Re: [OSM-dev] OSM server installation - problems with create

2008-09-04 Thread Frederik Ramm
Hi, Stefan de Konink wrote: if (mmap2(...) == -1) exit(-1); The problem is that between our program code (which is of course fast and efficient) and the kernel (which offers the mmap2 system call) sits a fat beast (a.k.a dynamic programming language) named ruby that magically frees

Re: [OSM-dev] OSM server installation - problems with create

2008-09-04 Thread Rogier Wolff
On Thu, Sep 04, 2008 at 12:09:24AM +0100, Tom Hughes wrote: Rogier Wolff wrote: The weird thing is: It apparently doesn't check for memory allocation return codes for a while before it notices that it's out of memory. Ever heard of garbage collection? You know, that thing where when a

Re: [OSM-dev] OSM server installation - problems with create

2008-09-04 Thread Tom Hughes
Frederik Ramm wrote: Stefan de Konink wrote: if (mmap2(...) == -1) exit(-1); The problem is that between our program code (which is of course fast and efficient) and the kernel (which offers the mmap2 system call) sits a fat beast (a.k.a dynamic programming language) named ruby that

Re: [OSM-dev] OSM server installation - problems with create

2008-09-03 Thread Tom Hughes
Rogier Wolff wrote: The weird thing is: It apparently doesn't check for memory allocation return codes for a while before it notices that it's out of memory. Ever heard of garbage collection? You know, that thing where when a memory allocation fails the language does some cleanup and then has