Regards Ronny -----Ursprüngliche Nachricht----- Von: openbeos-bounce@xxxxxxxxxxxxx [mailto:openbeos-bounce@xxxxxxxxxxxxx] Im Auftrag von Jonathon Freeman Gesendet: Montag, 19. Februar 2007 18:28 An: openbeos@xxxxxxxxxxxxx Betreff: [openbeos] Re: Problems building today? I gave a shot at building this morning, without updating the tree or doing anything else, and it built fine. Very strange. Regards, Jonathon > -----Original Message----- > From: openbeos-bounce@xxxxxxxxxxxxx > [mailto:openbeos-bounce@xxxxxxxxxxxxx] On Behalf Of Ingo Weinhold > Sent: Sunday, February 18, 2007 6:14 PM > To: openbeos@xxxxxxxxxxxxx > Subject: [openbeos] Re: Problems building today? > > > On 2007-02-18 at 23:52:46 [+0100], Jonathon Freeman > <j_freeman@xxxxxxxxxxxxxx> wrote: > > > > I noticed earlier today that my automated builds on > HaikuHost started > > failing. The last successful build was at 18-Feb-2007 00:26:36 GMT. > > I've tried to build manually with no luck. Attached is the > jam log of > > my most recent attempt. > > > > Any ideas whatsoever would be greatly appreciated. > > The bfs_shell part of the build system is deliberately broken > to the extend that it should work again. In case of the > Volume.cpp which fails to compile according to your log, the > build system uses a good deal of target platform (aka Haiku) > headers although the file is compiled for the host platform > (apparently a RedHat Linux in your case). This deserves to be > fixed, but this is another story... (*) > > The problem here is simply that a GCC 3.4.6 for target > i386-redhat-linux is configured as Haiku cross-compiler, > which simply won't work (although ironically, it actually > looks a bit more correct in this context). Please have a look > at the ReadMe.cross-compile in the trunk's root directory how > to configure the build system for cross-compilation. > > CU, Ingo > > (*) As I see it, fixing the fs_shell/bfs_shell build would > involve 1) using the headers/build headers instead of the > Haiku headers (porting them where necessary), 2) adjusting > the fs_shell accordingly, and 3) and most importantly using a > modified version of the BFS code. > >