perl-embperl mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From BeeThere Admin <>
Subject Re: Intermittent Segfaults
Date Mon, 20 Mar 2006 23:37:31 GMT
I don't know the details, it seems to be a pretty general problem. Gerald seems to be on it
though. He even said that he was working through the weekend (bless his heart :-)
The symptom on the user side, at least for me is a completely blank page. It looks like the
server sends back a 200 OK, but then nothing else. There doesn't seem to be any way to debug
the problem at the Embperl-user side because none of the page code ever gets executed, not
even your base.epl (or whatever you call it). A page refresh will usually result in the expected
page rendering, but at least for us, this is a critical issue holding a major release, business
deals, etc.
I looked into the logs because a user told me they couldn't get into a section of the site
because they only got a blank page. The page seems to be random, but I can easily reproduce
the problem in my dev environment now.
Bee Software LLC 
See us in action at:

----- Original Message ----
From: Hartmaier Alexander <>
Sent: Monday, March 20, 2006 12:50:29 PM
Subject: RE: Intermittent Segfaults

I just found out that I have segfaults on my servers too.
OS is debian testing
The apache startup message should include all infos you might want to know too:
[Mon Mar 20 18:27:05 2006] [notice] Apache/2.0.55 (Debian) Embperl/2.1.0 mod_ssl/2.0.55 OpenSSL/0.9.8a
mod_apreq2-20051231/2.5.7 mod_perl/2.0.2 Perl/v5.8.8 configured -- resuming normal operations

I've recompiled Embperl today to make sure it uses the current software versions but that
didn't help.


-----Original Message-----
From: Gerald Richter [] 
Sent: Friday, March 17, 2006 9:16 PM
To: 'BeeThere Admin';
Subject: RE: Intermittent Segfaults

> I have continued testing, and I can confidently say that the 
> problem is definitely not an out-of-memory issue. 


> The new dev 
> version you sent is segfaulting about once for every 100 page 
> requests in dev now, with no load except me clicking around, 
> on a machine with 2G total RAM and plenty of free memory. 
> The segfaults have occurred during pretty much every type of 
> page request, and also not occurred on those same requests. 
> Hitting refresh will always result in the page being 
> correctly sent, with not segfault, on at most a couple pushes.
> Did you get the back-traces I sent? 

Yes. These backtraces are with the new version I sent, right? (I need to
make sure the line numbers of the backtrace matches my source)

> I sent them to your 
> personal account, but I realized that you may check the list 
> more. Every back-trace points to "EMBPERL2_Node_appendChild 
> .... at epdom.c:2835" as the point of segfaulting.
> The only Embperl tags I ever use are the PerlBlock '[* *], [+ 
> +], [- -]' tags. Is there someway I can bypass this 'epdom' 
> code all together? (I don't see any reason for dom-parsing)

This code also parses [..] tags, so you will need it. (It never builds a
full dom tree of your source, it only parses out the tags that you tell
Embperl to process)

I have to take a deeper look at the backtraces. I will come back to you as
soon as possible


** Virus checked by BB-5000 Mailfilter ** 

To unsubscribe, e-mail:
For additional commands, e-mail:

T-Systems Austria GesmbH
Rennweg 97-99, 1030 Wien
Telefon +43 (0)57057 0, Telefax +43 (0)57057 8155
Bank Austria AG, BLZ 12000, Kto. 233-111-674/00, IBAN AT291200023311167400
Handelsgericht Wien, FN 79340b  DVR Nr. 0713520  ARA-Lizenz Nr. 12554
ATU 15169001

Hinweis: Dieses E-mail kann vertrauliche und geschtzte Informationen enthalten.
Sollten Sie nicht der beabsichtigte Empfnger sein, verstndigen Sie bitte den Absender und
lschen Sie dieses E-mail dann sofort.

Notice: This e-mail contains information that is confidential and may be privileged.
If you are not the intended recipient, please notify the sender and then delete this e-mail

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message