no problem with the
my_dir/ entry, so I don't know if this is an Archive::Zip bug or a Par bug.
At least zip -r -D avoids the problem.
Alan Stewart
In tkpp, line 548 should be:
open CONFIGFILE, +$configfile;
vice:
open CONFIGFILE, +$configfile;
or the prefs get wiped out second time around.
Alan Stewart
::* modules).
It does require Tk::JComboBox. . .
Alan Stewart
?
Alan Stewart
, leaving nothing behind in a all cleartemp environment?
Also, this would be a nice time to implement the default value of PAR_CLEARTEMP in
META.yml as you suggest in the notes in PAR.pm.
Alan Stewart
with the default?
Alan Stewart
? Does the runtime value in META.yml drive that or just reflect that.
Looks like a mis-spelling of PAR_INITIALIZED on line 121 of utils.c.
Alan Stewart
On 4 Jan 2004 at 13:47, Alan Stewart wrote:
PAR_CLEAN=1 overrides pp -C but PAR_CLEAN=0 doesn't override pp without -C. In
static.c/main.c you are looking for any non-zero length value for PAR_CLEAN.
Shouldn't
you be looking for the 0 or 1 value?
I said that backwards:
PAR_CLEAN=1
to be:
if ( !par_env_clean() (par_lstat(progname, PL_statbuf) == 0)) {
instead of:
if ( par_env_clean() (par_lstat(progname, PL_statbuf) == 0)) {
Alan Stewart
:
CODE(0x13a8f68)CODE(0x1428570) .)
at Tk/Event.pm line 3
END failed--call queue aborted.
Two runs one after the other is fine. Two simultaneous copies with clean is OK.
Alan Stewart
On 6 Jan 2004 at 17:09, Alan Stewart wrote:
I changed par_env_clean to !par_env_clean in mktmpdir.c and I get the right type of
dirs with the right type of clean/not clean, but now if I try to run two
simultaneous
copies of a Tk program without clean/with mtime dir, I get this error
On 6 Jan 2004 at 21:39, Alan Stewart wrote:
On 6 Jan 2004 at 19:44, Alan Stewart wrote:
Seems to be a Tk problem only. This small program has the same problem with the
second
copy trying to run:
use Tk;
my $mw = MainWindow-new();
MainLoop;
The cache directory does contain
$^X to command string only if the $ppfile is not pp.bat.
I've only tested on XP. Please try elsewhere. File tkpp.zip attached...
Alan Stewart
The following section of this message contains a file attachment
prepared for transmission using the Internet MIME message format.
If you are using
Clark
It is in ActiveState 5.6.1, but for mktmpdir.c you'll need C code MD5 subroutines.
There are a couple of examples around. One I found at:
http://www.cr0.net:8040/code/crypto/
which mentions SHA-1 (also there) as preferred to MD5 because of suspected collisions,
discussed elsewhere.
Alan
dir as pp by default - these also are getting left behind)
Don't know about those, but the PAR_CLEAN=1 might make them go away also. Hmmm.
Alan Stewart
I noticed that if you use -d with pp, then -C has no effect. Setting PAR_CLEAN=1
before execution works. There is bug in main.c. Line 139:
par_init_env();
is too late and needs to be moved back before the perl_parse, to line 131 or earlier.
Alan Stewart
On 11 Jan 2004 at 11:46, Autrijus Tang wrote:
在 日, 2004-01-11 06:36, Alan Stewart 寫道：
Here is a patched par.pl and a patched Par.pm.
Woohoo!
It will work if the .pm can be found in the local perl installation
from which perl.exe is running. Of course mtime is still there as fallback
{
if ($ENV{PAR_CLEAN}) {
unlink @tmpfile if @tmpfile;
rmdir $par_temp;
$par_temp =~ s{[^\\/]*$}{};
rmdir $par_temp;
}
}
Alan Stewart
P.S.
How about the modified tkpp with Clean and output window? Did that look OK?
as an aid to my eyes. Autrijus, could you share your
perltidy.rc if not default?
