Archive

Archive for the ‘Open Source Software’ Category

CoreGTK 3.10.0 Released!

August 20th, 2015 No comments

The next version of CoreGTK, version 3.10.0, has been tagged for release today.

Highlights for this release:

  • Move from GTK+ 2 to GTK+ 3
  • Prefer the use of glib data types over boxed OpenStep/Cocoa objects (i.e. gint vs NSNumber)
  • Base code generation on GObject Introspection instead of a mix of automated source parsing and manual correction
  • Support for GTK+ 3.10

CoreGTK is an Objective-C language binding for the GTK+ widget toolkit. Like other “core” Objective-C libraries, CoreGTK is designed to be a thin wrapper. CoreGTK is free software, licensed under the GNU LGPL.

You can find more information about the project here and the release itself here.

This post originally appeared on my personal website here.




I am currently running a variety of distributions, primarily Linux Mint 18.
Previously I was running KDE 4.3.3 on top of Fedora 11 (for the first experiment) and KDE 4.6.5 on top of Gentoo (for the second experiment).

Taking a look at some Linux e-mail clients

July 12th, 2015 No comments

Many people now use a browser based solution, like Gmail, for all of their e-mail needs however there are still plenty of reasons why someone might want to use a local e-mail client as well. In this post I’m going to take a look at some of the graphical e-mail client options available on Linux.

Balsa

I have to admin that I hadn’t even heard of Balsa before looking up e-mail clients to include in this list. In my limited time using it Balsa seems to be a relatively simple e-mail client that still offers quite a few options (supports POP3 and IMAP as well as PGP/GPG and even includes a spell checker) while still maintaining a very low memory footprint (less than 7MiB of RAM for an empty inbox). However one thing I couldn’t seem to get working was actually sending e-mail – it’s not that it was difficult to setup, I just simply couldn’t get it to connect to my SMTP server to send the mail. It kept timing out without giving me a cause which was annoying.

Balsa

Balsa

Balsa Project Website

Claws Mail

Similar to Balsa, Claws is also a very lightweight e-mail client that offers quite a few standard features but can also be expanded upon via plugins. Interestingly I couldn’t figure out a way to compose a non-plaintext (i.e. HTML) e-mail so perhaps the developers are of the opinion that e-mail should only be sent as text?

Claws Mail

Claws Mail

Claws Mail Project Website

Evolution

Evolution is/was (depends on who you ask) the golden standard for what an e-mail client on Linux should be. You can think of it as a complete Outlook replacement as it does so much more than just e-mail (contacts, calendar, memos, etc.) all without the need for additional plugins. This does come at a bit of a price as Evolution certainly feels heavier and uses more memory than some other e-mail only clients.

Evolution

Evolution

Evolution Project Website

Geary

Geary is a relative newcomer and has been getting quite a bit of attention as it is included as the default e-mail client in elementaryOS. This application is beautiful however very, very streamlined. You won’t find things like plugins, PGP/GPG, or loads of configuration options here, instead Geary focuses on being the best user experience it can be out of the box.

Geary

Geary

Geary Project Website

GNUMail.app

GNUMail.app is quite a bit different from the other e-mail clients on this list. It is associated with the GNUstep project and runs on both Linux and Mac OS X. Unfortunately while trying to use it on Linux I found myself at a loss… I simply couldn’t figure out how to use the thing! I managed to configure my account settings but could never get it to actually download any e-mail. So without actually being able to use the application I don’t have much else to say about it.

GNUMail.app

GNUMail.app

GNUMail.app Wikipedia Page

KMail

KMail provides the e-mail duties for the Kontact Personal Information Manager collection of software. It is a fully featured e-mail client and, because of the other Kontact applications, offers a compelling pseudo-integrated alternative to something heavy like Evolution. This is especially true if you are using the KDE desktop environment where things feel even more integrated.

KMail

KMail

KMail Project Website

Slypheed

Slypheed and Claws Mail are very similar, which makes sense because they used to be the same project (one was simply a place to try new features before putting it into the “real” project). Even though they share a linage Slypheed and Claws Mail now have different code bases and development teams. That said there aren’t very many obvious differences between the two at this point.

Slypheed

Slypheed

Slypheed Project Website

Thunderbird

Thunderbird is one of the most popular free/open source e-mail clients around and for good reason. It offers a good amount of features and can make use of plugins to add even more functionality. While it may not quite match up to Evolution in terms of advanced functionality for most people, myself included, it works very well.

