Author: Jude Pereira

Posted on

Contributing to Go in 54 days

With absolutely zero knowledge of Go 54 days ago, I decided to contribute to the Go project. Why? Put simply, I was bored. The thrill of learning something new, and contributing to a massive OSS project like Go caught my attention.

How?

  1. Find an issue that’s tagged as HelpWanted.
    1. There’s a “HelpWanted” tag, which is applied to issues where the Go community is looking for somebody on the outside to fix. I found one such a issue, #21216 with the topic being x/build/cmd/cl: build broken. This seemed a great place to start.
  2. Go through their Contribution Guide.
  3. Although I skipped this part at first, the commenting guide.
    1. I split the issue at hand into two parts, one that provided the resource, and the other to actually fix the reported issue.
    2. On my very first CL (change list), my commenting style varied greatly. I was asked to review the commenting guide. Read it. Seriously, read it.
  4. A must read before starting, Effective Go.
  5. Take a tour of it, in A Tour of Go.
  6. Use Gogland (I love JetBrains for their outstanding IDEs).

Learning Go from scratch was a fairly simple task. It’s just a new syntax, nothing more. Moreover, there’s always Stack Overflow to help you out. Think of SO as a passive mentor, who gives you advice when it’s asked.

I’ve got to thank a couple of people who helped me along the path, @kevinburke, @bradfitz and @andybons. They reviewed my code, and gave my changes a +2, and submitted them.

What does it feel like?

It feels like the first time you try to dive into a swimming pool. You don’t know whether you can do it, but you do it nevertheless. Getting my first two CLs accepted was a little challenging, but definitely enthralling. Talking to other like-minded people across the globe, committed to fixing issues and innovating, is a completely new experience to me. I’m now set on a path to contribute to Go, as it’s a fun weekend exercise, and moreover, just because I can.

Posted on

LetsTuneup: A music chart with Arjit Singh in the lead

LetsTuneup has grown tremendously, and with it, we’ve introduced new features too. We identified that a few of our users couldn’t use the app to it’s full extent because they didn’t have music on their devices.

We’ve solved that. Users can now pick their favourite artists, powered by a location aware scoring algorithm, which recommends popular artists in their area.

Leading the recommendation list in Mumbai is Arjit Singh, followed by Eminem, Linkin Park, Coldplay and Pink Floyd. Honey Singh is #11 on the chart, and some nostalgic users love Akon, making him #28.

Arjit Singh in the lead, with Eminem, Linking Park, Coldplay and Pink Floyd following close
Arjit Singh in the lead, with Eminem, Linkin Park, Coldplay and Pink Floyd following close

Stay tuned and look forward to our next big feature, very soon.

Posted on

US toy giant threatens to sue Matchbox

TL;DR: A giant American toy company threatens to sue Matchbox. Having no resources at hand, Matchbox is forced to change it’s name to Tuneup.

On the 14th of March, I received an email from Apple (via the giant toy company) with the following content:

The developer of the reported application is using the registered
Matchbox trademark in the description of the application without
authorization from the right holder. Also he is offering services by
unlawfully using the registered trademark which is causing bad
influence for the Matchbox brand values. Furthermore, this application is
not authorized from the right holder. This is a trademark infringement
causing damages to our client xxxxxxxx, the right holder for the
Matchbox intellectual property. I would kindly ask you to proceed and
remove the infringement/application from all stores worldwide.

This company, had the trademark for Matchbox registered under the following classification codes: US 001 002 003 005 021 022 023 026 029 036 037 038 039 041. Each and every G & S description was related to toy products, fabrication, stationery products, and manufacturing processes.

Looking up these classification codes on the United States Patent and Trademark Office’s website, absolutely none of them are for computer software, or anything that even crosses paths with Matchbox.

 

The million dollar question: If there was absolutely no case for “confusingly similar”, why did Matchbox change it’s name?

It’s simple really. Big guys always bully the small ones. They threatened to sue us if we didn’t stop using it. Even after proving to them that it wasn’t even remotely infringing to their use of Matchbox, they wouldn’t budge. Had I fought them legally, I would’ve won. Easy peasy. However, due to the lack of resources, I had no choice but to change it’s name.

If I had to call Matchbox anything other than Matchbox, I’d call it Tuneup. “Tuneup” is imagined by Joelle Fernandes, the co-founder of Let’s Tuneup.

Posted on

Why Matchbox, and how it connects people through music

There’s no doubt that music defines us. It influences our moods, for example, making us happy by releasing a chemical named dopamine. It can affect what we wear, what we eat, and perhaps even who we enjoy being together with. It affects our thought process too (it’s well known that ambient noise can improve productivity).

