Steve Peters wrote:
This failure started after yum decided that I needed bind on my system
to run some other piece of software (hate!).
mmm, yum. (sorry)
Since then, the smokes started
failing. Apparently, there are some differences between the structures
used for the re-entrant functions
Andy Lester wrote:
These consts are getting tedious, but dammit, we're going to have better,
more solid code as a result.
I wish I could go through file by file, but it's really based on what
has the most dependencies and which files clear up the most compiler
warnings.
Thanks, applied as
Alexey Tourbin wrote:
On Fri, Mar 18, 2005 at 11:19:30PM +0100, Rafael Garcia-Suarez wrote:
The simplest patch is generally wrong as it is Linux-specific; strictly
speaking, it is glibc-specific, when dynamic linking against libpthread
is used. Is there a better way to fix
Ton Hospel wrote:
This patch allows len/format to work for any format type, not just strings.
Basically for the non-string types it tries to gobble up all arguments
that are left, unless you give an explicit repeat count. It also makes
the * optional in the format.
It's not paranoid about
[EMAIL PROTECTED] wrote:
The patch below against latest bleadperl is a quick and dirty first
attempt at such a fix, which passes all tests (including the new one) here;
a cleaner attempt (feel free, that's all I have time for today) would hoick
the calculations of o and b up a level to avoid
Alexey Tourbin wrote:
On Sat, Mar 19, 2005 at 04:37:12PM +0300, Alexey Tourbin wrote:
The problem is that $in varialbe is not reset and if conditons hold
true till the end of file. Here is a patch (which has not been tested
yet, but seems to be reasonable offhand).
Here is a test that
Ton Hospel wrote:
Currently unpack(A) strips \0 and all classic whitespace
from the end of the string. Now that pack/unpack are encoding
neutral, the question arises what whitespace is in case of
unicode. Unicode has a lot more whitespace than the classic set
recognized by C isspace(3).
On Mon, 21 Mar 2005 22:29:09 +0100, demerphq [EMAIL PROTECTED] wrote:
On Mon, 21 Mar 2005 11:57:23 +, Nicholas Clark [EMAIL PROTECTED] wrote:
The regexp trie code issues some interesting looking warnings on VMS, and
also fails due to some char/unsigned char pointer issues. I think I could
On Mon, 21 Mar 2005 21:31:37 + (UTC), Ton Hospel
[EMAIL PROTECTED] wrote:
In article [EMAIL PROTECTED],
Rafael Garcia-Suarez [EMAIL PROTECTED] writes:
I'd prefer 3, because it's the more conservative. However, having to
choose between 1 and 2, I think 1 is more right, since
On 19 Mar 2005 17:59:23 -, via RT Alexey Tourbin
[EMAIL PROTECTED] wrote:
h2ph is unable to parse the following construct (found in Linux
kernel headers):
This bug is already fixed in blead.
On Sun, 20 Mar 2005 14:00:53 +0100, Tels [EMAIL PROTECTED] wrote:
the attached patch syncs blead with v1.75. New in this version:
Thanks, applied as change 24048.
On Sat, 19 Mar 2005 03:53:30 +0300, Alexey Tourbin [EMAIL PROTECTED] wrote:
perlrun says:
If you're just trying to get a print out of each line of Perl code
as it executes, the way that sh -x provides for shell scripts, you
can't use Perl's -D switch. Instead do this
I
Alexey Tourbin wrote:
Hello,
I noticed that B/Disassembler.pm sets its version in wrong package.
--- perl-5.9.2.24039/ext/B/B/Disassembler.pm- 2003-08-22 06:32:12 +
+++ perl-5.9.2.24039/ext/B/B/Disassembler.pm 2005-03-17 15:57:22 +
Thanks, applied to bleadperl as #24040.
Andy Lester wrote:
Attached is more constifying. Most of it this time is in utf8.c, like
all the is_utf8_alnum() etc funcs. I also hit grok_(bin|hex|oct).
Finally, I (void)ed some calls to SvREFCNTinc that were causing warnings
in -pedantic mode.
Thanks, applied as change 24042.
