nixCraft

sparky · #1 (**permalink**) 04-16-2006, 09:17 PM

I am trying to change a line from a file
the line reads

Code:

URL=

it may have web url after it it may not .
Now i need to add a url to that line all one words the problem is it need to
start http://

Code:

URL=http://www.whatever.com

now it works fine if i dont add http:// to $url
but if i do i get error and file is deleted

Code:

sed: -e expression #1, char 48: unknown option to `s'

My code

Code:

 isp=$(cat file.txt | grep URL=*)


echo -n "Please enter new url"
echo "please include http:// in your address "
echo "EXAMPLE  http://noobtutorials.co.uk"
read url
URL="URL="$url""

sed -e "s/$isp/\"$URL\"/g"   file.txt > file1

mv file1  file.txt
echo -e "New url link is $url "
echo -e "If you do not see http:// before www. please try again"
cat server.ini  | grep URL=*

Thanks

monk · #2 (**permalink**) 04-17-2006, 12:31 AM

In string http:// double slash (//) confuses sed, as they are part of sed syntax. You need to replace // with \/\/ to disable its special meaning. For example:

Code:

URL="http://old.com"
olddomain="old.com"
newdomain="new.com"
echo $URL | sed "/http:\/\//s/$olddomain/$newdomain/g"

Modify your script and it should work now

sparky · #3 (**permalink**) 04-17-2006, 06:54 AM

Thanks for the reply but still can t get it working tried for hours now

i think the problem is i need it to work for both options
ie the file may say
URL=http://www.whatever.com
or it may just be blank
URL=

Also while i am here i need to find out where a program exists
so i found the find command works well but i need to turn it into a var
whats the best way to remove the . at the begining

Code:

[sparky@localhost ~]$ find -name test
./MyScripts/test

Any ways i am really loving linux scripting have more pleasure out of this than any thing else ive done on my pc
even ordered some books

nixcraft · #4 (**permalink**) 04-17-2006, 12:31 PM

I think your approach is bit wrong here to solve problem. Anyways I got working script for you, replace your existing code with following script:

Code:

#!/bin/sh
# mod() is function
# must be at top of script
mod(){
NCF="$CF.$$"
[ ! -f $CF ] && touch $CF || :
V="$1"
grep -v -E "^$V=" $CF > $NCF
mv -f $NCF $CF
eval echo "$V=\"'\${$V}'\"">> $CF

}

# config file
CF="file.txt"
URL=""

echo -n "Please enter new url"
echo "please include http:// in your address "
echo "EXAMPLE  http://noobtutorials.co.uk"
read URL
# now change value for url
mod URL

echo -e "New url link is $URL "
echo -e "If you do not see http:// before www. please try again"

Variable CF stores file name, which you want to modify
Variable URL stores actual URL you want to setup

So script will ask for input i.e. new URL
Then function mod is called whice opens file stored in CF variable and replace or update URL it may have prefix http:// or without prefix http://

Quote:

whats the best way to remove the . at the begining,

Use any one of the following (cut is easy to use and small code):

Code:

find -name test  | cut -d. -f2  
find -name test  | sed -e 's/\.//g'
find -name test | awk -F'.' '{ print $2} '

Quote:

Any ways i am really loving linux scripting have more pleasure out of this than any thing else ive done on my pc even ordered some books

Same here

Books are good, if you want to master programming, all you need to do is see good code (script

and practice it.

sparky · #5 (**permalink**) 04-17-2006, 04:48 PM

Thanks nixcraft
I got a lot of reading to do before i understand that code in the mod function
still got two problems
you code works GREAT except it had single quotes either side of the new URl

any ideas
thanks

nixcraft · #6 (**permalink**) 04-17-2006, 05:28 PM

If you don't wanna ' (single quote) then remove it, i.e. replace following line in mod()

Code:

eval echo "$V=\"'\${$V}'\"">> $CF

with

Code:

eval echo "$V=\"\${$V}\"">> $CF

sweta · #7 (**permalink**) 04-17-2006, 05:42 PM

I must agree with sparky the code for mod function is high funda code. I just run on my system and I had no clue how it works

I think nixcraft has written it using all shell-scripting short forms that makes to read it difficult but may works very fast (small code works fast after all)

I guess sparky’s problem is almost solved and it would be great if nixCraft explains us what is cooking inside mod() function?

sparky · #8 (**permalink**) 05-21-2006, 05:48 PM

even thou i don t understand it work great
But i still have one problems to iron out.
IT adds to URL=http:/www but it places at bottom of text doc how can i then move it back to line 17

rockdalinux · #9 (**permalink**) 05-22-2006, 02:13 AM

Ha ha

Old post just got some life again... entier stuff is overkilling, IMPO

calum · #10 (**permalink**) 07-13-2006, 10:21 PM

i had a similar problem with extracting lines from xml files. the solution is to change the delimiter that sed uses from the normal / to @ or any of following %,;:

in my case i was looking to remove the 3 characters from the end of a line in an xml file. The characters are: "/>
also note that i had to change to single quotes to ensure the characters are interpreted literally.

original failed implementation

Code:

grep include $filename | sed "s/"/>//g"

becomes:

Code:

grep include $filename | sed 's@"/>@@g'

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Open Source Trouble Ticket Application w/ Active Directory Integration	Johnny Utah	Linux software	3	03-05-2008 01:30 PM
oops proxy server on debian	maroon	Getting started tutorials	3	09-25-2007 06:37 PM

nixCraft

More trouble with sed :oops:

More trouble with sed :oops:

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)

Similar Threads