Configuration Information [Automatically generated, do not change]:
Machine: x86_64
OS: linux-gnu
Compiler: /usr/intel/bin/gcc
Compilation CFLAGS: -DPROGRAM='bash' -DCONF_HOSTTYPE='x86_64'
-DCONF_OSTYPE='linux-gnu' -DCONF_MACHTYPE='x86_64-unknown-linux-gnu'
-DCONF_VENDOR='unknown' -DLOCALEDIR='/usr/intel/00r1.1/share/locale'
-DPACKAGE='bash' -DSHELL -DHAVE_CONFIG_H -I. -I. -I./include -I./lib -O2
-fPIC -L/usr/intel/pkgs/readline/5.0-64/lib
-I/usr/intel/pkgs/readline/5.0-64/include
-I/nfs/pdx/disks/comp.build_system.1/bash/3.0/em64t_linux26/bash-3.0/include
-I/nfs/pdx/disks/comp.build_system.1/bash/3.0/em64t_linux26/bash-3.0
-L/usr/intel/pkgs/ncurses/5.4/lib -I/usr/intel/pkgs/ncurses/5.4/include
uname output: Linux filc9194 2.6.5-7.276.PTF.196309.1-smp #1 SMP Mon Jul 24
10:45:31 UTC 2006 x86_64 x86_64 x86_64 GNU/Linux
Machine Type: x86_64-unknown-linux-gnu
Bash Version: 3.0
Patch Level: 0
Release Status: release
Description:
Summary:
Characters aren't being properly escaped/evaluated in a Bash
Full Description:
I'm trying to take a series of lines in a file like so (ignore
1.-3. characters -- they're just there for
visual breaking):
1. This is a file with multiple
2. fields spread across multiple
3. lines.
and group it up into line by line groups, and analyze with
"for" in bash. The problem is that if I did this out of the box
like the following, it 'flattens' the list:
Example:
for j in `cat $i`; do
echo $j
done
Output:
This
is
a
file
[...]
According to the Advanced Bash Scripting Manual, the following
should be -- and is -- possible:
Example:
for j in "This is a file with multiple" "fields spread
across multiple" "lines."; do
echo $j
done
Output:
This is a file with multiple
fields spread across multiple
lines.
Ok, cool. Now, in order to do this I chose Perl as my method of
escaping the lines
(couldn't get GNU sed to work as expected and I don't know awk
/ tcl), as follows:
for j in `$PERL -e "chomp(\$lines=qx{cat $i});
foreach(split(\"\\n\", \$lines)) { printf(\"\'%s\'\", \$_); }"`; do
echo $j
done
Output:
filc9194[202]% ./scripts/mk_batch.sh vfe1 falcon2
syntax error at -e line 1, near "chomp(="
Execution of -e aborted due to compilation errors.
No valid lines in
/nfs/fm/disks/fm_vt_n19029/users/gwcooper/vibe_regress/scripts/../archived/esperanto_list.txt
(all commented and/or empty)
Interestingly enough, the script fails. This shouldn't be done
because I properly escaped the dollar signs in the
double quotes, thus the Perl variables shouldn't be interpreted
and evaluated by Bash, but instead be left alone
for Perl to do its bidding.
I know this solution isn't the most ideal one, and although I
might receive workarounds which use sed, the fact is
that this is a bug with Bash. This behavior fails not only in
older versions of Bash (2.0, the system Bash), but also
the version that Intel compiles for its Suse machines (3.0.0),
as well as the version available for Cygwin (3.1.7).
[EMAIL PROTECTED] ~
$ ./test.sh
syntax error at -e line 1, near "chomp(="
Execution of -e aborted due to compilation errors.
[EMAIL PROTECTED] ~
$ bash --version
GNU bash, version 3.2.17(15)-release (i686-pc-cygwin)
Copyright (C) 2005 Free Software Foundation, Inc.
Repeat-By:
Run the following lines (replace $i appropriately with a file of your
choosing):
#!/bin/bash
PERL=`which PERL`
i=/pick/a/file/with/lines/in/it
for j in `$PERL -e "chomp(\$lines=qx{cat $i}); foreach(split(\"\\n\",
\$lines)) { printf(\"\'%s\'\", \$_); }"`; do
echo $j
done
Fix:
The problem lies in the parser somewhere, but I can't look through the
code (at least not now) because I have my hands full with other things. If I do
find the issue and can submit a fix, I will.
_______________________________________________
Bug-bash mailing list
Bug-bash@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-bash
