Hi
I m somwwhat new to unix and i cannot figure out how do i extract the
values under the heading VALUE  (see below) for all the files ...i used
cut command as cut -f3 -d ' ' but couldnt get the desired output.
i need the output to be the list of all unique values ( we'll use uniq
command i think) not including .rodata or .data (grep -v i think ) ...i
need to execute this all from a PERL script ...it has to be done for
all object files ...i  wrote

$var = ".rodata\|.data";
system("objdump -r *.o|cut -f3 -d ' ' |grep -v $var");
i know that we'll have to remove statements like :
test1.o:     file format elf32-i386

RELOCATION RECORDS FOR [.text]:
OFFSET   TYPE              VALUE
from the output so that cut works coz it otherwise considers FOR as
the third field.

---------------------------------------------
[shell prompt]$ objdump -r *.o

test1.o:     file format elf32-i386

RELOCATION RECORDS FOR [.text]:
OFFSET   TYPE              VALUE
0000000a R_386_32          .rodata
0000000f R_386_PC32        printf
00000022 R_386_32          .rodata
00000027 R_386_PC32        printf
00000041 R_386_PC32        G
00000046 R_386_PC32        F
0000004e R_386_32          .rodata
00000053 R_386_PC32        printf



test.o:     file format elf32-i386

RELOCATION RECORDS FOR [.text]:
OFFSET   TYPE              VALUE
0000000a R_386_32          .rodata
0000000f R_386_PC32        printf
00000022 R_386_32          .rodata
00000027 R_386_PC32        printf
0000003a R_386_32          .rodata
0000003f R_386_PC32        printf
00000059 R_386_PC32        F1
0000005e R_386_PC32        F2

[ Post a follow-up to this message ]

i figured out a way :

[shell promp]$ objdump -r *.o | grep -v "RELOCATION\|OFFSET\|file
format"
| grep -v ".rodata\|.data"


0000000f R_386_PC32        printf
00000027 R_386_PC32        printf
00000041 R_386_PC32        G
00000046 R_386_PC32        F
00000053 R_386_PC32        printf




0000000f R_386_PC32        printf
00000027 R_386_PC32        printf
0000003f R_386_PC32        printf
00000059 R_386_PC32        F1
0000005e R_386_PC32        F2

now two problems remain :
a) using cut command n uniq command to extract unique third field
values
b)incorporating this command in a PERL system command

[ Post a follow-up to this message ]

jeniffer wrote:
> Hi
> I m somwwhat new to unix and i cannot figure out how do i extract the
> values under the heading VALUE  (see below) for all the files ...i used
> cut command as cut -f3 -d ' ' but couldnt get the desired output.
> i need the output to be the list of all unique values ( we'll use uniq
> command i think) not including .rodata or .data (grep -v i think ) ...i
> need to execute this all from a PERL script ...it has to be done for
> all object files ...i  wrote
>
> $var = ".rodata\|.data";
> system("objdump -r *.o|cut -f3 -d ' ' |grep -v $var");
> i know that we'll have to remove statements like :
> test1.o:     file format elf32-i386
>
> RELOCATION RECORDS FOR [.text]:
> OFFSET   TYPE              VALUE
>  from the output so that cut works coz it otherwise considers FOR as
> the third field.
>
> ---------------------------------------------
> [shell prompt]$ objdump -r *.o
>
> test1.o:     file format elf32-i386
>
> RELOCATION RECORDS FOR [.text]:
> OFFSET   TYPE              VALUE
> 0000000a R_386_32          .rodata
> 0000000f R_386_PC32        printf
> 00000022 R_386_32          .rodata
> 00000027 R_386_PC32        printf
> 00000041 R_386_PC32        G
> 00000046 R_386_PC32        F
> 0000004e R_386_32          .rodata
> 00000053 R_386_PC32        printf
>
>
>
> test.o:     file format elf32-i386
>
> RELOCATION RECORDS FOR [.text]:
> OFFSET   TYPE              VALUE
> 0000000a R_386_32          .rodata
> 0000000f R_386_PC32        printf
> 00000022 R_386_32          .rodata
> 00000027 R_386_PC32        printf
> 0000003a R_386_32          .rodata
> 0000003f R_386_PC32        printf
> 00000059 R_386_PC32        F1
> 0000005e R_386_PC32        F2
>

You might try using tr to compress the spaces out

... | tr -s " " | cut -f3 -d " " | sort | uniq

--larry

[ Post a follow-up to this message ]

Thanks a lot !!! 
[shell prompt]$ objdump -r *.o | grep -v "RELOCATION\|OFFSET\|file
format"
| grep -v ".rodata\|.data"|tr -s " "|cut -f3 -d " "|sort|uniq

by this i m getting :
F
F1
F2
G
printf

now only 1 thing remains how to use this inside a PERL script...

[ Post a follow-up to this message ]

"jeniffer" <zenith.of.perfection@gmail.com> writes:

> Thanks a lot !!! 
> [shell prompt]$ objdump -r *.o | grep -v "RELOCATION\|OFFSET\|file
> format"
>  | grep -v ".rodata\|.data"|tr -s " "|cut -f3 -d " "|sort|uniq
>
> by this i m getting :
> F
> F1
> F2
> G
> printf
>
> now only 1 thing remains how to use this inside a PERL script...

Perl already has much more powerful tools built in:

my %values;
open OD, "objdump -r *.o |" or die;
while(<OD> ){
/^[[:xdigit:]]+\s+\w+\s+(\S+)$/ or next;
next if $1 =~ '\.rodata' or $1 =~ '\.data';
$values{$1} = 1;
}
close OD;

$, = $\ = "\n";
print sort keys %values;

--
Måns Rullgård
mru@inprovide.com

[ Post a follow-up to this message ]

i appreciate ur way but i wud have to do thru the system command only
in PERL coz that wud be easy n more understandable for me ...thnx a lot

[ Post a follow-up to this message ]