Thunderbird

Thunderbird

Thunderbird Project Website




I am currently running a variety of distributions, primarily Linux Mint 18.
Previously I was running KDE 4.3.3 on top of Fedora 11 (for the first experiment) and KDE 4.6.5 on top of Gentoo (for the second experiment).

Big distributions, little RAM 8

July 11th, 2015 2 comments

It’s been a while but once again here is the latest instalment of the series of posts where I install the major, full desktop, distributions into a limited hardware machine and report on how they perform. Once again, and like before, I’ve decided to re-run my previous tests this time using the following distributions:

  • Debian 8 (Cinnamon)
  • Debian 8 (GNOME)
  • Debian 8 (KDE)
  • Debian 8 (MATE)
  • Debian 8 (Xfce)
  • Elementary OS 0.3 (Freya)
  • Kubuntu 15.04 (KDE)
  • Linux Mint 17.1 (Cinnamon)
  • Linux Mint 17.1 (KDE)
  • Linux Mint 17.1 (MATE)
  • Linux Mint 17.1 (Xfce)
  • Mageia 4.1 (GNOME)
  • Mageia 4.1 (KDE)
  • OpenSUSE 13.2 (GNOME)
  • OpenSUSE 13.2 (KDE)
  • Ubuntu 15.04 (Unity)
  • Ubuntu Mate (MATE)
  • Xubuntu 15.04 (Xfce)

I also attempted to try and install Fedora 21 and Linux Mint 17.2 (KDE) but it just wouldn’t go.

All of the tests were done within VirtualBox on ‘machines’ with the following specifications:

  • Total RAM: 512MB
  • Hard drive: 8GB
  • CPU type: x86 with PAE/NX
  • Graphics: 3D Acceleration enabled

The tests were all done using VirtualBox 4.3.30, and I did not install VirtualBox tools (although some distributions may have shipped with them). I also left the screen resolution at the default (whatever the distribution chose) and accepted the installation defaults. All tests were run prior to June 2015 so your results may not be identical.

Results

Just as before I have compiled a series of bar graphs to show you how each installation stacks up against one another. Measurements were taken using the free -m command for memory and the df -h command for disk usage.

Like before I have provided the results file as a download so you can see exactly what the numbers were or create your own custom comparisons (see below for link).

Things to know before looking at the graphs

First off if your distribution of choice didn’t appear in the list above its probably not reasonably possible to be installed (i.e. I don’t have hours to compile Gentoo) or I didn’t feel it was mainstream enough (pretty much anything with LXDE). As always feel free to run your own tests and link them in the comments for everyone to see.

First boot memory (RAM) usage

This test was measured on the first startup after finishing a fresh install.

First_Boot_AllMemory (RAM) usage after updates

This test was performed after all updates were installed and a reboot was performed.

After_Updates_AllMemory (RAM) usage change after updates

The net growth or decline in RAM usage after applying all of the updates.

Usage_Changes_AllInstall size after updates

The hard drive space used by the distribution after applying all of the updates.

Install_SizeConclusion

Once again I will leave the conclusions to you. Source data provided below.

Source Data




I am currently running a variety of distributions, primarily Linux Mint 18.
Previously I was running KDE 4.3.3 on top of Fedora 11 (for the first experiment) and KDE 4.6.5 on top of Gentoo (for the second experiment).

CoreGTK now supports GTK+ 3 and is built from GObject Introspection

July 1st, 2015 No comments

It has been quite a while since the first release of CoreGTK back in August 2014 and in that time I’ve received a lot of very good feedback about the project, what people liked and didn’t like, as well as their wishlists for new features. While life has been very busy since then I’ve managed to find a little bit of time here and there to implement many of the changes that people were hoping for. As mentioned in my previous post here are the highlighted changes for this new version of CoreGTK:

Move from GTK+ 2 to GTK+ 3

GTK+ 3 is now the current supported widget toolkit and has been since February 2011. Now that GTK+ 3 is supported on all platforms (Windows, Mac and Linux) it makes sense to move over and take advantage of the updated features.

Additionally this allows for a natural break in compatibility with the previous release of CoreGTK. What that means for the end user is that I currently don’t have any plans on going back and applying any of these new ideas/changes to the old GTK+ 2 version of the code base, instead focusing my time and effort on GTK+ 3.