Gisle Aas wrote:
Seems like feedback on this has stopped, so for the convenience of the
pumpkin here is the combined patch. We (ActiveState) want to add this
to future versions of ActivePerl and it would less risky for us if
bleadperl could reserve the -f switch for this purpose by adopting
demerphq wrote:
Also, i discovered I omitted some new test files, and have a small
patch to apply on top of my previous. Should I prepare a new bundle
overall or just an addendum patch?
Ok, here is Rc8_trie.patch, with the extra tests, documentation,
changes to re.pm, and
On Sat, 19 Mar 2005 01:12:39 +0300, Alexey Tourbin [EMAIL PROTECTED] wrote:
The simplest patch is generally wrong as it is Linux-specific; strictly
speaking, it is glibc-specific, when dynamic linking against libpthread
is used. Is there a better way to fix this?
Yes, use a hint file, as many
I have the impression that t/uni/class.t became much slower with the
integration of the trie patch. I've no time to investigate further
now, but could someone take a look at this and confirm / infirm this
impression ?
On 17 Mar 2005 20:49:16 -, via RT Matthew Harris
[EMAIL PROTECTED] wrote:
Why does the following command not print 5,6,7?
perl -e '$,=,; print grep { $_ 4 .. $_ 7 } 1..10'
Because every value 4 is the beginning of a range.
You want :
$ perl -e '$,=,;print grep{$_==5 .. $_==7}1..10'
Sargent, John wrote:
In op.c, line 3856 (ish) the cast to a LOOP* is incorrect. in that the
memory allocated inside Perl_convert when the OP type is NOT OP_LIST is
I can't figure out a case where this wouldn't be a list op.
Have you code to reproduce this ?
less than the size of a LOOP
David Nicol wrote:
okay Dan, how does this grab you?
--- perlfunc.podTue Mar 8 12:23:37 2005
+++ perlfunc.pod.newTue Mar 8 18:22:41 2005
@@ -2273,7 +2273,8 @@
Hex strings may only represent integers. Strings that would cause
integer overflow trigger a warning. Leading
Andy Lester wrote:
I've spent the last couple of days going through internals, looking for
likely candidates for const qualifiers. Functions like av_len() are
ideal: They don't modify their parameter, so the compiler/lint will
(should) know a little bit more about the value that gets passed
On Mon, 14 Mar 2005 15:14:17 -0500, Horsley, Tom [EMAIL PROTECTED] wrote:
I wish people would quit spreading the myth that const helps
compilers do any kind of optimizations.
Well, my opinion is that const helps mostly humans, as every C++
programmer knows; and the perl 5 code could use some
On Sat, 12 Mar 2005 22:48:05 -0600, Andy Lester [EMAIL PROTECTED] wrote:
As it is now, if you do -DPERL_GCC_PEDANTIC, the pedantic options double
define PERL_GCC_BRACE_GROUPS_FORBIDDEN, and builds give:
In file included from toke.c:24:
perl.h:207:1: warning:
On 13 Mar 2005 08:48:05 -, via RT Jarkko Hietaniemi
[EMAIL PROTECTED] wrote:
I definitely don't think these are high-priority patches but I thought
they might help to clarify the rules. I better admit that the {} rule
bit me the other day, even after all these years :-)
Thanks, applied as
I applied the patch below to perldelta. Comments welcome.
Change 24035 by [EMAIL PROTECTED] on 2005/03/13 21:14:36
Document pack changes in perldelta
Affected files ...
... //depot/perl/pod/perl592delta.pod#3 edit
Differences ...
//depot/perl/pod/perl592delta.pod#3 (text)
On Fri, 11 Mar 2005 15:56:11 -0800, Michael G Schwern [EMAIL PROTECTED] wrote:
Is extended global? Lexical?
That's precisely why I want to add the possibility to write lexical
pragmas: people could write things like extended and put them on CPAN
without messing with the core.
OTOH if the
Thanks, indent patch applied as 24030 and blead test patch as 24031.
(leaving the maint patch for Nicholas.)
