bug-gnubg
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bug-gnubg] Compile latest version under Windows


From: Jim Segrave
Subject: Re: [Bug-gnubg] Compile latest version under Windows
Date: Wed, 27 Nov 2002 13:23:55 +0100
User-agent: Mutt/1.4i

On Wed 27 Nov 2002 (08:39 +0000), Joern Thyssen wrote:
> On Tue, Nov 26, 2002 at 11:31:16PM +0100, Jim Segrave wrote
> > On Tue 26 Nov 2002 (20:43 +0100), Holger wrote:
> > > Jim Segrave wrote:
> > > 
> > > > If not, then maybe:
> > > > 
> > > > Start with a copy of (it's short) br1.c. You may already have it, if
> > > > not, it's under CVS
> > 
> > I stand corrected - it's always generated. I can send you a copy of 
> > either the empty (1.5K) or a fully generated (huge version of br1.c)
> > if that would help.
> 
> err, br1.c should not be in CVS... Empty version? I'm a bit confused...

No, it's not under CVS, that was my mistake (I carelessly looked at
br1.c, saw the $Id line but didn't look at it carefully and assumed
I'd got it from CVS). What was actually happening was:

make was running and got to the step where it runs 
  ./makebearoff -o 6 -s 7999999 | ./makebearoff1 -o br1.c

But makebearoff.c had a couple of erroneous single quotes in the
printfs, so it core dumped. makebearoff1 therefor received no data
from makebearoff, so it built a C file with

static unsigned char acBearoff1[] = { 
};

since makebearoff1 ran to completion without errors, the make
continued. 

There's no way that I know of to run the above pipeline and detect a
failure in makebearoff. The options are to change the makefile to run
makebearoff sending it's output to a temporary file, then run
makebearoff1 process the temporary file (so that errors in either will
be seen by make), or change makebearoff1.c to check for valid input
from makebearoff. The second may be harder (makebearoff can core dump
before anything is output - this happened to me with gcc 2.95, or it
can output a small amount and then core dump - gcc 3.2), so I'd
suggest the temp file solution. 

Or perhaps someone knows how to run a pipeline with /bin/sh and return
the logical or of all the exit statuses.

Once I fixed  makebearoff, br1.c got filled with the data (although
masking the values with 0xff saves a few thousand warnings - I'm
guessing that by default chars are unsigned in your system and signed
in mine)

-- 
Jim Segrave           address@hidden




reply via email to

[Prev in Thread] Current Thread [Next in Thread]