Prefer the use of glib data types over boxed OpenStep/Cocoa objects (i.e. gint vs NSNumber)

When originally designing CoreGTK I decided to put a stake in the ground and simply always favour OpenStep/Cocoa objects where possible. The hope was that this would allow for easier integration with existing Objective-C code bases. Unfortunately good intentions don’t always work out in the best way. One of the major pieces of feedback I got was to take a less strict approach on this and drop the use of some classes where it makes sense. Specifically keep using NSString instead of C strings but stop using NSNumber in place of primitives like gint (which itself is really just a C int). The net result of this change is far less boilerplate code and faster performance.

So instead of writing this:

/* Sets the default size of the window */
[window setDefaultSizeWithWidth: [NSNumber numberWithInt:400] andHeight: [NSNumber numberWithInt:300]];

you can now simply write this:

/* Sets the default size of the window */
[window setDefaultSizeWithWidth: 400 andHeight: 300];

Base code generation on GObject Introspection instead of a mix of automated source parsing and manual correction

The previous version of CoreGTK was, shall we say, hand crafted. I had written some code to parse header files and generate a basic structure for the Objective-C output but there was still quite a bit of manual work (days/weeks/months) involved to clean up this output and make it what it was. Other than the significant investment in time required to make this happen it was also prone to errors and would require starting back at square one with every new release of GTK+.

This time around the output is generated using GObject Introspection, specifically by parsing the generated GIR file for that library with the new utility CoreGTKGen. The proccess of generating new CoreGTK bindings using CoreGTKGen now takes just a couple of seconds and produces very clean and simple source code files. This is also really just the start as I’m sure there are plenty of improvements that can be made to CoreGTKGen to make it even better! Perhaps equally exciting is that once this process is perfected it should be relatively easy to adapt it to support other GObject Introspection supported libraries like Pango, Gdk, GStreamer, etc.

Let’s have an example shall we?

While there are a couple of good examples over at the Getting Started page of the Wiki and even within the CoreGTK repo itself I figured I would show something different here. It has always been my goal with this project to make it as easy as possible for existing Objective-C users to port their applications to GTK+. Perhaps you were previously using a widget toolkit like Cocoa on the Mac and now you want to release your application on more platforms. What better way than to keep your existing business logic and swap out the GUI (you do practice good MVC right? :P).

So going with this idea here is a tutorial of porting the “Start Developing Mac Apps Today” example from Apple’s developer website here. This application is incredibly simplistic but basically lets you set a “volume” value either by typing in a number in the text box at the top, moving the slider up and down, or pressing the Mute button. Regardless of which action you take the rest of the GUI is updated to match.

Step 1) Setup the GUI

For this I will be using GLADE as a replacement for the Xcode Interface Builder but you could always program your GUI by hand as well.

From the Apple website we are trying to re-create something that looks like this:

apple_exampleThankfully in GLADE this is relatively easy and I was able to do a quick and dirty mock up resulting in this:

glade_mockup

 

Step 2) Configure GUI signals (i.e. events)

GLADE also makes this easy, simply click on the widget, flip over to the Signals tab and type in your handler name.

textEntrySignal

Here are the ones I created:

  • window (GtkWindow)
    • Signal: destroy
    • Handler: endGtkLoop
  • entry (GtkEntry)
    • Signal: changed
    • Handler: takeValueForVolume
  • scale (GtkScale)
    • Signal: value-changed
    • Handler: sliderValueChanged
  • mute_button (GtkButton)
    • Signal: clicked
    • Handler: muteButtonClicked

Step 3) Create classes

Even though Cocoa and GTK+ don’t map exactly the same I decided to follow Apple’s conventions where it made sense just for consistency.

AppDelegate.h

#import "CoreGTK/CGTKEntry.h"
#import "CoreGTK/CGTKScale.h"

#import "Track.h"

@interface AppDelegate : NSObject
{
    CGTKEntry *textField;
    CGTKScale *slider;
    Track *track;
    BOOL updateInProgress;
}

@property (nonatomic, retain) CGTKEntry *textField;
@property (nonatomic, retain) CGTKScale *slider;
@property (nonatomic, retain) Track *track;

/* Callbacks */
-(void)mute;
-(void)sliderChanged;
-(void)takeValueForVolume;

/* Methods */
-(void)updateUserInterface;

-(void)dealloc;

@end

AppDelegate.m

#import "AppDelegate.h"

