[Nottingham] Slow bash-ings for file line search

Martin martin at ml1.co.uk
Tue Apr 5 23:44:00 BST 2005

Peter Chang wrote:
> [snipped]
>>         if [ "$id" -ge "0" ]
> I don't understand this line. You start with id=0 and increment it so this
> test is always true.

True (:-))

But not always true. With the MS-headers crap, the id count can start 
from "-2" for example to synchronise the two streams.

> The best way for doing this in awk is to go through the file and store
> lines that are between those markers and then on a transition point, spit
> out your tag and the stored lines.

Fair idea. They could easily be dumped to a tmp file and then appended 
when ready.

I'll see how painful the present hack gets before rewriting.

Could be a good excuse to upgrade to a dual core Opteron (:-))


Martin Lomas
martin at ml1.co.uk

More information about the Nottingham mailing list