Bug#492025: Timezone database is corrupt
Package: php5 Version: 5.2.6-2 Severity: critical Hello, after upgrading from php5 5.2.5-3 to 5.2.6-2 we experienced the following errors: Wordpress: Fatal error: strftime() [a href='function.strftime'function.strftime/a]: Timezone database is corrupt - this should *never* happen! in /var/www/www.domain.tld/wp-content/plugins/timezone.php on line 150 Typo3: Fatal error: mktime() [a href='function.mktime'function.mktime/a]: Timezone database is corrupt - this should *never* happen! in /var/www/www.domain.tld/typo3conf/temp_CACHED_ps4760_ext_tables.php on line 148 I'm sure other programs also have the same problem. We did not change anything with the php.ini config files. After downgrading to 5.2.5-3 again (this was just a show stopper) everything works fine again. Regards, Martin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#492025: [php-maint] Bug#492025: Timezone database is corrupt
Hello Sean!
Wordpress:
Fatal error: strftime() [a
href='function.strftime'function.strftime/a]: Timezone database is
corrupt - this should *never* happen! in
/var/www/www.domain.tld/wp-content/plugins/timezone.php on line 150
oy. this definitely needs to be fixed then, as we're going to ship 5.2.6
with lenny.
This would be a good idea then. But not everybody seems to have the
problem.
is there any chance you could provide us with a short php script to
replicate this problem (or the one from typo3, i suspect it's the same
problem)? i.e. if you could give us what's being executed on line 150
with any surrounding context removed (i.e. expand the variables being
passed to strptime() etc)?
Here's the last part of the code:
--
// Get the GMT offset from TZ
function option_gmt_offset()
{
$offset = strftime('%z', time());
if (substr($offset, 0, 1) == '-') {
$dir = -1;
} else {
$dir = 1;
}
$hr = intval(substr($offset, -4, 2));
$mn = intval(substr($offset, -2, 2));
return $dir * ($hr + ($mn / 60));
}
}
$wppTimeZone = new TimeZone;
--
You can download the whole script here: http://kimmo.suominen.com/sw/timezone/
I can also provide you with Typo3 code. But it should be the same.
also, can you provide the output of dpkg -l tzdata ?
I suppose you just need the version. 2008c-1 is installed and it is
correctly installed.
If there is anything else I can help you with feel free to contact me.
Regards,
Martin
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#492025: [php-maint] Bug#492025: Bug#492025: Timezone database is corrupt
Hello Thijs,
how do I check the timezonedb file? I think it should be ok because php
5.2.5-3 runs fine. But I will test the file anyway to be sure.
Is this timezone file new with 5.2.6-2? If yes and if it is located at a
new location php might not have the right to access it on our system. We
are preventing that php scripts can access everything on our servers.
I only tried with php5-cgi under Apache 2.2, so I cannot tell you if
php5-cli is doing the same thing. Do you have an example script for me to
test on command line?
Regards,
Martin
Mittwoch, 23. Juli 2008, 15:00:21 [GMT +0200 (CEST)]:
On Wed, July 23, 2008 14:34, Martin Sebald wrote:
function option_gmt_offset() {
I've tried to reproduce this with a PHP file containing just this function
and a call var_dump(option_gmt_offset()). Running that file with
php5-cli just yields the expected int(2).
So I can't reproduce this here (this is on powerpc). Can you confirm that
the setup as I described fails for you?
Just a wild guess, does debsums verify that the timezonedb isn't indeed
corrupted with you locally?
Thijs
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#492025: [php-maint] Bug#492025: Timezone database is corrupt
Hello Sean, well, we are using some kind of hybrid system. But basicly it is a system based on the unstable tree, but I'm sure there are some stable packages installed as well. The stable packages are replaced by unstable versions if apt-get wants to. So some are left, but not many. And we are using official Debian packages only, no packport packages. Where do I set a time zone? When installing a Debian machine it asks for a time zone. Here I always set this value to Europe/Berlin. I thought this was saved in /etc/timezone but I cannot find this file on any of our servers anymore. It seems to be moved somewhere else. Regards, Martin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#492025: [php-maint] Bug#492025: Bug#492025: Bug#492025: Timezone database is corrupt
Hello Sean, Hello Thijs! my guess is that your security rules are blocking access to /etc/localtime and files under /usr/share/zoneinfo then. Yes, that was it! I unblocked these files/directories and everything works fine now! :-) Maybe this information should be put somewhere for other security geeks like us?!? Regards and thank you very much for the help, Martin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#410561: Fw: Re: Error 500 when trying to edit tt_news entry
Hi all, me and most of all my customers having serious problems with a Typo3 extension called tt_news. When trying to edit a tt_news entry by clicking on the edit icon in the backend the server produces a 500 error. I contacted Rupert Germann in December 2006, the author of tt_news and after trying and sorting out he believes that this is a PHP 5.2.0 issue. Attached the mail Rupert sent me. As I'm not a programmer I cannot say for sure if Rupert is right. But at least it reads comprehensible to me. As tt_news is a very popular extension plus I sure that also other extensions or complete different PHP applications have problems with this xml parsing bugs I hope that this is fixed very soon. I was waiting for the release of PHP 5.2.1 and after that the release of the Debian package since then. Thanks and regards, Martin ===8===This is a forwarded message Von: Rupert Germann An: Martin Sebald Datum: Dienstag, 