Alan Stewart
On 20 Jan 2004 at 17:46, Edward S. Peschko wrote:
hey all,
Attached is a modified Module::ScanDeps::DataFeed which makes
-c and -n work on pre-5.8.3 perl (it uses a somewhat convoluted but
back-portable way to get shared object filenames).
I didn't see an attachment ?!
Alan Stewart
and
install it manually. And your problem should go away :)
Alan Stewart
/Packer/Backend/PAR.pm $_); },
Was there supposed to be a change here or not?
Alan Stewart
install PAR
Nice site and up-to-date. If you don't want to download separately, a slightly
different page:
http://www.bribes.org/perl/ppmdir.html
gives you directions for letting ppm do the download and the install.
BTW, is Jean-Louis Morel a reader here ? If so, thanks !
Alan Stewart
recall details.
Alan Stewart
On 16 Feb 2004 at 7:31, Autrijus Tang wrote:
http://aut.dyndns.org/dist/Win32-Exe-0.04.tar.gz
MD5 (Win32-Exe-0.04.tar.gz) = b2f60472ee77f746c8a5d17dacc23a4e
I can make Win32-Exe OK, but I am failing all tests. What version/author of
Test::Simple/More/Builder are you using?
Alan
On 16 Feb 2004 at 15:16, Autrijus Tang wrote:
On Sun, Feb 15, 2004 at 09:35:02PM -0800, Alan Stewart wrote:
On 16 Feb 2004 at 7:31, Autrijus Tang wrote:
http://aut.dyndns.org/dist/Win32-Exe-0.04.tar.gz
MD5 (Win32-Exe-0.04.tar.gz) = b2f60472ee77f746c8a5d17dacc23a4e
I can make
FOR A PARTICULAR PURPOSE.
-Original Message-
From: Alan Stewart [mailto:[EMAIL PROTECTED]
Sent: Monday, February 16, 2004 10:27 PM
To: Jesse Schoch; [EMAIL PROTECTED]
Subject: Re: exec problems
On 16 Feb 2004 at 21:47, Jesse Schoch wrote:
Here is a simple test. Let me know if anyone
in progress and foo.exe isn't
creating it's own cache directory, so there are collisions when foo.exe unpacks. By
cleaning out the PAR_xx variables, I can get bar/foo to run with and without -C and -d.
On 16 Feb 2004 at 21:30, Alan Stewart wrote:
I think I see a clue. When I build bar and foo
and PAR_DEBUG getting to the second exe.
On 17 Feb 2004 at 0:27, Alan Stewart wrote:
I haven't figured out how to fix this the right way yet, but here is a work around
that
may work for you.
Put this:
for (keys %ENV) { delete $ENV{$_} if /^PAR_/ }
right before:
Win32::Process
On 18 Feb 2004 at 3:11, Autrijus Tang wrote:
On Tue, Feb 17, 2004 at 10:25:25AM -0800, Alan Stewart wrote:
The work-around I mentioned below is overkill and prevents having systems defaults
for
PAR_TEMP, PAR_CLEAN and PAR_DEBUG getting to the second exe.
Nice catch.
So all we have
to see if I am still in the loop with the list. Could be
my
ISP :(
Alan Stewart
for the next couple weeks, so I won't be doing anything soon.
There is always PAR 0.80_01 :)
Alan Stewart
. If the temp dir name generated automatically by
PAR
is OK with you, then you don't need $ENV{PAR_GLOBAL_TEMP} either. If you do specify
PAR_GLOBAL_TEMP, that overides any -C or PAR_GLOBAL_CLEAN anyway.
Alan Stewart
days
ago,
deleting all PAR_xxx. With the patch I sent in, you won't have to do that to keep
PAR_TEMP from foo carrying over to bar and messing it up. Nor do you need to restore
it. It has already done it's work.
See my other reply for the remaining gotcha.
Alan Stewart
That was intentionally not done based on the assumption that PAR itself does not
create
any subdirs when using -C. All the extraction is flat.
I know it was discussed. Autrijus, what was the reasoning for that again?
