[Lsb-infrastructure] First look at April 18 updates

[Stew Benedict]

On Sat, 19 Apr 2008, Stew Benedict wrote:

> 
> On Sat, 19 Apr 2008, Wichmann, Mats D wrote:
> 
> > 
> > core test 
> > 
> > can't quite see the fail, which is that chmog didn't
> > get installed. I don't see any errors in building it 
> > in the log that was emailed out.
> > 
> 
> Yep, if chmog or any of that pre-build stuff fails, things pretty much go 
> downhill from there. I can probably do something to try and capture this a 
> little better for post-mortem analysis. The is a journal of sorts from 
> that part of the build, iirc.
> 
> > this build is littered with warnings:
> > 
> > /opt/lsb/include/signal.h:25:11: warning: extra tokens at end of #endif
> > directive
> > 
> > which is the comments on the end of the version #ifdefs
> > I guess core test compiles in an old-fashioned mode which
> > doesn't understand the // comment convention? 
> 
> Yes, I think it's the -ansi flag it uses.
> 

So the "tools" prebuild fails with:

In file included from /tmp/chd.ss.c:1:
/opt/lsb/include/stdio.h:95: error: expected identifier or ( before / 
token

/opt/lsb/include/stdio.h:95:
// Function prototypes

"log" file is results/install.0001

So I either need to get the core test build to accept the "//" style 
comments or we need to change the headers.

========================================

libbat/readline is failing with:

lsbcc -c -DHAVE_CONFIG_H -I. -I. -DRL_LIBRARY_VERSION='"5.2"' -fPIC -g -O2 
rltty.c
rltty.c: In function save_tty_chars:
rltty.c:423: error: dereferencing pointer to incomplete type
and more...

I'm not really seeing what's happening with this one.

===========================================

trying to dig through libstdcpp-test. It's failing early on building 
libv3test.a, but this is another one that obscures what's really going on.

-- 
Stew Benedict
The Linux Foundation