Dear List,
H'mmm - I tried searching for a prior posting on this subject and got
Index file error: Index file 'lfs-dev.swish' and property file
'lfs-dev.swish.prop' are not related.
But that's by the way.
The problem was discovered while working through blfs (building libgpg-1.0),
but the problem is with the original lfs-built tools. I've been able to
reproduce it with less input and a simpler script as follows.
Set up files for testing as follows:
cat > test.in<<EOF
7 GPG_ERR_E2BIG
114 GPG_ERR_EALREADY
52 GPG_ERR_EBADE
9 GPG_ERR_EBADF
77 GPG_ERR_EBADFD
EOF
cat > test.awk <<EOF
BEGIN {
FS="[ \t]+GPG_ERR_";
}
//^#/ { next; }
{
print " { " $1 ", \"GPG_ERR_" $2 "\" },";
}
EOF
cat test.in | gawk -f test.awk
This is expected to produce
{ 7, "GPG_ERR_E2BIG" },
{ 114, "GPG_ERR_EALREADY" },
{ 52, "GPG_ERR_EBADE" },
{ 9, "GPG_ERR_EBADF" },
{ 77, "GPG_ERR_EBADFD" },
but in fact it produces
{ 7, "GPG_ERR_E2BIG" },
{ 114, "GPG_ERR_EALREADY" },
{ 52 EBADE, "GPG_ERR_" },
{ 9 EBADF, "GPG_ERR_" },
{ 77, "GPG_ERR_EBADFD" },
Clearly, two of the lines are clearly being split wrongly (in fact,
not at all, so $1 matches the entire line and $2 is empty).
There is no special problem with the original whitespace
(single normal space characters throughout).
I'm running a Pentium 4 box with an Intel 845PE chipset:
both my host environments are i686-pc-linux, built by me.
There's an 'LFS' set-up, exactly following the LFS book
as far as the kernel goes, but with glibc-2.3.5 and
binutils-2.15 . I also have gcc-3.4.4 installed on it.
The other system is (or was) Mandrake-10.1, but with my
own unmodified builds of binutils-2.16. Linux-2.6.11.6,
and gcc-3.4.3 . The odd one out is that I'm still using the
heavily Mandrake-patched version of glibc-2.3.3 .
I've compiled gawk-3.1.3 and gawk-3.1.4 with all of these
gcc/glibc combinations, and executed each of the resulting
executables in both systems (so using the host glibc and
ld-linux.so.2 in each case: no special load paths set up).
The *only* failing combination is
gawk-3.1.4 with the LFS system (so glibc-2.3.5), and this
breaks with all of the compilers.
I doubt very much whether the kernel or binutils version
matters a damn, so it looks like a mysterious combination
bug for gawk-3.1.4 and glibc-2.3.5 together.
Has anyone else seen this?
Can anyone else reproduce it?
Unless someone tells me I've painted myself into
an unreproducible private corner I'll raise it as
a bug with gawk and glibc.
Bernard Leak.
--
Before they made me, they broke the mould.
--
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
