I have an assignment, and I am looking for some suggestions....

I am given an html file and have to write a script in the BASH shell to parse out some information (IE: an IP address).

My only problem is that sed, grep and all those commands grab the whole line that the pattern is found on. In this case, each line of the html file is very lengthy, there are probably like 3 lines in the whole file.

So using these commands doesn't get me very far. I still have a large block of text stored in a new file, but it does me no good.

Any suggestions?

Thanks

Rez

My suggestion:
Use grep to find the line in the file, then pipe the output from that to sed to remove the tags and other extraneous junk.

How do I tell sed not to delete the full line though? If it finds say &nbsp in the text line it kills the whole line by default to my understanding.

Thanks.

I'm not to keen on regular expression syntax anymoe, but basically tell it to replace everything that meets a regular expression that denotes an HTML tag with nothing.

I was actually thinking of first deleting anything in between < >...

sed 's/<*>//` thefile

I'm not to sure on syntax either unfortuanately, we started learning sed this week. Let me know if this is right...

Thanks

I am doing the same assignment, and I was able to use grep -o [0-9A-Z][0-9A-Z]-go on for 5 more. This works fine, but its the IPs that I can't get. If I do [0-9][0-9][0-9].[0-9] and go on for the rest of the IP search it gives me 255.255.255.255 fine, but it wont give me anything else that has less digits in the IP (24.24.5.22).

My question is, is it possible to put some sort of OR statement between the [0-9]s? for each part of the IP so that it can match [0-9][0-9] for some of them? If I can get this thats all I need.

Thanks.

Rezon :::
I think this should do it for you

open(FILE,"<$ARGV[0]");
while(<FILE>)
{
s/\<[\d\D]*?\>//g;
s/\s+/ /g;
print "$_";
}

If you don't understand anything in this feel free to ask me

suprax::

I think | does that think for you [ a broken pipe | is like an "or"]