns are repetitions of
#pattern2 and pattern3. Also takes optional -s switch for stripping
#whitespace.

sub chmp_whitespace {
  $cpyfn=shift @​_;
  $file="";
  @​concat=split /\s+/, $cpyfn;
  foreach $baz (<@​concat>) {
  chomp $baz;
  print "$baz\n";
  $file="$file$baz";
  print "file $file\n";
  }
  return $file;
}

($#ARGV>-1) or die "You must provide cmd line args!";

$regexp = shift @​ARGV;

#if -s switch is present, strip $fns of whitespace
$strpspc=0;
if($regexp eq "-s") {
  $strpspc=1;
  $regexp=shift @​ARGV;
}

foreach $fn (<*>) {
  print "foreach $fn\n";
  #Strip whitespace
  if($strpspc) {
  $file=chmp_whitespace $fn;
  } else { $file=$fn }

  #proceed to replace regexps
  @​tmp=@​ARGV;
  while($foo=shift @​tmp) {
  $bar=shift @​tmp;
  if($file=/$regexp/ && $file=/$foo/) {
  $file=~s/$foo/$bar/g;
  rename("$fn", "$file") || print "Error renaming $fn​: $!\n";
  }
  }
}

Perl config under HP-UX​:

Summary of my perl5 (5.0 patchlevel 5 subversion 3) configuration​:
  Platform​:
  osname=hpux, osvers=11.00, archname=PA-RISC2.0
  uname='hp-ux ss b.11.00 u 9000800 680309313 unlimited-user license '
  hint=recommended, useposix=true, d_sigaction=define
  usethreads=undef useperlio=undef d_sfio=undef
  Compiler​:
  cc='cc', optimize='-O', gccversion=
  cppflags='-D_HPUX_SOURCE -Aa -I/usr/local/include'
  ccflags ='-D_HPUX_SOURCE -Aa -I/usr/local/include'
  stdchar='unsigned char', d_stdstdio=define, usevfork=false
  intsize=4, longsize=4, ptrsize=4, doublesize=8
  d_longlong=undef, longlongsize=, d_longdbl=define, longdblsize=16
  alignbytes=8, usemymalloc=y, prototype=define
  Linker and Libraries​:
  ld='ld', ldflags =' -L/usr/local/lib'
  libpth=/usr/local/lib /lib/pa1.1 /lib /usr/lib /usr/ccs/lib
  libs=-lnsl -lnm -lndbm -ldld -lm -lc -lndir -lcrypt
  libc=/lib/libc.sl, so=sl, useshrplib=false, libperl=libperl.a
  Dynamic Linking​:
  dlsrc=dl_hpux.xs, dlext=sl, d_dlsymun=undef, ccdlflags='-Wl,-E -Wl,-B,deferred '
  cccdlflags='+z', lddlflags='-b -L/usr/local/lib'

Characteristics of this binary (from libperl)​:
  Built under hpux
  Compiled at Aug 16 1999 18​:05​:55
  @​INC​:
  /opt/perl5/lib/5.00503/PA-RISC2.0
  /opt/perl5/lib/5.00503
  /opt/perl5/lib/site_perl/5.005/PA-RISC2.0
  /opt/perl5/lib/site_perl/5.005
  .

From vanbalen@mc.edu

Summary of my perl5 (5.0 patchlevel 5 subversion 3) configuration​:
  Platform​:
  osname=linux, osvers=2.2.14-1mdksmp, archname=i386-linux
  uname='linux jedi.mandrakesoft.com 2.2.14-1mdksmp #1 smp thu dec 2 01​:02​:03 cet 1999 i686 unknown '
  hint=recommended, useposix=true, d_sigaction=define
  usethreads=undef useperlio=undef d_sfio=undef
  Compiler​:
  cc='cc', optimize='-O3 -fomit-frame-pointer -fno-exceptions -fno-rtti -pipe -s -mpentium -mcpu=pentium -march=pentium -ffast-math -fexpensive-optimizations', gccversion=2.95.2 19991024 (release)
  cppflags='-Dbool=char -DHAS_BOOL -I/usr/local/include'
  ccflags ='-Dbool=char -DHAS_BOOL -I/usr/local/include'
  stdchar='char', d_stdstdio=undef, 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='cc', ldflags =' -L/usr/local/lib'
  libpth=/usr/local/lib /lib /usr/lib
  libs=-lnsl -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'

Characteristics of this binary (from libperl)​:
  Built under linux
  Compiled at Dec 17 1999 17​:16​:45
  @​INC​:
  /usr/lib/perl5/5.00503/i386-linux
  /usr/lib/perl5/5.00503
  /usr/lib/perl5/site_perl/5.005/i386-linux
  /usr/lib/perl5/site_perl/5.005
  .