@implementation AppDelegate

@synthesize textField;
@synthesize slider;
@synthesize track;

/* Callbacks */
-(void)mute
{
    if(!updateInProgress)
    {
        updateInProgress = YES;
        
        [self.track setVolume:0.0];
    
        [self updateUserInterface];
        
        updateInProgress = NO;
    }
}

-(void)sliderChanged
{
    if(!updateInProgress)
    {
        updateInProgress = YES;
        
        [self.track setVolume:[self.slider getValue]];
    
        [self updateUserInterface];
        
        updateInProgress = NO;
    }
}

-(void)takeValueForVolume
{
    NSString *text = [self.textField getText];
    if([text length] == 0)
    {
        return;
    }
    
    if(!updateInProgress)
    {
        updateInProgress = YES;
        
        double newValue = [[self.textField getText] doubleValue];
    
        [self.track setVolume:newValue];
    
        [self updateUserInterface];
        
        updateInProgress = NO;
    }
}

/* Methods */
-(void)updateUserInterface
{
    double volume = [self.track volume];
    
    [self.textField setText:[NSString stringWithFormat:@"%1.0f", volume]];
    
    [self.slider setValue:volume];
}

-(void)dealloc
{
    [textField release];
    [slider release];
    [track release];
    [super dealloc];
}

@end

Track.h

/*
 * Objective-C imports
 */
#import <Foundation/Foundation.h>

@interface Track : NSObject
{
    double volume;
}

@property (assign) double volume;

@end

Track.m

#import "Track.h"

@implementation Track

@synthesize volume;

@end

Step 4) Wire everything up

In order to make everything work, load the GUI from the .glade file, connect the signals to the AppDelegate class, etc. we need some glue code. I’ve placed this all in the main.m file.

main.m

int main(int argc, char *argv[])
{    
    NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init];
    
    /* This is called in all GTK applications. Arguments are parsed
    * from the command line and are returned to the application. */
    [CGTK autoInitWithArgc:argc andArgv:argv];
        
    /* Create a builder to load GLADE file */
    CGTKBuilder *builder = [[CGTKBuilder alloc] init];
    
    if([builder addFromFileWithFilename:@"mac_app.glade" andErr:NULL] == 0)
    {
        NSLog(@"Error loading GUI file");
        return 1;
    }
    
    /* Create an AppDelegate to link to the GUI */
    AppDelegate *appDelegate = [[AppDelegate alloc] init];
    
    /* Get text field, wrapping returned Widget in new CGTKEntry */
    appDelegate.textField = [[[CGTKEntry alloc] initWithGObject:(GObject*)[[CGTKBaseBuilder 
        getWidgetFromBuilder:builder withName:@"entry"] WIDGET]] autorelease];
    
    /* Get slider, wrapping returned Widget in new CGTKScale */
    appDelegate.slider = [[[CGTKScale alloc] initWithGObject:(GObject*)[[CGTKBaseBuilder 
        getWidgetFromBuilder:builder withName:@"scale"] WIDGET]] autorelease];
    
    /* Create track class for AppDelegate */
    Track *track = [[Track alloc] init];
    appDelegate.track = [track autorelease];
    
    /* Pre-synchronize the GUI */
    [appDelegate updateUserInterface];
    
    /* Use signal dictionary to connect GLADE signals to Objective-C code */
    NSDictionary *dic = [[NSDictionary alloc] initWithObjectsAndKeys:
                     [CGTKCallbackData withObject:[CGTK class] 
                         andSEL:@selector(mainQuit)], @"endGtkLoop",
                         
                     [CGTKCallbackData withObject:appDelegate 
                         andSEL:@selector(mute)], @"muteButtonClicked",
                         
                     [CGTKCallbackData withObject:appDelegate 
                         andSEL:@selector(sliderChanged)], @"sliderValueChanged",
                         
                     [CGTKCallbackData withObject:appDelegate 
                         andSEL:@selector(takeValueForVolume)], @"takeValueForVolume",
                     nil];

    /* CGTKBaseBuilder is a helper class to maps GLADE signals to Objective-C code */
    [CGTKBaseBuilder connectSignalsToObjectsWithBuilder:builder andSignalDictionary:dic];
    
    /* Show the GUI */
    [[CGTKBaseBuilder getWidgetFromBuilder:builder withName:@"window"] showAll];
    
    /*
     * Release allocated memory
     */
    [builder release];
            
    /* All GTK applications must have a [CGTK main] call. Control ends here
     * and waits for an event to occur (like a key press or
     * mouse event). */
    [CGTK main];
    
    /*
     * Release allocated memory
     */    
    [appDelegate release];
    [pool release];
    
    // Return success
    return 0;
}

 

Step 5) Compile and run

coregtk_result

So while this is a very basic, quick and dirty example it does prove the point. As for CoreGTK this release is still under development as I try and flush out any remaining bugs but please give it a shot, submit issues or pitch in to help if you’re interested! You can find the CoreGTK project at http://coregtk.org.

