Performing pattern match for a string that might be intermingle with other strings

Shell Programming and Scripting
1

I have a log file that display the serial output coming from different places. Sometime the string in search gets clobbered with the other strings and consequently change form. For example:

serial ouput:
--------------

hello world!
done with network
configuring asic registers
comJan 1 00:00:11 network: checksum passed
plete

---------------

In this case I want to check for "complete" string but since it is clobbered with the network stuff, i cant verify its existence. Is there a way I can use some sort of string matching to detect such strings?

2

If your file is complete.txt

final=`tail -1 complete.txt`

sets the variable final to plete
(assuming your file ends at the complete message??)

cat complete.txt | grep -B 1 $final | head -1

gives the preceeding line (although lots of other ways)

now you are left with how to figure out if complete is spelled on those two.

thought about using

tr -cd [complet]

to get rid of everything but those characters

but am now thinking to count the number of characters on last line (final variable) and subtract from the 8 in complete. and then check to see if that number of characters is how the previous line begins.

3

"Complete" can be anywhere in the file not necessarily at the end of the file... This approach might work but not sure how to implement it: accepting arbitrarily many characters between each char in the pattern c*o*m*p*l*e*t*e