amd64-build problems//now problem with TCL

blux.net at lineone.net blux.net at lineone.net
Sun Jan 6 11:00:52 PST 2008


On Sunday 06 January 2008 13:45:55 Ken Moffat wrote:
> On Sun, Jan 06, 2008 at 11:19:52AM +0100, blux.net at lineone.net 
wrote:
> > checking for gcc... gcc
> > checking for C compiler default output file name... a.out
> > checking whether the C compiler works... configure: error: cannot 
run
> > C compiled programs.
> > If you meant to cross compile, use `--host'.
> >
> > ######### I check config.log and found the following snippet
> > configure:1848: ./a.out
> > ./configure: ./a.out: /tools/lib64/ld-linux-x86-64.so.2: bad ELF
> > interpreter: No such file or directory
> > configure:1851: $? = 126
> > configure:1860: error: cannot run C compiled programs.
> >
> >
> >
> > ####### I did the following as super user
> > [root at bigworda cooker]# cd '/mnt/lfs/tools/lib'
> > [root at bigworda lib]# ln -s ld-2.7.so  ../lib64/ld-linux-x86-64.so.
2
> > [root at bigworda lib]# chown -R lfs:lfs /mnt/lfs
> > [root at bigworda lib]
> >
> > ######## and as lfs  i ran the configure script again
> > checking for gcc... gcc
> > checking for C compiler default output file name... a.out
> > checking whether the C compiler works... configure: error: cannot 
run
> > C compiled programs.
> > If you meant to cross compile, use `--host'.
> > See `config.log' for more details.
> > lfs:/mnt/lfs/sourcesTSjn08/tcl8.5.0/unix$
>

>  I don't think there is a lot I can suggest here.  The short answer 
is
> to point you to again look at config.log - presumably, the error is
> something different this time (either that, or your symlink is
> somehow broken).  Mind you, the only time I remember seeing a 'bad
> ELF interpreter' message was in a test to see if a 32-bit kernel
> could run 64-bit binaries (it couldn't).
>
>  The longer answer is to point out that getting a working build
> method is difficult.  Technically, you are straying from the svn
> book to even attempt x86_64.  That is not a bad thing, neither is
> trying newer versions, but both lead to frequent breakage.  So,three
> alternatives for you:
>
>  If you have svn, download the jh branch, plus the patches it refers
> to, and render it - you can then read it to learn what has been
> discovered and recommended in building natively for x86_64.
>
>  Alexander keeps releasing versions of the Live CD with an
> unofficial x86_64 book.  Google for a mail headed 'Stable LFS
> LiveCD 6.3-r2160 released' (It was sent to -dev at the end of
> December.  If you are playing with the development book and newer
> versions of packages, you should probably be subscribed to -dev, or
> following its archives.)
>
>  In many ways, following clfs is a workable solution.  I can't say
> if the current pure64 book will cause problems or not, I'm too busy
> with a couple of other architectures.
>

> to point you to again look at config.log - presumably, the error is
> something different this time (either that, or your symlink is
> somehow broken).

####first good news  tcl 8.5 now compiles  when link correctly made as 
in:

cd   /tools/lib/lib64  
ln -s /tools/lib/ld-2.7.so  ld-linux-x86-64.so.2

tcl-8.4.16  STILL DOES NOT COMPILE.

### now the not so good:  expect will not compile and patches only   
after repeated efforts as below:

#########  expect-5.43 patching
lfs:/mnt/lfs/sourcesTSjn08/expect-5.43$ patch -Np1 -i ../expect-5.43.0-
spawn-1.patch
patching file exp_chan.c
Hunk #1 succeeded at 622 (offset 103 lines).
patching file exp_command.h
Hunk #1 succeeded at 30 with fuzz 1 (offset 5 lines).
Hunk #2 succeeded at 99 (offset 1 line).
patching file expect.c
Hunk #2 succeeded at 1628 (offset 1 line).
Hunk #4 succeeded at 1730 (offset 1 line).
Hunk #6 succeeded at 2864 (offset 5 lines).
lfs:/mnt/lfs/sourcesTSjn08/expect-5.43$ cp configure{,.bak}
lfs:/mnt/lfs/sourcesTSjn08/expect-5.43$ sed 's:/usr/local/bin:/bin:' 
configure.bak > configure
lfs:/mnt/lfs/sourcesTSjn08/expect-5.43$
                                          

########## expect-5.43 file compile  (fails)
exp_inter.c: In function 'intMatch':
exp_inter.c:294: error: 'TCL_REG_BOSONLY' undeclared (first use in 
this function)
exp_inter.c:294: error: (Each undeclared identifier is reported only 
once
exp_inter.c:294: error: for each function it appears in.)
exp_inter.c: In function 'Exp_InteractObjCmd':
exp_inter.c:866: error: 'TCL_REG_BOSONLY' undeclared (first use in 
this function)
make: *** [exp_inter.o] Error 1
lfs:/mnt/lfs/sourcesTSjn08/expect-5.43$


as everall helpful suggestions are welcomed

====

blux



__________________________________________________
Get up to £150 by recycling your old mobile - visit www.tiscali.co.uk/recycle




More information about the lfs-support mailing list