In a study conducted amongst couples who were eighteen years old, it’s been found to predict personality traits. According to the same study, it’s what we’re most likely to discuss about when we meet somebody new, within the first few weeks. Psychologically, men and women who listen to similar music tend to be better communicators, and have longer lasting relationships.

It’s probably one of the most important things in our lives. If I were to place music on Maslow’s hierarchy of needs, I’d place it at the physiological stage. It’s a fundamental part of our society. Even the Hollywood movie directors (e.g. the scene from Interstellar) would agree.

Why not extend this to the social discovery apps we use today? None of them base their core on this. One of the most popular apps for social discovery, Tinder, uses Facebook page likes and interests, to match people together.

This is why Matchbox was created. It bridges the gap between “truly anonymous“, and “hey there“. The app shows you the top ten artists that are common between you and the person you’re looking at, giving you a fair knowledge of what that person would be like:

Matchbox showing the top 10 artists
Matchbox showing the top 10 artists

You’re more likely to be at ease knowing that the opposite person is a little similar to you. Matchbox was crafted with the sole intention that music is the key that connects us, and binds us together. It has evolved for over 9 months, before being made available to the world.

As it stands right now, Matchbox has a hundred active users, and is growing slowly.

Go ahead and test drive the app, and see for yourself how Matchbox re-defines the social discovery platform.

Download on the App StoreGet it on Google Play
Posted on

A letter to Tim Cook regarding iOS 10, with love for Apple

Tim, iOS has always been known for it’s user interface, until recently. Your new take on notifications have not only made them very loud, and unsettling, but do not flow with the entire look and feel of Apple:

No Mr. Tim, no. My notifications will not be waiting for me. They're crying out for my attention.

No Mr. Tim, no. My notifications aren’t waiting for me. They’re crying out for my attention.
 

See that white background? While you’ve done a good job of highlighting the content, it breaks continuity. There was nothing wrong to start with. The way notifications used to render on my iPhone was simply perfect! The content did not need to be highlighted the way you’ve done so in iOS 10. I really loved iOS for not cluttering my life, and making it simpler, compared to Android (the new style does look like inspiration from Android, doesn’t it?).

Comparing this to iOS 9, what you’ve lost is absolute ingenuity:

Yes Tim, the beauty of your beloved platform has been lost, and Steve is waiting for you.

See? Nothing was ever wrong.
 

I’ve been using the beta version of iOS 10 for a little more than a few hours now, and the Today screen has also been ruined. Oh! I almost forgot – what you’ve done to the control centre is horrible. Do you know what this reminds me of? It reminds me of the Red Wedding – Lord Bolton murdering the King of the North, Robb Stark. That was truly gruesome, wasn’t it?

Tim, iOS 9 was the epitome of creation.

I truly hope that this letter talks solely about iOS 10 preview, and that it doesn’t speak for iOS 10 final.

 

Sincerely,
Jude

Posted on

Compile LESS on the fly for your exploded WAR in IntelliJ

At CleverTap, we’ve recently started using LESS for dynamic CSS. While it has it’s upsides, the biggest downside was that most of our developers couldn’t use the hot deploy feature for their local deployments.

After an hour or so, we came up with a neat solution.

 

There are two parts to this:

  1. Just before deploying the app into the web container, compile all the LESS files within the exploded artifact output directory
  2. Have the File Watcher plugin re-compile a modified LESS file within the IDE, and copy it over to the artifact output directory

Both parts above utilize a bash script (since everybody uses a Mac for development, it’s cool).

Prerequisites:

  1. The LESS compiler – can be installed using npm (npm install -g less). If you don’t have the Node Package Manager, just search on how to install it (most likely you’d use Homebrew)
  2. Install the File Watcher plugin in IntelliJ
    1. Go to Preferences in IDEA, then to Plugins
    2. Hit the “Install JetBrains plugin…” button, and search for “file watchers”.
    3. Install the plugin and restart the IDE
  3. A run configuration that is configured to deploy an exploded WAR (can be either Tomcat/Jetty/anything)
  4. Knowing where your exploded artifact resides (in my case, it is /Users/jude/developer/WizRocket/out/artifacts/Dashboard_war_exploded). If you don’t know how to get this, follow these steps:
    1. Go to File -> Project Structure
    2. Click on Artifacts (in the left menu)
    3. Select your exploded WAR artifact
    4. On the right, you’ll see the output directory

Part 1: Compile the LESS into CSS just before deployment

Copy the following script and save it as /Users/username/bin/lessc-idea:


#!/bin/bash
exploded_artifact_path=/Users/jude/developer/WizRocket/out/artifacts/Dashboard_war_exploded
less=/usr/local/bin/lessc
function update {
target=`echo $1 | sed s/web\\// | sed s/.less/.css/`
echo "Generating $exploded_artifact_path/$target"
$less $1 $exploded_artifact_path/$target
}
function all {
find $exploded_artifact_path -name *.less | while read path; do
output=`echo $path | sed s/.less/.css/`
echo "Generating $output"
$less $path $output
done
}
$1 $2

view raw

lessc-idea.sh

hosted with ❤ by GitHub

Note: You will need to update the variable exploded_artifact_path in the script above.

Make it executable:

$ chmod +x /Users/username/bin/lessc-idea

Now, open up your run configuration, and scroll all the way to the bottom (where it says Make, followed by Build artifact …). Hit the “+” button, and select “Run External Tool”.

Hit the “+” button to add a new External Tool, and configure it as follows:

External Tool configuration for compiling LESS files before deployment

Ensure that the build order in your run configuration is as follows:

Build order for LESS compilation

Once this is done, your LESS files should be automatically generated when you deploy your web app. Go ahead and give it a shot.

 

Part 2: Configure the File Watcher plugin to re-compile LESS files edited:

Go to Preferences, and navigate to File Watchers under Tools (left menu). Hit the “+” button and select “Less”.

Configure your new watcher as shown in the screenshot below:

File Watcher configuration for LESS files

Before your hit the OK button, a few things to do:

  1. Clear any output filters added automatically: Press the Output Filters… button, and remove anything inside there.
  2. Select your scope: Select the CSS/LESS directory within your web module (ensure you click on Include Recursively after you’ve selected the directory)

You’re all set. Hit OK, then Apply, and OK.

Test drive your new setup. The moment you change a LESS file, it’ll get re-compiled into the corresponding CSS file within the corresponding directory in the artifact output, and you’ll be able to see the changes immediately.

Posted on

Sending notifications via Apple’s new HTTP/2 API (using Jetty 9.3.6)

HTTP/2 is still very much new to Java, and as such, there are just two libraries who support it – Jetty (from 9.3), and Netty (in alpha). If you’re going the Jetty way (as I have), you’ll need to add their ALPN library to your boot classpath.

Note: Jetty 9.3.x requires the use of Java 8.

A full library for this is available here, on GitHub.

Here’s a quick example:


package com.judepereira.jetty.apns.http2;
import org.eclipse.jetty.client.HttpClient;
import org.eclipse.jetty.client.api.ContentResponse;
import org.eclipse.jetty.client.api.Request;
import org.eclipse.jetty.client.util.StringContentProvider;
import org.eclipse.jetty.http2.client.HTTP2Client;
import org.eclipse.jetty.http2.client.http.HttpClientTransportOverHTTP2;
import org.eclipse.jetty.util.ssl.SslContextFactory;
import java.io.FileInputStream;
import java.security.KeyStore;
public class Main {
public static void main(String[] args) throws Exception {
HTTP2Client http2Client = new HTTP2Client();
http2Client.start();
KeyStore ks = KeyStore.getInstance("PKCS12");
// Ensure that the password is the same as the one used later in setKeyStorePassword()
ks.load(new FileInputStream("MyProductionOrDevelopmentCertificate.p12"), "".toCharArray());
SslContextFactory ssl = new SslContextFactory(true);
ssl.setKeyStore(ks);
ssl.setKeyStorePassword("");
HttpClient client = new HttpClient(new HttpClientTransportOverHTTP2(http2Client), ssl);
client.start();
// Change the API endpoint to api.development.push.apple.com if you're using a development certificate
Request req = client.POST("https://api.push.apple.com")
// Update your :path "/3/device/<your token>"
.path("/3/device/b2482deaf55521b2ccd755d5817a39784cc0044e24s3523a4708c2fa08983bdf")
.content(new StringContentProvider("{ \"aps\" : { \"alert\" : \"Hello\" } }"));
ContentResponse response = req.send();
System.out.println("response code: " + response.getStatus());
// The response body is empty for successful requests
System.out.println("response body: " + response.getContentAsString());
}
}

view raw

Main.java

hosted with ❤ by GitHub

Posted on

openFrameworks and AppCode

Developing an openFrameworks app with AppCode is pretty easy. However, if you just open and run the project created by the project generator, you might see the following errors:

Building a stock openFrameworks app results in these errors
Building a stock openFrameworks app results in these errors

Why doesn’t it just work?