Alan Stewart
changes, so they don't accumulate.
Alan Stewart
\\ in your full path above vice \.
Alan Stewart
%PAR::LibCache) {
print $_\t$PAR::LibCache{$_}\n;
}
print $0 done\n;
-
Alan Stewart
the pp'd script is called t2.pl.
Alan Stewart
(as he's under RedHat).
OK. I assumed that it was a Windows OS, since he had an .exe file extension.
Alan Stewart
On 3 Mar 2004 at 14:07, Alan Stewart wrote:
Autrijus, I see the same fix in several places:
#if defined(__APPLE__) defined(__DYNAMIC__)
static char ** environ;
environ = *_NSGetEnviron();
#else
extern char ** environ;
#endif
I'm not a Mac user, so I can't check out the #ifdef values
On 3 Mar 2004 at 14:28, Alan Stewart wrote:
Here are the similar lines from the perl source:
#if defined(__APPLE__) defined(PERL_CORE)
# include crt_externs.h /* for the env array */
# define environ (*_NSGetEnviron())
#endif
which define environ rather than setting
, 1);
return;
}
Alan Stewart
for you and -C wouldn't (they are
one and the same option). But be aware that previously, setting PAR_TEMP=xxx (or
setting PAR_GLOBAL_TEMP=xxx now) overrides all clean options.
Try the latest snap and see if it helps.
Alan Stewart
pp.
Does that seem possible to you, or am I out of my mind?
Not mutually exclusive, given the state of my mind :)
Alan Stewart
OK.
Alan Stewart
On 7 Apr 2004 at 22:41, Rick Fitzsimmons wrote:
Hi,
Been having problems since updating to 0.80. One user gets error message
with Windows 2000 [Version 5.00.2195] when running a pp'd executable,
but another user has no trouble at all. After many attempts, I've also
seen this happen
.
and he patched it already. This is different from the extra long SHA1 name Rick is
seeing. In the patch, the Control-L becomes an underscore.
Alan Stewart
On 14 Apr 2004 at 15:27, the.noonings wrote:
On SUN machines, LD_LIBRARY_PATH is a system wide dynamic variable that PAR
dare not change. I don't know about Linux except that on my current Linux
machine it does not exist. In any case, I don't think any environment
setting could help because
there.
OTOH, static.c could possibly set LD_LIBRARY_PATH as long as that only affects the
environment for the PAR execution.
Alan Stewart
On 15 Apr 2004 at 9:10, Roderich Schupp wrote:
OTOH, static.c could possibly set LD_LIBRARY_PATH as long as that only affects the
environment for the PAR execution.
PAR does set LD_LIBRARY_PATH, the code is in myldr/mktmpdir.c.
You are right. In mktmpdir.c, there is code that prepends
into the ActivePerl html dir, but
PAR doesn't have them (yet).
I have no idea what could be different about the bribes compile. I am using VC 6.0 and
I think Autrijus is also.
Has anyone seen this problem with anything other than the bribes package??
Alan Stewart
it. Is
it something that we should have concern
over?
Do you mean the signature test, or the first test of basic.t? I've never installed the
signature modules, so for my Win32, signature is skipped, and all basic.t passes.
Alan Stewart
::SHA1 installed?
Of course, if you just ignore this error, you won't know if you are using the one true
PAR, or some insidious variation ;) Most of the current snapshots don't include a
SIGNATURE file.
Alan Stewart
both have for 5.8.3.
You may find that you have no problem running your application on a machine that
doesn't have perl installed (or at least not in c:\perl).
Or your problem may go away if you download the snapshot and compile yourself.
Alan Stewart
it break for you?
Alan Stewart
. This is on WinXP, but my limited testing suggests that it appears
on other versions as well.
Did the same with no fault here, build 809, XP, and the latest snap of PAR compiled
with VC 6. Are you self compiling PAR or using the binary download?
Alan Stewart
);
Alan Stewart
-o t9.exe t9.pl
The stat preceding -l _ wasn't an lstat at c:/ActivePerl/lib/File/Find.pm line 252.
