Mailing-List: contact users-help@tomcat.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Tomcat Users List" <users@tomcat.apache.org>
Received-SPF: softfail (nike.apache.org: transitioning domain of aw@ice-sa.com
 does not designate 199.46.245.198 as permitted sender)
Message-Id: <1783560185.1286716248634@ace.smi>
Date: Sun, 10 Oct 2010 13:10:48 +0000
From: =?ISO-8859-1?Q?Andr=E9_Warnier?= <aw@ice-sa.com>
To: Tomcat Users List <users@tomcat.apache.org>
Subject: Re: Disable class monitoring for reloading container classes
Cc: "Caldarale, Charles R" <Chuck.Caldarale@unisys.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Importance: Normal

This email contained a .zip file attachment. Raytheon does not allow email =
attachments that are considered likely to contain malicious code. For your =
protection this attachment has been removed.

If this email is from an unknown source, please simply delete this email.

If this email was expected, and it is from a known sender, you may follow t=
he below suggested instructions to obtain these types of attachments.

+ Instruct the sender to enclose the file(s) in a ".zip" compressed file, a=
nd rename the ".zip" compressed file with a different extension, such as, "=
.rtnzip".  Password protecting the renamed ".zip" compressed file adds an a=
dditional layer of protection. When you receive the file, please rename it =
with the extension ".zip".

Additional instructions and options on how to receive these attachments can=
 be found at:

http://security.it.ray.com/antivirus/extensions.html
http://security.it.ray.com/news/2007/zipfiles.html

Should you have any questions or difficulty with these instructions, please=
 contact the Help Desk at 877.844.4712

---

Caldarale, Charles R wrote:
> From: Andr=E9 Warnier [mailto:aw@ice-sa.com]=20
> Subject: Re: Disable class monitoring for reloading container classes
>=20
>> thanks for the perl program.
>=20
> I would not introduce yet another variable into situation, since perl is =
not involved in any way with your Tomcat installation, and we don't know ho=
w a particular perl implementation might decide to handle timestamps.  I re=
commended C and Java programs since those underly Tomcat, and trying to pin=
point the exact location of the time discrepancy is likely crucial to eithe=
r fixing it or working around it.  Until the actual source of the problem i=
s determined, you're just throwing darts - blindfolded.
>=20
Charles, I would not dream of disagreeing with you on a Java matter, but in=
 this case, I will.
The perl interpreter is usually as close as you can get to the underlying s=
ystem's=20
libraries without resorting to C, and a perl script is an order of magnitud=
e easier to=20
handle than a C program.

What would be really nice, is if someone wrote a quick Java equivalent to t=
he perl script=20
I submitted.


Now if you *really* insist, the modified version of the perl program, below=
, will=20
explicitly use a couple of C functions, themselves using the builtin C stru=
ctures to get=20
the file's "last modified" time.

Running C in perl, scary stuff..

;-)


The interesting bit however, is when I run this program alternatively on my=
 Windows XP PC=20
and on a Linux server, both set for the same "Berlin" timezone (GMT+1, curr=
ent +1h of DST,=20
so GMT+2), I get different answers.

For a file in each case modified a few seconds ago, which would be around 1=
2:00 GMT,

a) Windows shows :

C:\Programs\local>perl showmodified.pl showmodified.pl
System raw timestamp : 1286712069
Current GMT time (no DST) : 2010/10/10-12:01:09
Current local time (with DST) : 2010/10/10-14:01:09
File [showmodified.pl]:
last modified timestamp : 1286712049
the file was last modified 20 seconds ago
last modified as GMT time (no DST) : 2010/10/10-12:00:49
last modified as local time (with DST) : 2010/10/10-14:00:49

b) Linux shows :

aw@colin:~$ perl showmodified_C.pl showmodified_C.pl
System raw timestamp : 1286712131
Current GMT time (no DST) : 2010/10/10-12:02:11
Current local time (with DST) : 2010/10/10-14:02:11
File [showmodified_C.pl]:
last modified timestamp : 1286708460
the file was last modified 3671 seconds ago
last modified as GMT time (no DST) : 2010/10/10-11:01:00
last modified as GMT time (no DST), as C would have it : 2010/10/10-11:01:0=
0
last modified as local time (with DST) : 2010/10/10-13:01:00
last modified as local time (with DST), as C would have it : 2010/10/10-13:=
01:00
aw@colin:~$


Even more surprising, for once it seems (to me) that Windows is actually mo=
re logical.
Or else, the underlying C libraries of Linux somehow adjust the values I'm =
getting.

Windows appears to record (and return) the last modified time of a file as =
GMT/UTC, while=20
Linux apparently records it as "local time", but without the DST adjustment=
.
Or is it recording it as GMT, plus the DST adjustment ?
(Unfortunately, the normal timezone here is GMT+1, and the DST adjustment r=
ight now is=20
also +1 h, so I do not know which is which.)
In any case, I get + one hour difference as compared to Windows, and what I=
 get is nonsense.
The file was not modified 3671 seconds ago, it was modified 71 seconds ago.