From vanbalen@mc.edu

frename.pl

From vanbalen@mc.edu

Summary of my perl5 (5.0 patchlevel 5 subversion 3) configuration​:
  Platform​:
  osname=hpux, osvers=11.00, archname=PA-RISC2.0
  uname='hp-ux ss b.11.00 u 9000800 680309313 unlimited-user license '
  hint=recommended, useposix=true, d_sigaction=define
  usethreads=undef useperlio=undef d_sfio=undef
  Compiler​:
  cc='cc', optimize='-O', gccversion=
  cppflags='-D_HPUX_SOURCE -Aa -I/usr/local/include'
  ccflags ='-D_HPUX_SOURCE -Aa -I/usr/local/include'
  stdchar='unsigned char', d_stdstdio=define, usevfork=false
  intsize=4, longsize=4, ptrsize=4, doublesize=8
  d_longlong=undef, longlongsize=, d_longdbl=define, longdblsize=16
  alignbytes=8, usemymalloc=y, prototype=define
  Linker and Libraries​:
  ld='ld', ldflags =' -L/usr/local/lib'
  libpth=/usr/local/lib /lib/pa1.1 /lib /usr/lib /usr/ccs/lib
  libs=-lnsl -lnm -lndbm -ldld -lm -lc -lndir -lcrypt
  libc=/lib/libc.sl, so=sl, useshrplib=false, libperl=libperl.a
  Dynamic Linking​:
  dlsrc=dl_hpux.xs, dlext=sl, d_dlsymun=undef, ccdlflags='-Wl,-E -Wl,-B,deferred '
  cccdlflags='+z', lddlflags='-b -L/usr/local/lib'

Characteristics of this binary (from libperl)​:
  Built under hpux
  Compiled at Aug 16 1999 18​:05​:55
  @​INC​:
  /opt/perl5/lib/5.00503/PA-RISC2.0
  /opt/perl5/lib/5.00503
  /opt/perl5/lib/site_perl/5.005/PA-RISC2.0
  /opt/perl5/lib/site_perl/5.005
  .

From [Unknown Contact. See original ticket]

My perl config (both Linux and HP-UX) is attached, along with the
malfunctioning prog (which seems to WORK JUST FINE under HP-UX). They are
also as follows, in case you don't like attachments​:

That's right. Here's the story.

If you just use the standard built-in glob, then for compatibility
with Perl of Old, it will split on white space because it doesn't
know that you aren't asking for more than one fileglob.

  #!/bin/sh -x
  rm -rf /tmp/fred "/tmp/fred stuff"
  mkdir "/tmp/fred stuff"
  touch "/tmp/fred stuff/a"
  touch "/tmp/fred stuff/b"
  perl -We '
  @​a = </tmp/fred stu*>;
  print "Normal built-in globbed @​a\n"
  ';

That prints just "/tmp/fred", which isn't there. But that's
what you have to do for backwards compat.

The work around is simple. Just use File​::Glob "​:glob" in that
package, or even use File​::Glob "​:globally" to redefine the world​:

  #!/bin/sh -x
  rm -rf /tmp/fred "/tmp/fred stuff"
  mkdir "/tmp/fred stuff"
  touch "/tmp/fred stuff/a"
  touch "/tmp/fred stuff/b"
  perl -We '
  use File​::Glob "​:glob";
  @​a = </tmp/fred stu*>;
  print "File Glob globbed @​a\n"
  ';

Now it reports having globbed "/tmp/fred stuff", as you
would like.

Note that when you pull in :glob, or :globally for the
whole program, you get sh semantics from there on
for <*> and glob("*"). You need to add options, or
call csh_glob for the old stuff. But it still is white-space
clean.