12. Dezember 2006, 10:34:17 [GMT +0100] Betreff: Re: Error 500 when trying to edit tt_news entry ===8==Original message text=== hi Martin, I also tried it with PHP 5.2.0-7 on Debian and after I wasted some hours with unsuccesful testing/debugging I downgraded to the former 5.1.6 version. The error seems to be located in the xmlparser of PHP. The extension manager from TYPO3 could not parse the extensions.xml file and also the parsing of big flexform xml files (like tt_news uses) failed. In the TYPO3 mailinglists there are several theads about the PHP 5.2 problems, f.i.: http://lists.netfielders.de/pipermail/typo3-english/2006-November/033111.html http://lists.netfielders.de/pipermail/typo3-english/2006-November/033273.html http://lists.netfielders.de/pipermail/typo3-english/2006-November/033480.html conclusion is that you either should downgrade to 5.1.6 or you could try the 5.2.1dev version where people say that this problem is fixed. greets rupert ===8===End of original message text===
Bug#338006: Postfix problem still there
Hello Kurt, will/could the changes in 0.9.8a-7 have any effect on the postfix issues I'm experiencing? Is it worth a try? A new postfix version has been uploaded to unstable, which has as changelog entry: * New upstream, fixes various TLS/SASL bugs. I suggest you try that. Thank you for this information. Postfix unstable 2.2.9 up and running, activated tls stuff in configuration again. I'll keep you up to date, but this will take one or two days until I see if the errors won't appear again. If the errors stay I will try the things you mentioned here: will/could the changes in 0.9.8a-7 have any effect on the postfix issues I'm experiencing? Is it worth a try? It might be. It seems that postfix is using the BIO interface, so it might be your problem. It never hurts to check, so we can rule things out. If you need any detailed information about the postfix configuration I'm running I will be happy to hand it over to you. What I need is a way to reproduce it. I don't think having your setup is going to help me, but feel free to send me the details. Maybe I should look at setting up postfix with ssl myself. Can you reproduce it when using s_client? Like: openssl s_client -starttls smtp -crlf -connect localhost:25 -bugs -no_ssl2 Try playing with some options, like adding -ssl3/-no_ssl3 or -tls1. Also see man s_client(1). Also try the same thing several times, it might work with some luck, and might fail the next time. Regards, Martin
Bug#338006: Postfix problem still there
Hello Kurt, will/could the changes in 0.9.8a-7 have any effect on the postfix issues I'm experiencing? Is it worth a try? If you need any detailed information about the postfix configuration I'm running I will be happy to hand it over to you. Regards, Martin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#338006: Postfix problem still there
Hello again, I'm sorry, but looking into the logs the last days and trying a few thing I have to say the following: The Postfix bugs are still there, either without the special smtpd_tls_cipherlist parameter from Yari and with this parameter. Without I get some the following errors in mail.log: Feb 2 06:50:53 servername postfix/smtpd[23187]: warning: TLS library problem: 23187:error:1408F10B:SSL routines:SSL3_GET_RECORD:wrong version number:s3_pkt.c:288: And with the parameter set I get many many of the following errors: Feb 7 06:30:19 servername postfix/smtpd[20597]: warning: TLS library problem: 20597:error:14094417:SSL routines:SSL3_READ_BYTES:sslv3 alert illegal parameter:s3_pkt.c:1057:SSL alert number 47: Well, I guess I have to turn off some features in Postfix to get rid of the errors (and the functionality I wanted to achieve). Regards, Martin
Bug#338006: Postfix problem still there
Hello! We believe that the bug you reported is fixed in the latest version of openssl, which is due to be installed in the Debian FTP archive: I upgraded to openssl 0.9.8a-6 but the Postfix errors stayed. It works now. I only upgraded openssl and forgot to upgrade libssl0.9.8. After upgrading I tried to comment out the smtpd_tls_cipherlist and it worked. Great! Why are the two packages openssl and libsslx.x.x not depending on each other? Regards, Martin
Bug#338006: Postfix problem still there
Hello! We believe that the bug you reported is fixed in the latest version of openssl, which is due to be installed in the Debian FTP archive: Was this meant for me? Otherwise sorry for spamming around... ;-) I upgraded to openssl 0.9.8a-6 but the Postfix errors stayed. I also got a mail from a nice guy named Yari about my Postfix problem: As I had the same problem, the workaround I've found is the following: in the main.cf configuration file add this directive for openssl: smtpd_tls_cipherlist = SSLv2:-LOW:-EXPORT:RC4+RSA in this way it works always, with TLS v1 and RC4 128 bit. Best Regards Yari Melzani Because I did not know what he wanted to say when talking about TLS v1 and RC4 128 bit (I need this to work with all TLS connections, not only some) I did not try it yet. Regards, Martin
Bug#338006: Postfix problem still there
Hello all, I think that the following Postfix problem is related to this bug. I found several entries in newsgroups linking to this bug report page. Jan 19 16:12:55 mailserver postfix/smtp[13270]: setting up TLS connection to mail.domain.tld Jan 19 16:12:55 mailserver postfix/smtp[13270]: SSL_connect error to mail.domain.tld: -1 Jan 19 16:12:55 mailserver postfix/smtp[13270]: warning: TLS library problem: 13270:error:1408F455:SSL routines:SSL3_GET_RECORD:decryption failed or bad record mac:s3_pkt.c:426: I'm running postfix 2.2.6-5 and openssl 0.9.8a-5. I just disabled TLS in Postfix because it produces errors 9 of 10 times (note that the error does not appear every time). Is there any timetable for the bugfix? :-) Regards, Martin
