Messages in this thread |  | | Date | Mon, 29 Jun 2009 10:51:46 +0800 | From | Amerigo Wang <> | Subject | Re: [PATCH] kconfig: simplification of scripts/extract-ikconfig |
| |
On Wed, Jun 24, 2009 at 03:31:48PM +0200, Dick Streefland wrote: >On Wednesday 2009-06-24 13:46, Dick Streefland wrote: >| Because I don't want to rely on a particular version of grep, I've >| modified the script to use perl instead. I hope that's OK. Below is >| a new patch. ><snip>
Sorry for the delay, I missed this patch...
> >I found a way to work around the differences in behavior of grep. >Older versions of grep report the byte offset of the start of the >line instead of the byte offset of the pattern. By using tr to >make sure the pattern is always at the start of a "line", both old >and new versions of grep can be used.
This is very tricky...
<snip>
>+ >+gz1='\037\213\010' >+gz2='01' >+cf1='IKCFG_ST\037\213\010' >+cf2='0123456789' >+ >+dump_config() > { >- echo " usage: extract-ikconfig [b]zImage_filename" >-} >- >-clean_up() >-{ >- if [ "$TMPFILE" != "" ]; then >- rm -f $TMPFILE >+ if pos=`tr "$cf1\n$cf2" "\n$cf2=" < "$1" | grep -abo "^$cf2"`
Does this work? My quick test shows no...
I still can't get the right offset number with this trick.
P.S. My grep is 2.5.1.
|  |