Here are other examples​:

  use File​::Glob '​:glob';
  @​list = glob('*.[ch]'); # won't do tildes here anymore1
  $homedir = glob('~daemon', GLOB_TILDE | GLOB_ERR);
  if (GLOB_ERROR) {
  # an error occurred reading $homedir
  }

  use File​::Glob 'csh_glob';
  $homedir = csh_glob("~root"); # now you can
  print "home is $homedir\n";

  ## override the core glob
  use File​::Glob '​:globally';
  my @​sources = <*.{c,h,y}>;

  ## override the core glob, forcing case sensitivity
  use File​::Glob qw(​:globally :case);
  my @​sources = <*.{c,h,y}>;

  ## override the core glob forcing case *in*sensitivity
  use File​::Glob qw(​:globally :nocase);
  my @​sources = <*.{c,h,y}>; # gets *.C also!!

And here's the full thing, the default globbage in 5.6.

--tom

NAME
  File​::Glob - Perl extension for BSD glob routine

SYNOPSIS
  use File​::Glob '​:glob';
  @​list = glob('*.[ch]');
  $homedir = glob('~gnat', GLOB_TILDE | GLOB_ERR);
  if (GLOB_ERROR) {
  # an error occurred reading $homedir
  }

  ## override the core glob (even with -T)
  use File​::Glob '​:globally';
  my @​sources = <*.{c,h,y}>

  ## override the core glob, forcing case sensitivity
  use File​::Glob qw(​:globally :case);
  my @​sources = <*.{c,h,y}>

  ## override the core glob forcing case insensitivity
  use File​::Glob qw(​:globally :nocase);
  my @​sources = <*.{c,h,y}>

DESCRIPTION
  File​::Glob implements the FreeBSD glob(3) routine, which is a
  superset of the POSIX glob() (described in IEEE Std 1003.2
  "POSIX.2"). The glob() routine takes a mandatory `pattern'
  argument, and an optional `flags' argument, and returns a list
  of filenames matching the pattern, with interpretation of the
  pattern modified by the `flags' variable. The POSIX defined
  flags are​:

  `GLOB_ERR'
  Force glob() to return an error when it encounters a
  directory it cannot open or read. Ordinarily glob()
  continues to find matches.

  `GLOB_MARK'
  Each pathname that is a directory that matches the pattern
  has a slash appended.

  `GLOB_NOCASE'
  By default, file names are assumed to be case sensitive;
  this flag makes glob() treat case differences as not
  significant.

  `GLOB_NOCHECK'
  If the pattern does not match any pathname, then glob()
  returns a list consisting of only the pattern. If
  `GLOB_QUOTE' is set, its effect is present in the pattern
  returned.

  `GLOB_NOSORT'
  By default, the pathnames are sorted in ascending ASCII
  order; this flag prevents that sorting (speeding up
  glob()).

  The FreeBSD extensions to the POSIX standard are the following
  flags​:

  `GLOB_BRACE'
  Pre-process the string to expand `{pat,pat,...}' strings
  like csh(1). The pattern '{}' is left unexpanded for
  historical reasons (and csh(1) does the same thing to ease
  typing of find(1) patterns).

  `GLOB_NOMAGIC'
  Same as `GLOB_NOCHECK' but it only returns the pattern if
  it does not contain any of the special characters "*", "?"
  or "[". `NOMAGIC' is provided to simplify implementing the
  historic csh(1) globbing behaviour and should probably not
  be used anywhere else.

  `GLOB_QUOTE'
  Use the backslash ('\') character for quoting​: every
  occurrence of a backslash followed by a character in the
  pattern is replaced by that character, avoiding any
  special interpretation of the character. (But see below
  for exceptions on DOSISH systems).

  `GLOB_TILDE'
  Expand patterns that start with '~' to user name home
  directories.

  `GLOB_CSH'
  For convenience, `GLOB_CSH' is a synonym for `GLOB_BRACE |
  GLOB_NOMAGIC | GLOB_QUOTE | GLOB_TILDE'.

  The POSIX provided `GLOB_APPEND', `GLOB_DOOFFS', and the
  FreeBSD extensions `GLOB_ALTDIRFUNC', and `GLOB_MAGCHAR' flags
  have not been implemented in the Perl version because they
  involve more complex interaction with the underlying C
  structures.

DIAGNOSTICS
  glob() returns a list of matching paths, possibly zero length.
  If an error occurred, &File​::Glob​::GLOB_ERROR will be non-zero
  and `$!' will be set. &File​::Glob​::GLOB_ERROR is guaranteed to
  be zero if no error occurred, or one of the following values
  otherwise​:

  `GLOB_NOSPACE'
  An attempt to allocate memory failed.