Alan Stewart
Packer.pm has a typo on line 872:
$alias = s{^[a-zA-Z]:}{} if $^O eq 'MSWin32';
should be:
$alias =~ s{^[a-zA-Z]:}{} if $^O eq 'MSWin32';
Alan Stewart
\automated_pp_test.pl
at line 7563)
Autrijus has identified using -M to add an ordinary files as deprecated in favor of
-a. I would suggest removing these tests and just leaving the ones that use -M to add
modules and -a for ordinary files.
Alan Stewart
...
HTH,
Robert
There should be a new version of Module::ScanDeps that will eliminate the need for
this
extra use statement, coming real soon. Autrijus is on travel and p4.elixus.org has
been down, so I don't know when real soon is :(
Alan Stewart
Tk/DragDrop/Win32Drop.pm Tk/DragDrop/Win32Site.pm
Tk/DragDrop/XDNDDrop.pm Tk/DragDrop/XDNDSite.pm
) ],
at Module::ScanDeps.pm line 292.
See if that helps.
Alan Stewart
downloaded the SDK or the
additional include files, you hit some more missing stuff next (see my earlier post).
Alan Stewart
On 5 Jul 2004 at 10:47, Knut Virow wrote:
Alan Stewart wrote: (on 03.07.2004)
As mentioned elsewhere, you can download the precompiled packages from:
http://aut.dyndns.org/par/
and manually put it in the PAR-0.xx directory before making PAR.
But if you are Internet connected when you
of the binary is still automatic.
The readme doesn't really say that having a C compiler overrides having the pre-built.
We probably need an install query if both are present. Autrijus??
Alan Stewart
or
Windows native widget app.
Alan Stewart
people :)
However, the purpose of Obfuscate and Bytecode is not to process pod. Size reduction
(if any) is an unplanned side effect.
upx (upx.sourceforge.net) will shrink the size of the exe, but won't help speed it up,
unless you are fetching the exe from slow media.
Alan Stewart
. Under a default C compile,
this means [^a-zA-Z0-9], so non-English letters become underscores also. This is to
ensure a valid path name. The path preceeding par-$USER is not constructed, it pre-
exists, so it has or doesn't have spaces as you wish.
Alan Stewart
a
bit cleaner. No traces of PAR left, if all PAR exe's are -C.
That's not possible if some PAR's are cached, so you may have empty par-$USER dirs
left laying around.
Alan Stewart
the parl.pl frontend in memory, but
that
shouldn't be very large.
Still digging
Alan Stewart
, version 1.0, so if
you'd like to improve it, feel free.
Autrijus, for the contrib directory, if I may...
Alan Stewart
The following section of this message contains a file attachment
prepared for transmission using the Internet MIME message format.
If you are using Pegasus Mail, or any another MIME
On 18 Jul 2004 at 21:37, Autrijus Tang wrote:
On Sun, 18 Jul 2004 17:43:45 -0700, Alan Stewart [EMAIL PROTECTED] wrote:
I do a lot of little Tk apps with PAR, and although disk space is cheap, I
sometimes
have to send these by very slow means. My parsimonious soul doesn't like to send
extract a .par archive built from the same script (including base modules)
in under a second.
How does the first time startup compare to the speed if you set PAR_GLOBAL_CLEAN=1
before running the app, so it's not cached?
Alan Stewart
with these patches, if a user is only going to run your app once, you
should get a slight increase in startup with pp -C since inc/ is not created at all.
Alan Stewart
The following section of this message contains a file attachment
prepared for transmission using the Internet MIME message format
the ASCII subset that just happens
to match utf8, or something.
Alan Stewart
? I remember a discussion about how to re-instate the console a while ago.
Alan Stewart
attachment par.dat to par.zip.
Alan Stewart
The following section of this message contains a file attachment
prepared for transmission using the Internet MIME message format.
If you are using Pegasus Mail, or any another MIME-compliant system,
you should be able to save it or view it from within
-threaded apps.
I will continue to look into this a bit later tonight.
Second caveat: I am only guessing.