On Sat, 12 Mar 2005 18:54:29 + (UTC), Ton Hospel wrote:
A missing test to blead for u too wide:
--- perl-dev/t/op/pack.tSat Jan 22 01:21:58 2005
+++ perl-dev2/t/op/pack.t
On Sat, 12 Mar 2005 23:35:53 +0100, demerphq wrote:
Currently in order to debug regexes you have to load the debug module
and then review copious debug statements written directly to STDERR.
My trie patch includes some changes to control which output is emitted
Currently the re-debug mechanism
On Sat, 12 Mar 2005 17:23:18 -, Paul Marquess wrote:
Minor change to fix a DBM Filter issue.
Thanks, applied to blead as 24028.
Michael G Schwern wrote:
On Thu, Mar 10, 2005 at 11:26:25PM +, Nigel Sandever wrote:
The backwards compatibility argument is bunk: we can do exactly what was
done
for err() and lock(). Use a weak keyword. Any existing routine called
say()
will continue to work as before.
I
Nicholas Clark wrote:
New features are possible. I'd quite like to be able to bind in Perl 5.
Wasn't there a proposal for :
my \$a = \$b;
although I don't know what you call binding exactly.
[snip words of wisdom that follow...]
Andy Lester wrote:
On Thu, Mar 10, 2005 at 03:07:07PM -0500, Andrew Dougherty ([EMAIL
PROTECTED]) wrote:
The problem is the declaration of a new variable on line 1589:
AV* av = newAV();
Here we go. Third time should be the charm.
Thanks, applied as 24023.
Nigel Sandever wrote:
This is patched against 5.8.6, but I thought if I sent it in, someone might
look
it over and tell me
1) Would the idea ever be likely to be accepted?
No. Adding new keywords breaks backwards compatibility. So you must have
a very good reason to do so, and I don't
Nicholas Clark wrote:
I wonder how hard it would be to fix the limitation in the perl core noted
in Perl6::Say:
http://search.cpan.org/~dconway/Perl6-Say-0.02/Say.pm#BUGS_AND_IRRITATIONS
Quite hard, I expect. The indirect object notation is already the origin
of lots of special cases in the
Robin Barker wrote:
perl -MPerl6::Say -we 'say Hello World; say STDERR Warning'
Clever, but :
say { $fh } willow
still doesn't work.
Robert wrote:
Andy Lester wrote:
Attached is a patch of my day of playing with adding const qualifiers to
many underlying C functions.
This changes the ABI, so already compiled XS modules may no longer function.
We already broke binary compatibility.
However, already written XS modules
demerphq wrote:
Hi,
Ive been having a lot of annoynances related to trailing whitespace in
the .h and .c files in the perl distro.
What kind of annoyances ?
Im wondering if this could be dealt with by those with commit rights
ensuring that they dont check in code with trailing whitespace.
Brendan O'Dea wrote:
Attached are the patches to Debian's 5.8.4 package not currently covered
by my latest rsync of perl-5.8.x. The patches have been re-worked to
apply cleanly to that branch, but all should be applicable to
perl-current as well.
00_fix_instmodsh_doc
Already done in
Steve Hay wrote:
op/pack.t...FAILED 14028-14604
This is presumably caused by 24010.
Please check whether 24011 fix it, or part of it.
On Wed, 9 Mar 2005 21:18:04 + (UTC), Ton Hospel
[EMAIL PROTECTED] wrote:
Please check whether 24011 fix it, or part of it.
Huh, funny how I consistently used two different variables named end
without noticing even if they were right next to each other :-)
Anyways, I see that the new
Ton Hospel wrote:
Here is a new patchset relative to bleed.
This one should basically be the final version unless people disagree or
find bugs.
Thanks, applied as change #24010. (plus a few very minor typos in the docs)
I think I'll also rip parts of your patch description for perldelta. If
On Tue, 8 Mar 2005 15:46:37 -0500 (EST), kj [EMAIL PROTECTED] wrote:
Is there a howto on writing compiler backends? (The documentation for
B is of little help.)
Look at http://www.faqs.org/docs/perl5int/compiler.html
Absent that, if someone can point me to a backend module that is
Here are the slides from the talk on Perl 5.10 I gave last week-end at
FOSDEM in Brussels :
http://people.mandrakesoft.com/~rgarciasuarez/slides/perl510/perl510.html
It notably outlines the TO-DOs and the plans for the next 5.9.x
releases.
