Need some help with sed

[biggyL]

Hello All,

Could you please give me some quick help here?

I've got this file:

[cwr]

You could do it all with sed, but I don't have a reference by me to check the exact
pattern match. The quick and dirty way is "grep verbose input.txt >output text",
and then use vim (or sed) to trim the lines in output.txt to the exact data required.

Will

[biggyL]

Sorry man I need this code to run inside a script to do it on the fly, so your solution is out of the question

Any other idea?

[ppurka]

[think4urs11]

[cwr]

OK, try:

sed -e "s/^.*=//" -e "s/.[0-9].*//" infile > outfile

It's often easier, I find, to take two cuts at the more complicated expressions.

Will

[ppurka]

[think4urs11]

[cwr]

Yes, I hadn't realised there might be digits in the name. However, full sed parsing
isn't hard if you have the docs by you (I was trying to do it from memory) and so
on the whole I'd use think4urs11's solution. But like most scripts, it needs to be
developed from actual test cases.

Will

[ppurka]

[biggyL]

Think4UrS11

Thank you very much!
It worked like a charm

[truc]

[biggyL]

Sorry to bother you again but it seems that I need additional help

running
# glsa-check -p $(glsa-check -t all)
I got this output

[truc]

keeping the same idea as above, this might work (just tested with your sample output)

[cwr]

Or feed the file through grep to remove the lines you don't want?
(grep -v pattern). It looks as if most of the unwanted lines start
the same way.

Will

[truc]

[biggyL]

[think4urs11]

[biggyL]

Thanks Think4UrS11

But I get:

[truc]

keeping the Think4UrS11 idea, this might work

[biggyL]

[truc]

wild guess: this line is probably echoed in stderr and not stdout..

try this:

[biggyL]

[truc]

just play with glsa until you find how to hide this line! :p

[biggyL]