Unfortunately also, changing the system time or its timezone, on the Linux =
system on which=20
I am testing, is not an option for now.

Finding information on the web specifically about different OS'es practices=
 in terms of=20
recording "last modified" times is apparently not easy.

But I did find this rather old discussion of a similar issue, which seems t=
o say the=20
opposite of what I claim above :
http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=3D4860999

Can anyone smarter than I am throw more light on the subject ?

Which, to bring it back closer to topic, could be rephrased as : does a JVM=
 used by Tomcat=20
always return a "correct" (or at least comparable) last modified timestamp =
for an=20
application file, independently of the platform, time zones, and DST adjust=
ments ?


Perl script (also attached, zipped) :

Note : you may need for this to first run :
perl -MCPAN -e "install Inline::C"
(and answer yes to all the questions)

---- start copying here ---------
#!/usr/bin/perl
# showmodified_C.pl
use warnings;
use strict;
my ($sec,$min,$hour,$mday,$mon,$year);
my $subject =3D $ARGV[0] or die "missing argument";
die "File not found or not readable : $!" unless ((-f $subject) && (-r $sub=
ject));

my $now =3D time();
print "System raw timestamp : $now\n";
($sec,$min,$hour,$mday,$mon,$year) =3D gmtime($now);
print "Current GMT time (no DST) :=20
",sprintf("%04d/%02d/%02d-%02d:%02d:%02d",$year+1900,$mon+1,$mday,$hour,$mi=
n,$sec),"\n";
($sec,$min,$hour,$mday,$mon,$year) =3D localtime($now);
print "Current local time (with DST) :=20
",sprintf("%04d/%02d/%02d-%02d:%02d:%02d",$year+1900,$mon+1,$mday,$hour,$mi=
n,$sec),"\n";

print "File [$subject]:\n";

my ($dev,$ino,$mode,$nlink,$uid,$gid,$rdev,$size,
        $atime,$mtime,$ctime,$blksize,$blocks)
            =3D stat($subject);
print "last modified timestamp : $mtime\n";
print "the file was last modified ",$now-$mtime," seconds ago\n";

($sec,$min,$hour,$mday,$mon,$year) =3D gmtime($mtime);
print "last modified as GMT time (no DST) :=20
",sprintf("%04d/%02d/%02d-%02d:%02d:%02d",$year+1900,$mon+1,$mday,$hour,$mi=
n,$sec),"\n";
($year,$mon,$mday,$hour,$min,$sec) =3D C_gmtime($mtime);
print "last modified as GMT time (no DST), as C would have it :=20
",sprintf("%04d/%02d/%02d-%02d:%02d:%02d",$year+1900,$mon+1,$mday,$hour,$mi=
n,$sec),"\n";

($sec,$min,$hour,$mday,$mon,$year) =3D localtime($mtime);
print "last modified as local time (with DST) :=20
",sprintf("%04d/%02d/%02d-%02d:%02d:%02d",$year+1900,$mon+1,$mday,$hour,$mi=
n,$sec),"\n";
($year,$mon,$mday,$hour,$min,$sec) =3D C_localtime($mtime);
print "last modified as local time (with DST), as C would have it :=20
",sprintf("%04d/%02d/%02d-%02d:%02d:%02d",$year+1900,$mon+1,$mday,$hour,$mi=
n,$sec),"\n";


use Inline C =3D> <<'END_OF_C_CODE';

#include <time.h>

void C_localtime(int utc) {
   struct tm *ltime =3D localtime(&utc);
   Inline_Stack_Vars;

   Inline_Stack_Reset;
   Inline_Stack_Push(sv_2mortal(newSViv(ltime->tm_year)));
   Inline_Stack_Push(sv_2mortal(newSViv(ltime->tm_mon)));
   Inline_Stack_Push(sv_2mortal(newSViv(ltime->tm_mday)));
   Inline_Stack_Push(sv_2mortal(newSViv(ltime->tm_hour)));
   Inline_Stack_Push(sv_2mortal(newSViv(ltime->tm_min)));
   Inline_Stack_Push(sv_2mortal(newSViv(ltime->tm_sec)));
   Inline_Stack_Push(sv_2mortal(newSViv(ltime->tm_isdst)));
   Inline_Stack_Done;
}

void C_gmtime(int utc) {
   struct tm *ltime =3D gmtime(&utc);
   Inline_Stack_Vars;

   Inline_Stack_Reset;
   Inline_Stack_Push(sv_2mortal(newSViv(ltime->tm_year)));
   Inline_Stack_Push(sv_2mortal(newSViv(ltime->tm_mon)));
   Inline_Stack_Push(sv_2mortal(newSViv(ltime->tm_mday)));
   Inline_Stack_Push(sv_2mortal(newSViv(ltime->tm_hour)));
   Inline_Stack_Push(sv_2mortal(newSViv(ltime->tm_min)));
   Inline_Stack_Push(sv_2mortal(newSViv(ltime->tm_sec)));
   Inline_Stack_Push(sv_2mortal(newSViv(ltime->tm_isdst)));
   Inline_Stack_Done;
}
END_OF_C_CODE


exit;
---- stop copying here ---------

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org