Yuval Kojman (via RT) wrote:
Given
TEST
use strict;
use warnings;
use Test::More tests = 5;
use_ok('Tie::RefHash');
tie my %hash, 'Tie::RefHash';
$hash{blah} = poot;
is($hash{blah}, poot, stored properly with scalar key);
is(delete($hash{blah}), poot, delete returns value with
demerphq wrote:
A trie is a way of storing keys in a tree structure where the
branching logic is determined by the value of the digits of the key.
Ie: if we have car, cart, carp, call, cull and cars we can
build a trie like this:
c + a + r + t
| | |
| | + p
| |
Nicholas Clark wrote:
On Wed, Feb 23, 2005 at 11:52:12PM -0800, Yitzchak Scott-Thoennes wrote:
On Wed, Feb 23, 2005 at 01:14:01PM -, Nils wrote:
Switch should be fixed or removed from Core. It is a serious liability.
Umm, since removing it would also break programs that say use
[EMAIL PROTECTED] (via RT) wrote:
It looks like dynamic scoping of $digit is not working properly when
recursion is used.
Consider:
sub b {
print B $1\n;
$_[0] =~ /(l+)/;
b(aocate) if $_[0] eq hello;
print E $1\n;
}
b(hello);
This prints
B
B ll
E
E
demerphq wrote:
Some of Abigails japhs dont work on win32 or depend on fixed bugs.
This patch skips the relevent tests as per the file convention already
established.
Thanks, applied as #23987.
Craig A. Berry wrote:
The bleadperl build currently fails using Compaq C V6.5-001 on OpenVMS
Alpha V7.3-1 because there are lots of signed/unsigned mismatch warnings
in pp_pack.c. The attached patch adds a bunch of casts to make explicit
the assumption that as long as we shovel 8 bits at a
Dintelmann, Peter wrote:
In Perl 5.6.1 we have
$ perl561 -MPOSIX -e 'require unistd.ph'
Are your unistd.ph files the same ?
can you reproduce this without POSIX and/or unistd.ph ?
Nicholas Clark wrote:
Currently IPC::Run fails tests with PERL_UNICODE set and a UTF-8 locale.
The appended patch is one way to fix this. I'm not sure if it's the best
way, so didn't apply it. Thoughts?
Thanks, applied as #23989 to blead.
Quoting the -I../.. was confusing the spawned perl
On Sat, 19 Feb 2005 00:27:57 +, Nicholas Clark [EMAIL PROTECTED] wrote:
Currently IPC::Run fails tests with PERL_UNICODE set and a UTF-8 locale.
The appended patch is one way to fix this. I'm not sure if it's the best
way, so didn't apply it. Thoughts?
This looks good to me, except that
On Sun, 20 Feb 2005 22:47:33 +, Nicholas Clark [EMAIL PROTECTED] wrote:
If it's VMS that's going to be the pain, don't I actually want
push @perl, q(-I../..) if $ENV{PERL_CORE};
to stop it changing the case of the I?
er, right, sorry :)
[EMAIL PROTECTED] wrote:
The patch below adds processing of the -path, -ipath and -iname
options to find2perl (including with Pod change). Also, in the
x2p Makefile some dependencies was missing and will be added by the
patch.
Thanks, applied as 23978 and 23979.
All is in the subject :)
The goal is to make it lexical one day.
Enjoy. And thanks to Autrijus.
Gisle Aas wrote:
A better patch would probably be:
--- ext/IO/lib/IO/Socket/INET.pm.cur2005-02-03 12:09:58.774436606 +0100
+++ ext/IO/lib/IO/Socket/INET.pm2005-02-03 12:10:16.011910233 +0100
@@ -132,7 +132,7 @@
$proto ||= (getprotobyname('tcp'))[2];
my $pname =
demerphq wrote:
On Thu, 17 Feb 2005 19:44:19 +0100, Rafael Garcia-Suarez
[EMAIL PROTECTED] wrote:
I've upgraded bleadperl to PathTools 3.04
Cwd.pm fails with the following results:
not ok 27 - abs_path() works on files in the root directory
# Failed test (cwd.t at line 210
Stephen McCamant wrote:
It appears there were two problems with walkoptree and pmreplroot:
first, there was a logic error in the previous addition of pmreplroot
support: the code was claiming that the pmreplroot kid of the OP was a
always PMOP, while in general it isn't: it's only the *parent*
demerphq wrote:
The win32 Makefiles weren't modified when corelist was added causing
complaints during install.
