Friday, February 2. 2007More on DSLs: awk, part II
To reward those of you who have struggled through my most recent posts regarding DSLs in general and awk in particular, I give you a short post.
How I'm using awk (in conjunction with other tools) to block the botnet attacking us: I was able to find a regular expression that lets me match attempts by the botnet. So I told my computer to watch the logfile forever, looking for matching lines and blocking the IP address that made the request. Here's what the command line looks like (I've left out the actual grep command so that spammers don't pick up on it). tail -f /var/log/apache2/sites/thehomestarmy.com | grep ' Here, awk is being used for its simplest, but perhaps most popular, feature. It's very intelligent about providing ways to break a piece of text into chunks and making those chunks available for printing. What I've done is asked awk "Please read you input, break it into fields, and print the first field." That first field from my apache logfile format happens to be the IP address of the user requesting the page. That's it. To demonstrate how effective these "little languages" are at solving this kind of task, I went ahead and wrote the same tool in perl and C. Using DSLs (awk, grep, bash): 1 line, 20 words, 156 bytes Perl solution: 20 lines, 62 words, 566 bytes C solution: 367 lines, 926 words, 9718 bytes. DSLs clearly won the day here. Trackbacks
Trackback specific URI for this entry
Comments
Display comments as
(Linear | Threaded)
Mad Giggler on :
You're a very thorough man. The only problem with DSLs is that many of us are lazy and don't want to go to the trouble of learning a new tool for everything. For instance, while I'm familiar with grep, I haven't taken the time to learn awk. :(
|
Items of InterestTrekWest5 Subscriber Count: Syndicate This Blog |