  `GLOB_ABEND'
  The glob was stopped because an error was encountered.

  In the case where glob() has found some matching paths, but is
  interrupted by an error, glob() will return a list of
  filenames and set &File​::Glob​::ERROR.

  Note that glob() deviates from POSIX and FreeBSD glob(3)
  behaviour by not considering `ENOENT' and `ENOTDIR' as errors
  - glob() will continue processing despite those errors, unless
  the `GLOB_ERR' flag is set.

  Be aware that all filenames returned from File​::Glob are
  tainted.

NOTES
  * If you want to use multiple patterns, e.g. `glob "a* b*"', you
  should probably throw them in a set as in `glob "{a*,b*}'.
  This is because the argument to glob isn't subjected to
  parsing by the C shell. Remember that you can use a
  backslash to escape things.

  * On DOSISH systems, backslash is a valid directory separator
  character. In this case, use of backslash as a quoting
  character (via GLOB_QUOTE) interferes with the use of
  backslash as a directory separator. The best (simplest,
  most portable) solution is to use forward slashes for
  directory separators, and backslashes for quoting.
  However, this does not match "normal practice" on these
  systems. As a concession to user expectation, therefore,
  backslashes (under GLOB_QUOTE) only quote the glob
  metacharacters '[', ']', '{', '}', '-', '~', and backslash
  itself. All other backslashes are passed through
  unchanged.

  * Win32 users should use the real slash. If you really want to
  use backslashes, consider using Sarathy's File​::DosGlob,
  which comes with the standard Perl distribution.

AUTHOR
  The Perl interface was written by Nathan Torkington
  <gnat@​frii.com>, and is released under the artistic license.
  Further modifications were made by Greg Bacon
  <gbacon@​cs.uah.edu> and Gurusamy Sarathy
  <gsar@​activestate.com>. The C glob code has the following
  copyright​:

  Copyright (c) 1989, 1993 The Regents of the University of California.
  All rights reserved.
 
  This code is derived from software contributed to Berkeley by
  Guido van Rossum.

  Redistribution and use in source and binary forms, with or without
  modification, are permitted provided that the following conditions
  are met​:

  1. Redistributions of source code must retain the above copyright
  notice, this list of conditions and the following disclaimer.
  2. Redistributions in binary form must reproduce the above copyright
  notice, this list of conditions and the following disclaimer in the
  documentation and/or other materials provided with the distribution.
  3. Neither the name of the University nor the names of its contributors
  may be used to endorse or promote products derived from this software
  without specific prior written permission.

  THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  SUCH DAMAGE.

From [Unknown Contact. See original ticket]

Tom Christiansen <tchrist@​chthon.perl.com> wrote

If you just use the standard built-in glob, then for compatibility
with Perl of Old, it will split on white space because it doesn't
know that you aren't asking for more than one fileglob.

That's true about splitting on spaces *within the glob argument*.
But David's example

  foreach $fn (<*>) {

doesn't have have spaces in the glob argument, just in the filename.

Also I don't see the effect David reports. Trying

  touch "foo bar baz"

then in the Perl debugger I see

  DB<1> x <foo*>
0 'foo bar baz'
  DB<2>

A list of length one, not three.

This is with any Perl version up to 5.005_03.

So it seems that HP-UX, Solaris 2.6 and SunOS4 do the right thing here,
while Linux gets it wrong.

Something to do with the Linux shell?

Mike Guy

From [Unknown Contact. See original ticket]

Tom Christiansen <tchrist@​chthon.perl.com> wrote

If you just use the standard built-in glob, then for compatibility
with Perl of Old, it will split on white space because it doesn't
know that you aren't asking for more than one fileglob.

That's true about splitting on spaces *within the glob argument*.

But David's example
foreach $fn (<*>) {
doesn't have have spaces in the glob argument, just in the filename.

Oh, ok.

Also I don't see the effect David reports.

Neither do I.

So it seems that HP-UX, Solaris 2.6 and SunOS4 do the right thing here,
while Linux gets it wrong.

No, I can't reproduce it under "Linux" (whatever that means, since its
not the kernel, surely) either.

Something to do with the Linux shell?

