Bug #1197

SIGABRT on MacOS X + select

Added by Anonymous over 3 years ago. Updated almost 2 years ago.

Status:Missing Feedback Start:
Priority:Normal Due date:
Assigned to:jan % Done:

0%
Category:core
Target version:1.5.0
Missing in 1.5.x:

Description

OS X V. 10.3.9

lighttpd-1.4.15

Python 2.5.1

web.py framework (http://webpy.infogami.com/install)

Shell output:

% sudo lighttpd -Df code.py

Abort trap

%

Abort trap occurs when I visit the page http://localhost:8080 or remotely at http://ip:8080. This same setup works fine on my other mac, which has the following versions:

OS X V. 10.4.9

Python 2.5

I have the following files:

code.conf (lighttpd config)

code.py (simple hello world using web.py)

code.trace.txt (a kdump of a ktrace ; The first request attempt from localhost worked, the second attempt from a remote computer caused the crash )

gdb_bt.txt (a gdb backtrace)

code.conf (11.3 KB) Anonymous, 05/24/2007 06:50 pm

code.py (305 Bytes) Anonymous, 05/24/2007 06:51 pm

code.trace.txt (12.9 KB) Anonymous, 05/24/2007 06:51 pm

gdb_bt.txt (867 Bytes) Anonymous, 05/24/2007 06:52 pm

Associated revisions

Revision 2589
Added by stbuehler about 1 year ago

Change mod_expire to append Cache-Control instead of overwriting it (fixes #1197)

History

Updated by Napalm-Llama over 3 years ago

Yup, I'm seeing this exact same problem (SIGABRT on first TCP connection, and an identical gdb backtrace).

I'm on a uClibc-based embedded Linux box though. The processor, however, is PPC. I'm guessing that could be what's causing the problem. I'm using the -mcpu=603e option for GCC, if that's any help.

I don't know much about coding, (at least not in C), but I'm willing to do what it takes to help get this bug ironed out if you can't reproduce it at your end.

Cheers :)

Updated by jan about 3 years ago

does the same happen with kqueue or poll as fdevent handler ?

Updated by jan about 3 years ago

  • Status changed from New to Assigned

Updated by admin about 3 years ago

Replying to jan:

does the same happen with kqueue or poll as fdevent handler ?

I'm sorry, I don't know what that means. How would I find out?

Updated by Anonymous almost 3 years ago

I'll add my comment to the list -- I'm seeing the same issue on MacOS, albeit with an FCGI handler implemented in Haskell (and then compiled to native). For me, the server will take ~500 hits before crashing.

This is with the freebsd-kqueue event handler setting.

-- paulrbrown

Updated by Anonymous almost 3 years ago

Configuring the "poll" event handler instead yields an apparently stable server (under hammering by ab).

-- paulrbrown

Updated by stbuehler almost 2 years ago

  • Status changed from Assigned to Missing Feedback
  • Pending changed from Yes to No
  • Patch available set to No

Backtrace to old now.

Also available in: Atom