Day 5 is online. It's about how to rock out with your friend, xargs(1).

day 5; xargs

I've been on the fence about getting a new system for many many months now. I finally caved. It was a bit more than I wanted to pay, but it'll be worth it. It'll be running vmware server and a pile of virtual machines, all in the name of doing more with less. Stats:

Two dual core Opteron 270, 4GB ram, tyan thunder k8we, three 500gb sata 3gb/s drives (who doesn't need 1.5T of storage?), and a fancy antec case.

It's been 6+ years since I've purchased computer hardware, so there was much learning to be had. Since I intend on running vmware server on this, I really wanted one of the new Intel chips that had virtualization. I also wanted two dual core processors. So, I found a nice Xeon 3000 that suited my price and power needs, but it turns out nobody makes dual-socket motherboards for Xeon 3000-series. Intel's website explicitly says thes are single-processor systems only. Sigh. Stupid Intel, they'd have my money now if they didn't make that (bad) decision.

So, the next chip up is the Xeon 5000-series. These processors are more than twice the cost of a Xeon 3000-series, and are very not worth it to me just to have virtualization. That pushed me towards AMD. I looked at benchmarks and features and ended uplooking at AMD's dual core Opteron line. Opteron 275's were at my price point, but the best price was at Newegg and they were sold out. Opteron 270's were $30ish cheaper, and only 200mhz slower, so I bought two of them. If I'm feeling adventurous, I'll overclock them and save myself time and $60, right? ;)

I'm really looking forward to having some new hardware. Wee!

On another note, I surfed the internets looking for items of need, reviews, and prices. Everything I found was either only $5 cheaper than newegg, or the cheapest was at newegg. I heart newegg. 6ish years ago, that's where I bought my 1.4gHz, then-nearly-top-of-the-line system. I'm glad to see they're still around and doing quite well.

I've had to buildworld on a test machine atleast 4 times today. After the 2nd time, I got fed up and installed ccache to make the build go faster. I'm on the 4th buildworld now, and it is running much faster now.

ccache in FreeBSD, when installed from ports, comes with examples on how to get ccache to work with src and ports building.Copy and paste, and you're ready to go. This tool rocks.

dev# ccache -s
cache directory                     /root/.ccache
cache hit                           7868
cache miss                         20433

So far so good. Most of the cache misses were from the first buildworld with ccache. This dev box is running in vmware on a very, very cheap Dell workstation, and it's not the fastest thing to run vmware, so any speed boosts will help.

Also, day 5 of the week of unix is coming tomorrow. I'm slowly finishing it.

Day 4 is finally ready for consumption, a bit late ;)

This article touches: cat, nc, ssh, openssl, GET, wget, w3m, and others. It's designed to show you a pile of tools you can use to pull data from various places.

day 4; data sources

I'm tired of searching through my piles of open tabs in firefox using only my eyes. Make the computer do the hard work! To fix this, I wrote an extension that will let you quickly search (unobtrusively!) your open tabs. Super useful. Check out the project page:

View the screencast demo

Visit the firefox-tabsearch project page

Day 3 is ready for viewing. It's about awk.

This article has lots of usage examples for the many ways you can use awk to do hard work for you. Check out the article here:

day 3: awk

Sorry folks. I've spent most of today with the pager going off dealing with work, so I didn't get around to writing up today's tool. It'll show up soon!

On another note, I don't know where I'd be without xpathtool. I will never again parse web pages with sed, w3m, and awk.

Day 2 is ready for viewing. It's about cut and paste.

Candice suggested these two tools are underutilized and are very useful when you need the features they provide.

Check out the article here:

day 2: cut and paste

Day 1 is ready for viewing. It's about sed, something I feel many sysadmins (and others) neglect in favor of perl, awk, or other tools. It's a super useful tool. Check out the article here:

Day 1: sed

This past week of work had me stretching my piping-ninja skills to the limits. In the past 4 days, I have created a oneliner that invoked xargs 3 times in a series of piped expressions, a oneliner that had xargs calling xargs, a oneliner that was well over 1000 characters and invoked sed, xargs, grep, perl, awk, xpathtool and others, sometimes twice, and several other ninja-like uses of filter-fu involving unix pipes. It's been a fun week for oneliners.

So, I thought I might spend the next 7 days covering some of the tools I find myself working with almost every day. I hope to cover all the ways I use a given tool, when it should be used over any alternatives, things to look up in the manpage, etc.

I'll try to cover more than one tool per day, because some tools can be batched together because there isn't much to day other than "this is useful for foo, bar, anz baz".

Tentative, incomplete list:

data sources

cat, GET, echo, nc (netcat), shell command substitution

filters

cut, grep, paste, sed, sort, tr, uniq, wc

magical

xargs, sh, awk, perl (for oneliners)

Here's a sneak peek at what you might learn:

'sort' can sort lines, right? But how do you sort words on the same line?

% echo "one two three four five" \
  | xargs -n1 | sort | xargs
five four one three two

Some of you might notice that the xargs(1) invocations can be replaced with tr(1). Yes. I use xargs(1) because it's less typing and handles many case tr(1) won't. This is not to say tr(1) doesn't have it's wonderful uses. More on that later this week ;)

If you have suggestions, let me know.