Notes on Go, part 2 of ∞

Why aren’t my test running?

The command go test will only pick up tests in the directory that the command is being run in. To have it find tests in sub-folders, try the command go test ./.... This searches recursively through the sub-folders and finds the hiding tests and runs them.

Tests still aren’t running? If you are using godep to manage dependancies, you can try:

godep get

Still getting error messages á la:

somefolder/some_test.go:10:2: cannot find package "github.com/THING/code" in any of:
    /usr/local/Cellar/go/1.5.1/libexec/src/github.com/THING/code (from $GOROOT)
    ~/code/go/src/github.com/THING/code (from $GOPATH)

Try the following series of commands from the root directory of your project.

godep restore
rm -rf Godeps
godep save ./...
godep go test ./...
go test ./...

Stay tuned for an actual explanation of why any of this works when I figure it out!

Notes on Go, part 1 of ∞

When writing tests in Go, there are a few useful ways to get info printed out to the screen.

  • t.log

Inside the test, you can log stuff using t.log. This will only show up if you use the -v flag when running go test.

  • fmt.PrintLn

Inside the actual program itself, you can print things using an fmt.Print command, and running the tests with the -v flag.

Basically, the -v flag is your friend when you want to see things in your go tests!

Free private remote git repos

If you want a remote backup of your private git repos, a la Github but free, there are 3 options I know of.

  1. BitBucket. Has unlimited free private repos, but sets limits on the number of collaborators. Also lacks issue tracking.

  2. GitLab. Has unlimited free private repos, WITH free issue tracking that is very similar to how Github works.

  3. DIY. Host it yourself! Except you have to pay for hosting. So there’s that.

I’m using GitLab for my Ultra Top Secret projects right now. So far so good! Their interface is a little rougher than GitHub, but more than workable.

Tab stops in Microsoft Office 2011 OSX

Apologies for the COMPLETELY DIFFERENT topic, but I spent a lot of time trying to format some stuff, so I’m recording this for next time I want to do this.

Step 1: place the curser near the text you want to align, and double click in the ruler bar area. A small black bent arrow should appear. It will probably point toward the right.

Step 2: double click on the arrow in the ruler bar. A dialog box should appear. Depending on how many tab stops you have, one or more numbers will appear in a list on the left under ‘Tab stop position’. Select the tab stop of interest. In my case, it was the second/last stop. To get the arrow to point toward the left/align the text on the right side of the document, click the ‘right’ selector under ‘Alignment’.

Step 3: Hit save. Drag the arrow to where you want to text to line up. Be happy!

Writing regular expressions in the context of a new programming language

One of the things that is kind of annoying about regular expressions is that every programming language implements them slightly differently. If you can, find someone who can give you the low-down in this new language. Otherwise, you’ll have to stick with googling, which can take a while to figure out what you need. I’ll get you started with a few languages.

General Resources

Two of my favorite and most helpful resources:

  • RegExPlanet.com: lets you test regular expressions in many different languages.
  • Regex Cheat Sheet: has a pretty comprehensive general overview of regex syntax.

Regular Expressions in Ruby

One of the easiest ways to get started with regular expressions in Ruby is via Rubular.com. This site provides a way to test regular expressions against any text, as well as a quick cheat sheet to help. RegExPlanet.com also has a Ruby tester that is in beta.

Regular Expressions in Java

For help with Java, I really like using the tester at RegExPlanet.com. It does two really cool things.

  1. Different Java methods (apparently) use regular expressions differently. RegExPlanet.com shows if and how a regular expression will work with each of the methods.
  2. RegExPlanet.com also provides the ‘Java string’ for use in Java methods. In Java, we have to escape the backslashes with additional backslashes. This can get pretty confusing very quickly, so having RegExPlanet.comgenerate that string for me is very helpful.

Regular Expressions in JavaScript

Using regular expressions in JavaScript was where I really began using the pattern modifiers. If a regular expression is between two forward slashes, pattern modifiers are the letters that come after the last forward slash.

/hello/i

This regular expression will match the word ‘hello’ as well as capitalized ‘Hello’, and all caps ‘HELLO’. It will even match the super fancy ‘hElLo’, if you are into crazy stuff like that. The i modifier tells the regular expression to be case-insensitive.

Another really useful modifier for regular expressions in JavaScript is g, which stands for ‘global’. In JavaScript, regular expressions will find the first time it matches a pattern and then stop. Using the g modifier tells it to find ALL the instances where a string matches the given pattern.

Tips and Tricks: Useful Regular Expressions

I’ve learned a whole lot about regular expressions over the last year and a bit, and there are two ‘phrases’ for lack of a better term that are sort of like the nuclear option or like a blunt object, but will get me most of where I need to go.

Phrase 1: .*?

This is one of the most useful regular expression ‘phrases’ I’ve learned. It selects zero or more of any kind of character, including spaces, but isn’t greedy. (That just means that it’ll stop the first time it sees whatever is put after the question mark). The key limitation of this phrase is that it will not include new lines (though the exact implementation varies from language to language).

