[kaffe] Tiny steps towards tinderboxing kaffe

[Dalibor Topic]

Hi Karl,

Karl Trygve Kalleberg wrote:
> I talked to Dalibor a while back about writing some sort of tinderbox/testing
> harness for kaffe, to help in detecting compile-time (and in time, run-time)
> errors during development.

Always a good thing :)

> In Gentoo, we already have some tools for building packages from sources on a
> regular basis. This is what our package system was designed to do; build
> programs from source code, so people could build their own "proper"
> distribution.

Yeah, my thoughts were going in the direction of using gentoo ebuilds to 
  verify that kaffe stays useable for software it was useable before. 
Ito usually catches a log of 'drift' problems in our class libraries, 
but it'd be nice to automate that a little bit.

> Naturally, the kaffe developers do not want (nor should they need) to run Gentoo,
> and that is in no way a requirement for the system I'm working on. While my system
> is using Portage at its core, all development happens on Debian 3.0rc2 running
> on sparc64.

Oh, I don't mind running Gentoo, I just don't have a spare box with high 
speed inet access for it around ;)

> I have however, run into a snag while trying to compile HEAD of kaffe now:
> 
> -----
> gcc -DHAVE_CONFIG_H -I. -I. -I../../../../config -I../../../../include/kaffe
> -I. -I. -I../.. -I./../.. -I../../../../config -I../../../../config
> -I../../../../include -I../../../../include -I../../../../kaffe/xprof
> -I../../../../include -DTRANSLATOR
> -I/var/tmp/portage/kaffe-cvs-0/work/kaffe/./kaffe/kaffevm
> -I/var/tmp/portage/kaffe-cvs-0/work/kaffe/./kaffe/kaffevm/jit
> -I/var/tmp/portage/kaffe-cvs-0/work/kaffe/./kaffe/kaffevm/systems/unix-jthreads
> -I../../../../config -I../../../../include -DKAFFE_FEEDBACK -O2
> -fomit-frame-pointer -Wno-import -Wchar-subscripts -Wcomment -Wformat=2
> -Wimplicit-int -Wmain -Wmissing-braces -Wmultichar -Wparentheses -Wreturn-type
> -Wswitch -Wtrigraphs -Wuninitialized -W -Wundef -Wshadow -Wpointer-arith
> -Wbad-function-cast -Wcast-qual -Wcast-align -Wwrite-strings -Wconversion
> -Wsign-compare -Waggregate-return -Wstrict-prototypes -Wmissing-prototypes
> -Wmissing-declarations -Wmissing-noreturn -Wredundant-decls -Wnested-externs
> -Winline -Wlong-long -c signal.c -Wp,-MD,.deps/signal.TPlo -fPIC -DPIC -o
> .libs/signal.o
> 
> In file included from /usr/include/sys/socket.h:35,
>                  from ../../../../config/config-io.h:27,
>                  from jthread.h:36,
>                  from signal.c:18:
> /usr/include/bits/socket.h: In function `__cmsg_nxthdr':
> /usr/include/bits/socket.h:277: warning: cast increases required alignment of target type
> signal.c: In function `nullException':
> signal.c:98: structure has no member named `sigc_sp'

Intersting. I've looked a little bit around, and it seems to have been 
flushed from glibc by this patch: 
http://sources.redhat.com/ml/libc-hacker/2000-03/msg00068.html

so the solution wil be to try to find what the equivalent replacement 
fields in the new structs are. I'll look into it right now.

> PS. Please correct me if this is the wrong mailing list to spew out
> compile-time error logs. I couldn't find a developer-only mailing list.

One list to fix them all :)

cheers,
dalibor topic