Example Source Code
File name: mac_port_example.zip
File hashes: Download Here
License: (LGPL) View Here
File size: 5.3KB
File download: Download Here

This post originally appeared on my personal website here.




I am currently running a variety of distributions, primarily Linux Mint 18.
Previously I was running KDE 4.3.3 on top of Fedora 11 (for the first experiment) and KDE 4.6.5 on top of Gentoo (for the second experiment).

Adding GTK+ 3 support and building CoreGTK using GObject Introspection

May 3rd, 2015 No comments

It has been a while since I made any mention of my side project CoreGTK. I’m sure many people can relate that with life generally being very busy it is often hard to find time to work on hobby projects like this. Thankfully while that certainly has slowed the pace of development it hasn’t stopped it outright and now I am just about ready to show off the next update for CoreGTK.

First off thank you to everyone who took a look at the previous release. I received quite a few nice comments as well as some excellent feedback and hope to address quite a bit of that here. The feedback plus my own ideas of where I wanted to take the project defined the goal for the next release that I am currently working toward.

Goals for this release:

  • Move from GTK+ 2 to GTK+ 3
  • Prefer the use of glib data types over boxed OpenStep/Cocoa objects (i.e. gint vs NSNumber)
  • Base code generation on GObject Introspection instead of a mix of automated source parsing and manual correction

In order to explain the rationale behind these goals I figured I would address each point in more detail.

Move from GTK+ 2 to GTK+ 3

This one was pretty much a no-brainer. GTK+ 3 is now the current supported widget toolkit and has been since February 2011. Previously my choice to use GTK+ 2 was simply due to the fact that I wanted to make it as cross-platform as possible and at the time of release GTK+ 3 was not supported on Windows. Now that this has changed it only makes sense to continue forward using the current standard.

Additionally this allows for a natural break in compatibility with the previous release of CoreGTK. What that means for the end user is that I currently don’t have any plans on going back and applying any of these new ideas/changes to the old GTK+ 2 version of the code base, instead focusing my time and effort on GTK+ 3.

Prefer the use of glib data types over boxed OpenStep/Cocoa objects (i.e. gint vs NSNumber)

When originally designing CoreGTK I decided to put a stake in the ground and simply always favour OpenStep/Cocoa objects where possible. The hope was that this would allow for easier integration with existing Objective-C code bases. Unfortunately good intentions don’t always work out in the best way. One of the major pieces of feedback I got was to take a less strict approach on this and drop the use of some classes where it makes sense. Specifically keep using NSString instead of C strings but stop using NSNumber in place of primitives like gint (which itself is really just a C int). The net result of this change is far less boilerplate code and faster performance.

So instead of writing this:

/* Sets the border width of the window */
[window setBorderWidth: [NSNumber numberWithInt:10]];

you can now simply write this:

/* Sets the border width of the window */
[window setBorderWidth: 10];

Base code generation on GObject Introspection instead of a mix of automated source parsing and manual correction

The previous version of CoreGTK was, shall we say, hand crafted. I had written some code to parse header files and generate a basic structure for the Objective-C output but there was still quite a bit of manual work involved to clean up this output and make it what it was. Other than the significant investment in time required to make this happen it was also prone to errors and would require starting back at square one with every new release of GTK+. This time around the output is generated using GObject Introspection, specifically by parsing the generated GIR file for that library. Currently, and I must stress that there is still quite a bit of room for improvement, this allows me to generate CoreGTK bindings from scratch within an hour or so. With some of the final touches I have in mind the time required for this should hopefully be down to minutes (the auto-generation itself only takes seconds but it isn’t 100% yet). Better still once this process is perfected it should be relatively easy to adapt it to support other GObject Introspection supported libraries like Pango, Gdk, GStreamer, etc.

So where is this new release?

