[openbeos] Re: building obos

On 2005-02-25 at 20:15:49 [+0100], Daniel Furrer wrote:
>
> Ok, thanks, that helped a bit. There were two small errors in the
> gensyscalls.cpp which I fixed here. See below for the diff. But then it
> still fails with the following error messages:
> 
> ...patience...
> don't know how to make
> ./distro/x86.R1/beos/system/add-ons/kernel/drivers/dev/tty
> don't know how to make
> ./distro/x86.R1/beos/system/add-ons/kernel/drivers/dev/console
> don't know how to make
> ./distro/x86.R1/beos/system/add-ons/kernel/console/vga_text
> ...found 4923 target(s)...
> ...updating 45 target(s)...
> ...can't find 3 target(s)...
> ...can't make 1 target(s)...
> GenSyscallInfos1 objects/x86.R1/tools/gensyscalls/gensyscalls_infos.c
> 
> objects/x86.R1/tools/gensyscalls/gensyscallinfos
> objects/x86.R1/tools/gensyscalls/syscalls.h.pp
> objects/x86.R1/tools/gensyscalls/gensyscalls_infos.c
> 
> ...failed GenSyscallInfos1
> objects/x86.R1/tools/gensyscalls/gensyscalls_infos.c ...
> ...skipped <src!tools!gensyscalls>gensyscalls_infos.o for lack of
> <src!tools!gensyscalls>gensyscalls_infos.c...
> ...skipped gensyscalls for lack of
> <src!tools!gensyscalls>gensyscalls_infos.o...
> ...skipped <syscalls>syscall_dispatcher.h for lack of gensyscalls...
> ...skipped <src!kernel!core>syscalls.o for lack of
> <src!kernel!core>syscalls.c...
> ...skipped kernel_core.o for lack of <src!kernel!core>syscalls.o...
> ...skipped kernel for lack of kernel_core.o...
> ...skipped config.x86.ini for lack of kernel...
> ...failed updating 1 target(s)...
> ...skipped 7 target(s)...
> $
> 
> It tells me something of a gensyscalls_infos.c which I have no idea where
> this should come from since there is no such file anywhere in the cvs.

That's OK, since it is a generated file. The line in the jam output starting 
with `GenSyscallInfos1' shows the point at which jam tries to build it. The 
file is generated by a command `gensyscallinfos' (which is built before) from 
a file `syscalls.h.pp' (a preprocessed `headers/private/kernel/syscalls.h'). 
Since jam says that the execution of the `GenSyscallInfos1' failed, it 
directly before printed the command line it let the shell execute (in the 
mail listed as three lines).

So far so good. The only problem is, that `gensyscallinfos' shouldn't fail. 
At least it shouldn't fail without printing an error message (which it 
apparently didn't do). I just looked at the source and indeed it should print 
an error message in all cases of failure. If you like you can insert a few 
printf()s to understand what really happens. Though before it wouldn't harm 
to check whether the output file 
(`objects/x86.R1/tools/gensyscalls/gensyscalls_infos.c') really hasn't been 
generated.

> PS: The gensyscalls.cpp should either use only int and no int32's or
> otherwise include some definition for the int32 somewhere. Here's my patch:

Yep, thanks.

CU, Ingo

Other related posts: