Second attempt to reply, this time from a subscribed address... On Fri, Jun 06, 2008 at 02:56:20PM +0930, Sharif Oerton wrote: > > > > e.g. perhaps the '\' in the second sed was not the last character > > on a line (it's a continuation, just like in the configure command). > > Getting that wrong could conceivably wipe out gcc/Makefile.in - what > > do you get for > > ls -l ../gcc-4.1.2/gcc/Makefile.in* > > > > ? - if any of these files are zero length, you messed up the seds. > > 200, for all of them. Slight surprise that they are all identical-length: Makefile.in.orig should be about 4 characters longer because the first sed shortens it, and similarly the second sed adds about 21 characters to what is in Makefile.in.tmp (my maths might be out!). So, I don't think your seds worked. I'm less than familiar with gcc/Makefile.in, but 200 sounds a bit short, so I took a look at my own tarball:
[EMAIL PROTECTED] ~ $tar -tvf /sources/gcc-4.1.2.tar.bz2 gcc-4.1.2/gcc/Makefile.in -rw-rw-r-- mitchell/mitchell 198099 2006-11-01 14:40 gcc-4.1.2/gcc/Makefile.in [EMAIL PROTECTED] ~ $ A bit of a difference from 200. I think you need to confirm your gcc tarball is still intact (the md5sum is in section 3.2), then work _slowly_ through section 5.11.1. Before the first sed, look at the size of gcc/Makefile.in then copy it, run the sed, and check again. If the sizes are still sensible, do the second sed and again check. If, for either sed, the sizes are not sensible, diff the original and resulting files to see what changed, it might help identify what is going wrong. In fact, diff them anyway and look at what changed, so you can be sure it is correct. If it all looks good to you, apply the patch, wipe out any existing gcc-build directory, recreate gcc-build, and run the configure (logging it through 'tee' as I suggested earlier). Then run make, and again log it. If it fails in the same way, you'll need to go through the output from configure with the proverbial fine-toothed comb. Perhaps, you could post the interesting values from config.log (SET_MAKE, LANGUAGES) - I doubt the full log will be small enough to post. > > The only thing that I can see that might be relevant in configure's output is > this: > > *** This configuration is not supported in the following subdirectories: > target-libada gnattools target-libgfortran target-libffi target-boeh-gc > target-zlib target-libjava zlib fastjar target-libobjc > Funnily enough, that output is normal. ĸen -- das eine Mal als Tragödie, das andere Mal als Farce ----- End forwarded message ----- -- das eine Mal als Tragödie, das andere Mal als Farce -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