I am getting closer to showing off this new code but first I have to do a bit of cleanup on it. This hopefully won’t take too much longer and to show you how close I am here is a screenshot of CoreGTK running using GTK+ 3.

coregtk-3

This post originally appeared on my personal website here.




I am currently running a variety of distributions, primarily Linux Mint 18.
Previously I was running KDE 4.3.3 on top of Fedora 11 (for the first experiment) and KDE 4.6.5 on top of Gentoo (for the second experiment).

How to easily forward Firefox (PC & Android) traffic through an SSH tunnel

March 29th, 2015 No comments

Say you are travelling, or are at a neighbourhood coffee shop, using whatever unsecured WiFi network they make available. You could either:

  1. trust that no one is sniffing your web traffic, capturing passwords, e-mails, IMs, etc.
  2. trust that no one is using more sophisticated methods to trick you into thinking that you are secure (i.e. man in the middle attack)
  3. route your Internet traffic through a secure tunnel to your home PC before going out onto the web, protecting you from everyone at your current location

which would you choose?

VPNs and SSH tunnels are actually a relatively easy means for you to be more secure while browsing the Internet from potentially dangerous locations.

Making use of an SSH tunnel on your PC

There are many, many different ways for you to do this but I find using a Linux PC that is running on your home network to be the easiest.

Step 1: Install SSH Server

Configure your home Linux PC. Install ssh (and sshd if it is separate). If you are using Ubuntu this is as easy as running the following command: sudo apt-get install ssh

Step 2: Make it easy to connect

Sign up for a free dynamic DNS service like DynDNS or No-IP so that you know of a web address that always points to your home Internet connection. To do this follow the instructions at the service you choose.

Step 3: Connect to tunnel

On your laptop (that you have taken with you to the hotel or coffee shop) connect to your home PC’s ssh server. If you are on Windows you will need to get a program like PuTTY. See their documentation on how to forward ports. On Linux you can simply use the ssh command. The goal is to forward a dynamic port to the remote ssh server. For instance if you are using a Linux laptop and ssh then the command would look something like: ssh -D [dynamic port] [user]@[home server] -p [external port number – if not 22]. An example of one would be ssh -D 4096 user@example.com -p 4000

Step 4: Configure browser to use SSH tunnel proxy

In your browser open the networking options window. This will allow you to tell the browser to forward all of its traffic to a proxy, which in this case, will be our dynamic port that we set up in step 3. Here is an example of my configuration for the example above.
If you don’t feel awesome enough doing the above graphically you can also browse to “about:config” (without quotes) and set the following values:

  • network.proxy.proxy_over_tls
    • true
  • network.proxy.socks
    • Change to “127.0.0.1” with no quotes
  • network.proxy.socks_port
    • Change to the SSH Tunnel Local Port set above (4096)
  • network.proxy.socks_remote_dns
    • Change to true
    • Note: you cannot actually set this setting graphically but it is highly recommended to configure this as well!
  • network.proxy.socks_version
    • Change to 5
  • network.proxy.type
      Change to 1

Step 5: Test and use

Browse normally – you are now browsing the Internet by routing all of your traffic (in Firefox) securely through your home PC. Note that this doesn’t actually make web browsing any more secure beyond protecting you from people in your immediate vicinity (i.e. connected to the same insecure WiFi network).


What about Android?

Just like the PC you can also do it on Android even without root access. Please note that while I’m sure there are a few ways to accomplish this, the following is just one way that has worked for me. I’m also assuming that you already have an SSH server to tunnel your traffic through.

Step 1: Install SSH Tunnel

The first thing you’ll want to do is install an application that will actually create the SSH tunnel for you. One such application is the aptly named SSH Tunnel which can be found on the Google Play Store here.

Step 2: Configure SSH Tunnel

Next you’ll want to launch the application and configure it.

  • Set the Host address (either a real domain name, dynamic DNS redirector or IP address of your SSH server) and port to connect on.
  • You’ll also want to configure the User and Password / Passphrase.
  • Check the box that says Use socks proxy.
  • Configure the Local Port that you’ll connect to your tunnel on (perhaps 1984 for the paranoid?)
  • I would recommend checking Auto Reconnect as well, especially if you are on a really poor WiFi connection like at a hotel or something.
  • Finally check Enable DNS Proxy.

Step 3: Connect SSH Tunnel

To start the SSH tunnel simply check the box that says Tunnel Switch.

