Some Cool New Tools

I’ve written (and ported) some new tools and thought others might find them useful or inspiring.

In my ongoing project to write little tools in Go I have made a lot of progress lately. One fun little tool is replace-unzip, which reimplements the unzip(1) command but does not extract .DS_Store, __MAXOSX/, and extracts to a new directory if the files in the zipfile don’t already have a common root. This was originally Perl, but I ported it to go because the Archive::Zip module on CPAN is not core perl.

🔗 rss

I’ve been meaning to start reading RSS again since December. One of my goals is not to add new places to check (inboxes) but instead to bend things to my workflow. So instead of firing up an RSS client or finding a hosted one, I wrote a commandline client. It takes the url to check and a filename to store state:

$ rss afm.json
[Announcing shellquote](
[Detecting who used the EC2 metadata server with BCC](
[Centralized known_hosts for ssh](
[Buffered Channels in Golang](
[C, Golang, Perl, and Unix]( 

The state is simply a JSON list of GUIDs that the tool has seen before. I wrote a little vim command that lets me just type :RSS and my new RSS items will just show up as links in the current document. This is built around my notes system, discussed here.

A handy side effect is that because I’m just running a simple program I can easily add more intelligence. For example, I enjoy LWN, but I could not care less about new kernel releases, so I can filter them out trivially:

$ rss .rss/lwn.js |
      grep -Piv 'kernel (update|prepatch)'

Writing this highlighted to me, again, that the Go ecosystem is not mature yet. The RSS modules I found all force the user to implement part of the various specifications instead of doing the obvious work out of the box. Here’s an example: RSS allows the link itself to be treated like the GUID, but none of the modules I found would put the link in the GUID if the GUID was missing. There’s more than just that, but that’s a really simple improvement that would help almost everyone using the module.

🔗 dump-mozlz4

In a similar fashion to the above, I wanted to be able to insert links to all of the current tabs I have open. This assists me in “swapping out” such that the only thing open is stuff I actually am doing at the moment. I spelunked my Firefox profile directory and finally found a file that would have the data I was interested in. Here’s what I found:

$ file sessionstore-backups/previous.jsonlz4
sessionstore-backups/previous.jsonlz4: data

$ xxd sessionstore-backups/previous.jsonlz4 | head -3
00000000: 6d6f 7a4c 7a34 3000 d48e 0300 f221 7b22  mozLz40......!{"
00000010: 7665 7273 696f 6e22 3a5b 2273 6573 7369  version":["sessi
00000020: 6f6e 7265 7374 6f72 6522 2c31 5d2c 2277  onrestore",1],"w

So clearly it’s got some kind of header mozLz40 and then some other bytes before what is presumably lz4 compressed data. I tried using some lz4 compression cli tool but it errored saying incorrect magic number, which I expected. I also figured that the other bytes were probably a length and probably 4 bytes long.

I found a rust tool on github that confirmed my suspicions above. After that it was a matter of research, and I was able to build a mozlz4 package to assist in decompressing these files in the future.

Armed with this package and a very basic CLI wrapper I am now able to get a list of all the pages that are open in Firefox:

$ dump-mozlz4 sessionstore-backups/recovery.jsonlz4 |
   jq -r '.windows[].tabs[] | .entries[.index - 1] | " * ["+.title+"]("+.url+")"'

I found it interesting that I had to use .index in the above; without it you will get the wrong entry when you press back.

I have long thought that one of the super powers of software engineering is that our world is so much more mutable than others, and that we can build tools to solve our problems in ways that simplify our lives and reduce friction. I hope this post inspires you to do the same.

Nearly all of the code in this post is Go. If you want to learn Go, you should read The Go Programming Language. It’s not just a great Go book but a great programming book in general with a generous dollop of concurrency.

Another book that I recommend is The Unix Programming Environment. It goes through the motions of creating wrappers tools and tools afresh, diving into some of the operating system details that assist the toolmaker so much.

Posted Tue, Jul 17, 2018

Receive Blog Posts in Your Email