[erlang-questions] Binary pattern matching inconsistencies with R12B

Rory Byrne <>
Fri Feb 29 22:42:48 CET 2008


On Fri, Feb 29, 2008 at 09:32:24PM +0200, Kostis Sagonas wrote:
> 
> I confirm your experiences.  Mine are slightly different than yours, but 
> the end results are the same; see below.  This is with the most recent 
> development version.  I suspect a GC-related.
> 

Thanks Kostis. Bit of a relief really - I wasn't exactly making a 
whole lot of progess fixing my code!

> 
> PS. Surprisingly, I cannot manage to get a seg-fault if I compile
>      to native code. [using hipe:c() instead of c()]
> 

Excellent. I just tried hipe on the amd64 machine and no seg-fault.
Oddly hipe doesn't seem to be enabled on my i386 even though
I compiled it the same way as the amd64 version.

As an added bonus, I was stearing clear of hipe because I had 
read somewhere that there was problems with running it on 
a xen instance (due to the threading library used if I recall).
However, the amd64 machine I just tried it on is a xen 
instance, so that looks promising. So, thanks again!

Rory



More information about the erlang-questions mailing list