Attached patch fixes the problem.
Good catch. Actually I applied this, adding corelist to the clean target as
well :
Change 23984 by [EMAIL PROTECTED] on 2005/02/18 16:07:54
Nikolai Eipel wrote:
Hi,
I think that the problem is situated in the pp_sys.c file. When the
function
PP(pp_fttext) is used for a read protected file this function setted the
variable PL_Laststatval to -1. Than when the function tries to open the not
readable file we get a warning and
I've upgraded bleadperl to PathTools 3.04, but I've applied the following
patches on top of it :
The first one allows to run Cwd from miniperl
--- PathTools-3.04/Cwd.pm 2005-02-07 01:26:55.0 +0100
+++ bleadperl/lib/Cwd.pm2005-02-17 19:24:19.584093712 +0100
@@ -199,7 +199,7
Yitzchak Scott-Thoennes wrote:
Rats, I somehow overlooked a test failure. Please add this:
Thanks, applied as #23973.
--- perl/ext/B/t/stash.t.orig 2005-01-07 06:57:11.0 -0800
+++ perl/ext/B/t/stash.t 2005-02-15 20:08:59.063289600 -0800
@@ -61,7 +61,7 @@
@got = grep { !
Steve Hay wrote:
If IPC::Open3 does work then I guess CPANPLUS should use it, and we
don't need IPC::Run in the core.
If not then either someone needs to look at Nicholas' suggestions for
improving pipes on Win32 so that IPC::Run's requirement for Win32::
modules can be avoided, or
Justin Mason wrote:
FWIW, this looks like it'd be excellent for SpamAssassin ;)
I haven't had much time to look over the implementation, and I'm not
really any use for reviewing it from a p5p POV due to lack of familiarity
with perl internals, but the benchmark figures look fantastic and the
Jos I. Boumans wrote:
Basically, if we can either get:
a) pipes to work 'somehow' on win32 (that being ipc::run or a fix on
ipc::open3 or teaching IPC::Cmd to use IPC::Run3).
I don't know whether we can hope to get a reliable version of this in
the near future, so give me hints, windows
[re cc'ing p5p]
demerphq wrote:
On Wed, 16 Feb 2005 16:30:47 +0100, Rafael Garcia-Suarez
I was waiting for Hugo's advice on this patch.
Also, it would be nice to have something that works with /i, if that can
be done reliably.
Its done. I hope reliably too, but there is this annoying
demerphq wrote:
[re cc'ing p5p]
Dangit, sorry about that. Still getting used to writing mails in gmail.
Damn automatic gmail reply-to, as well.
At least on Linux, all tests pass.
Is that the second patch i posted or the first Raphael?
(trie_5.patch.gz is the latest)
The second.
Michael G Schwern wrote:
- Forwarded message from Alfie John [EMAIL PROTECTED] -
From: Alfie John [EMAIL PROTECTED]
Date: Tue, 15 Feb 2005 21:46:05 +1100
To: 'Michael G Schwern' [EMAIL PROTECTED]
Subject: RE: Symbol::delete_package() bug
X-Mailer: Microsoft Outlook, Build 10.0.2616
David Dyck (via RT) wrote:
The example given in perldoc -f getsockopt:
...
is incorrect, it should declare the constant it intends to import
I applied the fix below :
//depot/perl/pod/perlfunc.pod#453 (text)
@@ -2115,13 +2115,13 @@
An example testing if Nagle's algorithm is
[EMAIL PROTECTED] (via RT) wrote:
The each function loops infinitely on anonymous hashes.
This example:
while( my ($k, $v) = each %{{qw(one 1 two 2 three 3)}} ){
print $k $v\n;
}
prints three 3 forever.
