On Fri, 11 Mar 2011, Jeremy Evans wrote: > Is it better to show the wrong example and have something that works, or
That's were I don't agree. Because if it works, it's by luck. > to not show an example at all and have something that doesn't work? > What problems does it actually cause? The same problem that you outline below, there is a reason why we don't ship threaded perl and ruby. That is also the reason I marked devel/p5-Glib2 as BROKEN. I tried using -lpthread as well back in the time. For some stuffs it worked, for others it didn't. It's just too much of a mess and until we get rthreads, then I think we should not encourage using software that work most of the time by luck. Glib2 needs threads (for gthread), our perl is not threaded, and if you force -lpthread that means you will start messing with fds, it may work or not... Again as I said, it's just me, maybe others disagree. Also several of us spent quite some time moving hundreds of ports from using -lpthread into using -pthread (not all were done though). Again I'm not trying to be an ass here or prevent you from importing whatever you need, but I think this needed to be known. If other porters don't agree, just ignore me. > I was told that new software could not be imported if it required the > LD_PRELOAD hack. This is the only known way to work around it, unless > we start building ruby with pthread by default (which we don't for the > same reasons our perl isn't build with pthread). > > Jeremy > > -- Antoine