Good luck :)
Alan Stewart
to
recompile PAR to use it, snapshots and binaries become available again :(
Alan Stewart
and ran the exe or just
ran
wperl tk_stdio_demo.pl.
HTH
Alan Stewart
The following section of this message contains a file attachment
prepared for transmission using the Internet MIME message format.
If you are using Pegasus Mail, or any another MIME-compliant system,
you should be able to save
message.
Anyway, you can't rely on the constant being implemented as a sub, and as Rob said, it
is not neccessary.
Alan Stewart
?
Sure.
Alan Stewart
On 28 Jul 2004 at 14:00, Alan Stewart wrote:
I probably shouldn't have said broken, but not classic. PodStrip.pm is looking
specifically for the POD commands listed in the perlpod doc. The commands =header
and
=usage are not defined there. I'll look closer at what PodStrip does.
Well
On 28 Jul 2004 at 17:41, Glenn Linderman wrote:
So thinking about this some more, I guess I don't understand why the
cache-gibberish gibberish is so much longer than the filename gibberish.
Well, long is better, less chance of a duplicate. The cache dir is created by the
binary frontend
is
inheriting the file handles of the first. Beyond that I don't have a clue :(
Alan Stewart
for you.
Alan Stewart
is forgiving.
Alan Stewart
On 5 Aug 2004 at 19:35, Alan Stewart wrote:
It's not PAR. It's the C binary frontend compiled by VC++ 7.0 that links to
MSVCR70.dll. Bill wants you to upgrade all your systems to .NET where it will be
available. Either use -l or put it in the same dir as the exe or use VC++ 6.0.
BTW
On 22 Aug 2004 at 0:20, Pavel wrote:
Hello
I am trying to start perl script from PAR
script1.pl that runs without parameters and was converted to exe
script2.pl runs with parameters and should have the same environment
as script1.pl
I cannot chande or modify script2.pl so I need use
test.exe -l inside.pl outside.pl
and the output of test.exe is a dir list of c:\windows and c:\windows\system, so the
command interpreter is not the problem itself.
Is this anything like what you are trying to do? Are you trying to use pp -g on
script1.pl?
Alan Stewart
Perl modules that you are using in this app?
Alan Stewart
over after the crash somewhere.
-jess
-Original Message-
From: Alan Stewart [mailto:[EMAIL PROTECTED]
Sent: Friday, September 17, 2004 9:11 PM
To: Jesse Schoch; [EMAIL PROTECTED]
Subject: Re: Par locking, death, and purgatory
On 17 Sep 2004 at 12:31, Jesse Schoch wrote
for a gpp bug fix.
Alan Stewart
errors, so it silently gets what it
can.
Alan Stewart
On 19 Oct 2004 at 19:58, Edward Peschko wrote:
ok, then another question.. have you used vc6 to compile perl itself? For some
reason,
when I use vc6, ..\perl.h has a bunch of incorrect return codes (ie: it expects
'sys/time.h' which doesn't exist, and has 'long long', etc), and I looked
.
Robert
I saw your question on comp.lang.misc and interpreted the errors you got there
as a
possible icon problem, but that is not true here. I have had no problem with
the -d
option for any script I have. Can you give us a small example script that fails
for
you?
Alan Stewart
want obfuscation, you need to use a PAR obfuscation filter,
like:
pp -f Bleach
or:
pp -f Bytecode
By default, PAR packages can be read with any zip file utility. Even so,
obfuscation
only slows down the determined hacker. It won't stop him.
Alan Stewart
defaults. Do you need the original permissions or would a better
default or a pp option for the default permissions be acceptable?
BTW, my unzip (Info-ZIP 5.42) say that I need unzip -X to restore ACLs from
the zip
file when unzipping.
Alan Stewart
.
Robert
I saw your question on comp.lang.misc and interpreted the errors you got there
as a
possible icon problem, but that is not true here. I have had no problem with
the -d
option for any script I have. Can you give us a small example script that fails
for
you?
Alan Stewart
1 - 100 of 126 matches
Mail list logo