That's not a bug, the hash is recreated each time the while
condition is
Ton Hospel wrote in perl.perl5.porters :
Here is a redo of the unpack part of the patches as a result of the
discussion here. Things of note:
- There is a new unpack format W (wide char), an encoding neutral form
of C that's also able to handle char values = 256
- All formats are
Reini Urban wrote in perl.perl5.porters :
Great! Now we only have to provide better place for the docs.
See the attached patch for pods/perlcygwin.pod
How about that?
Thanks, applied to bleadperl as change #23962.
Sherm Pendley wrote in perl.perl5.porters :
Mac OS X no longer uses file type info to launch apps in a MacPerl
application. Updated, and added a reference to DropScript.
Thanks, all doc patches applied to bleadperl as change #23963,
sometimes with minor rectifications (since your wording seems
Tels wrote:
the attached patch should cure the problem and the smokes - finally. Steve
Hay was so kind to test it and it really seems to work now. Sorry that it
took so long,
Thanks, applied as change #23955.
Dave Rolsky wrote:
Still no fix for the AIX bug yet ...
1.112005-02-09
- Try to make detection of supported epoch range a little smarter.
The detection was allowing negative epochs on Win32 but apparently
this doesn't work, and trying to pass a pre-epoch date in just causes
a lot of
[EMAIL PROTECTED] (via RT) wrote:
perl -e 'not() || 1'
perl -e 'die unless not()'
...
Like in the two examples above, the evaluation of not()
produces a segmentation fault or a freeze of the interpreter.
It occurs in bleadperl as well; but that's not the evaluation
of not() that segfault,
Jos I. Boumans wrote:
On Jan 25, 2005, at 12:26 PM, Jos I. Boumans wrote:
Although things don't quite work on vms (yet), here is the patch to
integrate IPC::Run 0.79 with core perl. This allows us to at least
test it widely on other systems as well and is a prereq for CPANPLUS
to
H.Merijn Brand wrote:
Why not Cwd and abspath?
Good idea. So I applied Jos' patch (thanks) as change 23954 to
bleadperl, but I changed every test preamble to this :
BEGIN {
if( $ENV{PERL_CORE} ) {
chdir '../lib/IPC/Run' if -d '../lib/IPC/Run';
unshift @INC,
王晓哲 wrote:
When using split() function in (??{...}), the Perl interpreter died.
This error can be verified by the following program:
#!/usr/bin/perl
$a=aba;
print matched\n if $a=~/(??{aba})/; # OK
print matched\n if $a=~/(??{join(,split(,aba))})/; # Segmentation
Fault
The regexp
I've added Module::CoreList to bleadperl as change 23947.
I also made the following minor change, please integrate in the CPAN version :
Change 23948 by [EMAIL PROTECTED] on 2005/02/08 10:39:21
Fix typo in corelist
Affected files ...
... //depot/perl/lib/Module/CoreList/bin/corelist#2
Steve Hay wrote:
The attached patch (against Time-Local-1.10) fixes the example above,
and also silences a bucketload of similar noise that was previously
coming from Time::Local when running the LWP test suite on Win32.
Added in Time-Local rt.cpan queue, closing it here.
Alfie John wrote:
Hi,
In the Symbol.pm BUGS section, it specifies that
Symbol::delete_package() undefines every symbol that lives in the
specified package and in its sub-packages. I have found that this is
not the case. While it does undefine symbols and sub-package names with
the
A coworker found this oddity.
The object created by the closure returned by ok() is destroyed soon,
the one created by the closure returned by bad() is destroyed during
the END phase. I can't figure out why...