There is no Linux shell. :-( But here is perl5.005_03 on RedHate 6.0​:

  DB<1> open(FH, ">foo bar blarch") or die;
  DB<2> system "ls", "-l", "foo bar blarch"
-rw-r--r-- 1 tchrist root 0 Mar 4 06​:58 foo bar blarch
  DB<3> @​got = <foo*>
  DB<4> x @​got
0 'foo bar blarch'

or directly​:

redhat% perl -e '@​got = <foo*>; print "got ", scalar @​got'
1

Or more tellingly...

redhat# strace -s 60 -f perl -e '@​got = <foo*>; print "got ", scalar @​got'

fork() = 29368
[pid 29367] close(4) = 0
[pid 29367] fcntl(3, F_GETFL) = 0 (flags O_RDONLY)
[pid 29367] fstat(3, {st_mode=S_IFIFO|0600, st_size=0, ...}) = 0
[pid 29367] mmap(0, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0
) = 0x2aabf000
[pid 29367] _llseek(3, 0, 0x7ffff3dc, SEEK_CUR) = -1 ESPIPE (Illegal seek)
[pid 29367] fcntl(3, F_SETFD, FD_CLOEXEC) = 0
[pid 29367] read(3, <unfinished ...>
[pid 29368] close(3) = 0
[pid 29368] dup2(4, 1) = 1
[pid 29368] close(4) = 0
[pid 29368] execve("/bin/sh", ["sh", "-c", "/bin/csh -cf \'set nonomatch; glob foo*\' 2>/dev/null"], [/* 35 vars */]) = 0

That gives me an idea!

redhat% touch "foo ' stuff ' can be a problem"
redhat% perl -le '@​got = <foo*>; print "got ", scalar @​got'
got 2
redhat% touch "foo \\\' stuff \\\\\' can be a problem"
redhat% ls -l foo*
-rw-r--r-- 1 tchrist tchrist 0 Mar 4 07​:15 foo ' stuff ' can be a problem
-rw-r--r-- 1 tchrist tchrist 0 Mar 4 07​:15 foo \' stuff \\' can be a problem
-rw-r--r-- 1 tchrist tchrist 0 Mar 4 07​:06 foo bar blarch

redhat% perl5.003 -le '@​got = <foo*>; print "got ", scalar @​got'
got 2

Wicked! you *can* screw it up. Well, with that release. But not
with the current one, using the spiffy new built-in BSD globbing​:

redhat% perl5.5.670 -le '@​got = <foo*>; print "got ", scalar @​got'
got 3

Interestingly, I cannot using the same Linux-freaking sequence (or
vice versa :-) confuse even older Perl releases (5.004) under
OpenBSD, even older OpenBSDs.

openbsd% ktrace -i perl5.00404 -le '@​got = <foo*>; print "got ", scalar @​got'
got 3

openbsd% kdump | grep foo
  "@​got = <foo*>; print "got ", scalar @​got
  "@​got = <foo*>; print "got ", scalar @​got
13014 csh NAMI "foo bar blarch"
13014 csh NAMI "foo ' stuff ' can be a problem"
13014 csh NAMI "foo \' stuff \\' can be a problem"
  "foo ' stuff ' can be a problem\0foo \\' stuff \\\\' can be a problem\0foo ba\
  "foo ' stuff ' can be a problem\0foo \\' stuff \\\\' can be a problem\0foo ba\
19417 perl5.00404 NAMI "foo ' stuff ' can be a problem"
19417 perl5.00404 NAMI "foo \' stuff \\' can be a problem"

I leave you to draw your own conclusions on these various matters.

--tom

From [Unknown Contact. See original ticket]

On Sat, 4 Mar 2000, M.J.T. Guy wrote​:

Tom Christiansen <tchrist@​chthon.perl.com> wrote

So it seems that HP-UX, Solaris 2.6 and SunOS4 do the right thing here,
while Linux gets it wrong.

Something to do with the Linux shell?

I'm running bash under Linux Mandrake 7.0-2 if that helps an (also
running bash on HP-UX). I'm pretty sure it's not the shell though cuz
globbing to shell commands works fine (i.e. rm *, etc).
I tried Tom's suggestion and neither interpreter (Linux or HP-UX) could
find Glob.pm (don't have enough experience to know how the heck to fix
that) and so I couldn't try that to see if it works.

I'll be outa town for a few days and check back on this Thursday.

Mike Guy

David van Balen mailto​: vanbalen@​mc.edu
Box 5054 vanbalen@​rocketmail.com
Clinton, MS 39058 http​://www.mc.edu/~vanbalen

From The RT System itself

This issue has been resolved in perl5.6 with the new internal glob stuff.