This is because openFrameworks doesn’t support 64 bit builds yet on the Mac, due to a dependency on the deprecated QT framework. More on that here.

What’s the quick fix?

Set your project’s architecture to i386 (32 bit) in it’s build settings:

Ensure that you set both, your project's architecture, as well as openFrameworks' architecture to i386
Ensure that you set both, your project’s architecture, as well as openFrameworks’ architecture to i386

Once you’ve done this, your run configurations should shortly say 32 bit Intel instead of 64 bit Intel:

Run configurations now say 32 bit. Yay!
Run configurations now say 32 bit. Yay!

Kudos! Run your project now, and it will work right out of the box!

Posted on

OpenWRT won’t bring my WiFi interface up, unless the other is up

I recently bought a D-Link DIR 505 router. So far, I’ve got a DLNA server running on it, along with Transmission, a bit torrent client. Life is awesome so far.

I set it up to repeat another WiFi router in my house, the one connected to the internet – using a bridge. It works really well right now.

However, when that WiFi network is down, even the second WiFi network created by my new router won’t come up. I don’t know why as of yet, but I have a dirty hack – if within 10 seconds after boot the router cannot ping my other WiFi router, I will disable that interface and restart the WiFi. This brings the network created by the D-Link up, and I can continue to stream stuff off my hard drive from it.

I created an executable shell script, and placed it in /usr/bin. Then I added a link in rc.local, which is executed after the system is up:

# /usr/bin/wifi_failsafe.sh

logger "Waiting for 10 seconds for network to settle down"
sleep 10

if uci get wireless.@wifi-iface[0].disabled | grep 1; then
    logger "Primary interface is disabled"
else
    logger "Primary interface hasn't been disabled"
    logger "Checking for connectivity"
    if ping -c 1 192.168.2.1; then
        logger "Connectivity has been established"
    else
        logger "Connectivity lost. Disabling primary WiFi interface"
        uci set wireless.@wifi-iface[0].disabled=1
        uci commit wireless
        wifi
    fi
fi
Posted on

Discover – my second iOS app

iTunes Genius is a great feature. However, it lacks music discovery outside your own music library. Sure, you can always do a Google search for similar tunes, but let’s face it – who has time to do this anymore?

There weren’t any great music discovery apps on the App Store either. All of them either looked ugly, or had to be opened by the user. The content wasn’t available readily.

Then I thought of Discover. I wrote this app keeping in mind that the app would never have to be opened by the user, to see any content. Instead, why not present it in the Today screen itself? This way, the widget can refresh it’s content quickly and present it, in a beautiful manner.

Unobtrusive. Simply genius, isn’t it?

How can this be made any better? Provide buttons which directly search the iTunes Store or YouTube for the song recommended. This way, it’s easy for the user to try out new songs, with zero effort. Eureka! The effort of typing on the device is now gone!

I’ve submitted the app on the App Store for review, and I hope it will be accepted and published soon. Here’s a sneak peak of it:

Highlights of Discover
Highlights of Discover

Discover took a total of one month to complete. Although it was a simple app, I couldn’t give it much time day to day.

I love what it’s turned into. There’s so much that I’ve learned about iOS – auto layout, GCD, and the language itself.

Posted on

0f > Float.MIN_VALUE = false!

Updates:
Turns out that this is the expected behaviour from the java doc:
A constant holding the smallest positive nonzero value of type float, 2-149.

So now, how do I get the smallest negative value that a float can hold?

Just came across the most weirdest thing ever in Java – 0f > Float.MIN_VALUE returns false!

Similarly, anything less than 0, (say -10 for example) is surprisingly not greater than Float.MIN_VALUE.

However, 0 > Integer.MIN_VALUE returns true!

Want to try it out?

public class E {
    public static void main(String[] args) {
        System.out.println("1f > Float.MIN_VALUE: " + (1f > Float.MIN_VALUE));
        System.out.println("0f > Float.MIN_VALUE: " + (0f > Float.MIN_VALUE));
        System.out.println("-1f > Float.MIN_VALUE: " + (-1f > Float.MIN_VALUE));
        System.out.println("0f < Float.MIN_VALUE: " + (0f < Float.MIN_VALUE));
        System.out.println("0f == Float.MIN_VALUE: " + (0f == Float.MIN_VALUE));
    }
}

Result:

1f > Float.MIN_VALUE: true
0f > Float.MIN_VALUE: false
-1f > Float.MIN_VALUE: false
0f < Float.MIN_VALUE: true
0f == Float.MIN_VALUE: false