sub Oo::DESTROY {
warn DESTROY\n;
}
sub ok {
my ($b, $w,
[EMAIL PROTECTED] (via RT) wrote:
perl -le 'print unpack(H*, pack(P, undef))'
perl -le 'print unpack(H*, pack(P, $a))'
28391608
After the proposed patch:
perl -le 'print unpack(H*, pack(P, undef))'
perl -le 'print unpack(H*, pack(P, $a))'
Thanks, applied as
Orton, Yves wrote:
On Feb 8, 2005, at 12:15 PM, Rafael Garcia-Suarez wrote:
I also made the following minor change, please integrate in
the CPAN
version :
Change 23948 by [EMAIL PROTECTED] on 2005/02/08 10:39:21
Fix typo in corelist
Thanks, applied as 11945
Lukas Mai wrote in perl.perl5.porters :
#!/usr/local/bin/perl
use warnings;
use strict;
$\ = \n;
for $_ (qw/foo bar baz/)
{ print }
for $_ qw/you don't see this/
{ print }
__END__
First surprise: you don't need parens if you use qw() (maybe because it's
turned into (you, don't,
Sisyphus wrote:
On perl 5.8.6, however, with both Win32 and linux, \(0..127) becomes an
array of 128 scalar references.
Is that the way it should be ?
Yes, it's documented in perlref :
Taking a reference to an enumerated list is not the same
as using square brackets--instead it's
Ben Tilly (via RT) wrote:
sub DESTROY {print OK\n}
bless [];
That's a known bug (ticket merged with #10030, see discussion there.)
Ton Hospel wrote:
I'd still like to make U0 and C0 mode into my reversed interpretation,
but now that unpack(C*) ignores the mode I think the impact will be
very low and worth doing to make pack and unpack consistent.
I've now applied the four bugfixes (23915, 23922, 23923, 23924) with
Andy Lester wrote:
On Mon, Jan 31, 2005 at 10:21:29PM -, rirans @ comcast. net ([EMAIL
PROTECTED]) wrote:
The localtime documentation is misleading. The header
should indicate both prototypes.
Fixed, plus some other cleanup on the text to make it easier for the
reader to pick out
H.Merijn Brand wrote:
I cannot let you have a peek at the files that make the base of Configure.
There's almost 900 of them, in a special depot, to which only few have access
It would be interesting to release them somewhere publicly (just in case
some fools could be interested :) be it by
KF (Lists) wrote:
PERLIO_DEBUG
If set to the name of a file or device then certain operations of PerlIO
sub-system will be logged to that file (opened as append). Typical uses
are UNIX:
...
else you can do the following either via sperl or vial sperl:
...
I fixed the local root exploit by
Rafael Garcia-Suarez wrote:
KF (Lists) wrote:
PERLIO_DEBUG
If set to the name of a file or device then certain operations of PerlIO
sub-system will be logged to that file (opened as append). Typical uses
are UNIX:
...
else you can do the following either via sperl or vial sperl
Rafael Garcia-Suarez wrote:
Rafael Garcia-Suarez wrote:
KF (Lists) wrote:
PERLIO_DEBUG
If set to the name of a file or device then certain operations of PerlIO
sub-system will be logged to that file (opened as append). Typical uses
are UNIX:
...
else you can do the following
KF (lists) wrote:
Much thanks on the quick response and fix.
Attached are the PoC for each condition incase you were wondering how
exploitable this *was*.
Do you have any idea what the span of this bug is (like what versions
are affected)?
I'd say every release since we have PerlIO, e.g.
Ton Hospel wrote:
I can trivially reverse the meaning for U0 and C0 in my patch of course,
and make the default starting mode U0. But it still wouldn't give you what you
want since C0 mode would still work on the (at least conceptually) upgraded
string, and it still wouldn't see through the
Nicholas Clark wrote:
Vendors are of course encouraged to grab those patches.
This seems to be the best way. We probably should wrap up all 3 patches into
one single this is what you need patch. Historically, I'm not sure if there
is a precedent on where anything like this has been
Ton Hospel wrote:
Ok, how about the following:
The backward compatibility problem probably mainly exists for
unpack(C*) which has traditionally been used to get to the
underlying bytes of an encoded string. In my patch C was a bit
special anyways (I made = 256 a croak to make it easy to
Ton Hospel wrote:
Take the following three statements:
1) What pack in current perl does for C0 is basically ok:
perl -wle 'use Devel::Peek; Dump(pack(C0U*, 8188))'
SV = PV(0x8162464) at 0x817547c
REFCNT = 1
FLAGS = (PADTMP,POK,pPOK)
PV = 0x816b750 \341\277\274\0
CUR = 3
LEN =
501 - 600 of 949 matches
Mail list logo