[snap-0.11] Segmentation fault

[Björn Peemöller]

Hi,

I'm trying to deploy a snap-0.11 application under a sub-url and ran into a segmentation fault. I could reproduce it via:

    mkdir testapp
    cd testapp
    snap-0.11 init default
    cabal-dev install-deps
    # change src/Site.hs:85 from
    #   addRoutes routes
    # to
    #   addRoutes [("/foo/bar", route routes)]
    cabal-dev configure
    cabal-dev build
    dist/build/testapp/testapp

(I don't know if this is the intended way to install an application under a sub-url, but couldn't find documentation adressing this.)

While URLs like /foo/bar, /foo/bar/login etc. work as expected, the url /foo/ leads to a segmentation fault (using GHC 7.6.2):

    $ dist/build/snap-011-default/snap-011-default
    no port specified, defaulting to port 8000
    User JSON datafile not found. Creating a new one.
    Initializing app @ /
    Initializing heist @ /
    ...loaded 7 templates from ~/testapp/snaplets/heist/templates
    Initializing CookieSession @ /sess
    Initializing JsonFileAuthManager @ /auth

    Listening on http://0.0.0.0:8000/
    Segmentation fault

The log files also do not mention anything suspicious.

I really appreciate some advice on this issue. If someone needs more information about my environment, I will be happy to provide it.

Regards,
Björn

[Gregory Collins]

Hi,

More information would be helpful, especially:

  * your platform

  * can you get a backtrace from gdb?

  * what version of "hashable" do you have installed? Try "cabal install --constraint='hashable<1.2'" to see if it makes the problem go away.

G

--
 
---
You received this message because you are subscribed to the Google Groups "Snap Framework" group.
To unsubscribe from this group and stop receiving emails from it, send an email to snap_framewor...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
Gregory Collins <gr...@gregorycollins.net>

[Björn Peemöller]

Am Dienstag, 2. April 2013 16:34:47 UTC+2 schrieb Gregory Collins:

Hi,

More information would be helpful, especially:

  * your platform

I'm running Debian 6.0.7 (squeeze) on a i686 box:

$ uname -a
Linux palmer 3.2.0-0.bpo.4-686-pae #1 SMP Debian 3.2.39-2~bpo60+1 i686 GNU/Linux
 

  * can you get a backtrace from gdb?

I did, and it reports the following:

     Program received signal SIGSEGV, Segmentation fault.
    [Switching to Thread 0xb6bfeb70 (LWP 13972)]
    0x085bdf0a in hashable_siphash24_sse2 ()

A quick search led me to https://github.com/tibbe/hashable/issues/60, which seems to be the cause.

 

  * what version of "hashable" do you have installed? Try "cabal install --constraint='hashable<1.2'" to see if it makes the problem go away.

hashable-1.2.0.5. Configuring to hashable-1.1.* indeed solved the problem, so I think that hashable is to blame.

Thanks for your suggestions!
Björn