Phrase 2: .+?

Related to the first phrase, this one selects one or more of any kind of character, including spaces, but isn’t greedy. This is useful when something has to exist, but can be any number of things. Like the one before it, this (usually) doesn’t include new lines.

Phrase 3: [\s\S]*?

This ‘phrase’ is even more ‘powerful’ than the one above. This one selects zero or more of any kind of character, including all kinds of whitespace (yes, also new lines), but again, isn’t greedy. I used this one a LOT for a while, especially in Java.

Get creative

I was having trouble getting a regular expression to select an section of HTML that was broken up onto different lines irregularly 1. Finally I figured out that I wanted to use \s*? (zero or more whitespace characters, not greedy) to account for the different amounts of spacing, but not to pick up any letters or other characters.


  1. yes, I know regex is not the tool of choices for HTML, but my hands were a bit tied on this project 

Some things I learned about case statements in Ruby

I’ve been working with case statements in Ruby, and I was having a difficult time getting it to do what I wanted.

Things I learned:

  1. Case statements that take an argument after the case (ex: case SOME_VARIABLE) can’t take logic after the when part of the statement.

    Bad:

    case thingie
    when thingie.match(/some regex/)
    # something cool happens
    end
    

    Good:

    case thingie
    when "a perfectly boring string"
    # something cool happens
    end
    
  2. Case statements that do not take an argument after the case CAN take logic after the when

    Good:

    case
    when thingie.match(/some regex/)
    # something cool happens
    end
    

Tips and Tricks: Gem pristine

A while ago, some friends and I were doing some code exploration inside gems, trying to figure out how they worked. All was well and good until we had changed everything and broke the gem. Here are a few great commands for playing around with this stuff, and then resetting it when you are done so everything on your system still works the way it is supposed to.

To open a gem in your default text editor and poke around it’s squishy inner bits:

bundle open GEMNAME

EX:

bundle open rake

Once you are done playing around and just want things to go back to the way they were, there are two good commands to know.

For when you have bundler set up to install gems to a local vendor directory, and that is the gem you were playing with:

bundle exec gem pristine GEMNAME

Or if you want to restore all the possible versions of a gem:

gem pristine GEMNAME

Happy hacking!

Sublime text 2 keyboard shortcuts – delete an entire line

I’ve learned some new shortcuts for Sublime Text 2. I’ve been learning a few vim shortcuts, and I wanted to figure out how to do some of the things I can do in vim in Sublime.

  1. Ctrl + shift + k
    Deletes entire line. Ctrl is the correct key, even for OSX.

  2. Cmd + x
    Cuts the line if you want to move it somewhere. This does not require highlighting anything.

And as a note, many of the usual keyboard shortcuts work this way. Cmd + c will copy the line without needing to select anything.

Tips and Tricks: JSON in IRB or Pry

Trick one: Parse your raw JSON objects when you pull them in.

For our Ruby Gem project, we were pulling down JSON-formatted data from the Wikipedia  API. To successfully work with JSON in IRB, I imported the object with an HTTP gem and then parsed the JSON data with the JSON gem. EX:

require 'json' 
require 'rest_client' 
JSON.parse(RestClient.get <URL>)

Which gives us a nice Hash output that looks like this:

{"query-continue"=&gt;
    {"images"=&gt;
        {"gimcontinue"=&gt;"736|Citizen-Einstein.jpg"}},
    "query"=&gt;
        {"pages"=&gt;
            {"-1"=&gt;{"ns"=&gt;6, "title"=&gt;"File:1919 eclipse positive.jpg", "missing"=&gt;""},
             "-2"=&gt;{"ns"=&gt;6, "title"=&gt;"File:Albert Einstein's exam of maturity grades (color2).jpg", "missing"=&gt;""}}}}

Without parsing, the information was still in JSON format, a thing that looked like:

"{\"query-continue\":{\"images\":{\"gimcontinue\":\"736|Citizen-Einstein.jpg\"}},\"query\":  ...

with extra quotation marks and all the backslashes, and Ruby got pretty cranky about trying to work with that string.

Trick Two: ‘puts’ your JSON

For another project, I was converting hashes into JSON, and I was getting a bit frustrated. I was quite certain that I was converting the hash into JSON correctly, but I kept getting extraneous backslash-escaped quotation marks in my JSON returns like before.

"{\"query-continue\":{\"images\":{\"gimcontinue\":\"736|Citizen-Einstein.jpg\"}}, ...

Thanks to some StackOverflow googling, I realized/remembered that this was because I was directly calling the JSON in the console, instead of puts-ing it from within the script I was running. When I used a puts statement inside the script, I could see that my output was actually formatted correctly, as I expected. Using puts in the console also worked to show me the JSON with it’s correct formatting.

puts <JSON>

instead of

<JSON>