Step 4: Install Firefox

While you may have a preference for Google Chrome, Firefox is the browser I’m going to recommend setting up the tunnel with. Additionally this way if you do normally use Chrome you can simply leave Firefox configured to always use the SSH tunnel and only switch to it when you want the additional privacy. Firefox can be found on the Google Play store here.

Step 5: Configure Firefox to use SSH Tunnel

In order to make Firefox connect via the SSH tunnel you’ll need to modify some settings. Once you are finished the browser will only work if the SSH tunnel is connected.

  • In the Firefox address bar browse to “about:config” with no quotes.
  • In the page that loads search and modify the following values:
    • network.proxy.proxy_over_tls
      • true
    • network.proxy.socks
      • Change to “127.0.0.1” with no quotes
    • network.proxy.socks_port
      • Change to the SSH Tunnel Local Port set above (1984?)
    • network.proxy.socks_remote_dns
      • Change to true
    • network.proxy.socks_version
      • Change to 5
    • network.proxy.type
        Change to 1

Step 6: Test and browse normally

Now that you have configured the above you should be able to browse via the tunnel. How can you check if it is working? Simply turn off the SSH Tunnel and try browsing – you should get an error message. Or if you are on a different WiFi you could try using a service to find your IP address and make sure it is different from where you are. For example if you configured Firefox to work via the SSH tunnel but left Chrome as is then visiting a site like http://www.whatismyip.com/ should show different information in each browser.

This post is a complication of two posts which originally appeared on my personal website here.




I am currently running a variety of distributions, primarily Linux Mint 18.
Previously I was running KDE 4.3.3 on top of Fedora 11 (for the first experiment) and KDE 4.6.5 on top of Gentoo (for the second experiment).

Applying updates to Docker and the Plex container

December 7th, 2014 No comments

In my last post, I discussed several Docker containers that I’m using for my home media streaming solution. Since then, Plex Media Server has updated to 0.9.11.4 for non-Plex Pass users, and there’s another update if you happen to pay for a subscription. As the Docker container I used (timhaak/plex) was version 0.9.11.1 at the time, I figured I’d take the opportunity to describe how to

  • update Docker itself to the latest version
  • run a shell inside the container as another process, to review configuration and run commands directly
  • update Plex to the latest version, and describe how not to do this
  • perform leet hax: commit the container to your local system, manually update the package, and re-commit and run Plex

Updating Docker

I alluded to the latest version of Docker having features that make it easier to troubleshoot inside containers. Switching to the latest version was pretty simple: following the instructions to add the Docker repository to my system, then running

sudo apt-get update
sudo apt-get install lxc-docker

upgraded Docker to version 1.3.1 without any trouble or need to manually uninstall the previous Ubuntu package.

Run a shell using docker exec

Let’s take a look inside the plex container. Using the following command will start a bash process so that we can review the filesystem on the container:

docker exec -t -i plex /bin/bash

You will be dropped into a root prompt inside the plex container. Check out the filesystem: there will be a /config and a /data directory pointing to “real” filesystem locations. You can also use ps aux to review the running processes, or even netstat -anp to see active connections and their associated programs. To exit the shell, use Ctrl+C – but the container will still be running when you use docker ps -a from the host system.

Updating Plex in-place: My failed attempt

Different Docker containers will have different methods of performing software updates. In this case, looking at the Dockerfile for timhaak/plex, we see that a separate repository was added for the Plex package – so we should be able to confirm that the latest version is available. This also means that if you destroy your existing container, pull the latest image, then launch a new copy, the latest version of Plex will be installed (generally good practice.)

But wait – the upstream repository at http://shell.ninthgate.se/packages/debian/pool/main/p/plexmediaserver/ does contain the latest .deb packages for Plex, so can’t we just run an apt-get update && apt-get upgrade?

Well, not exactly. If you do this, the initial process used to run Plex Media Server inside the Docker container (start.sh) gets terminated, and Docker takes down the entire plex container when the initial process terminates. Worse, if you then decide to re-launch things with docker start plex, the new version is incompletely installed (dpkg partial configuration).

So the moral of the story: if you’re trying this at home, the easiest way to upgrade is to recreate your Plex container with the following commands:

docker stop plex

docker rm plex

# The 'pull' process may take a while - it depends on the original repository and any dependencies in the Dockerfile. In this case it has to pull the new version of Plex.
docker pull timhaak/plex

# Customize this command with your config and data directories.
docker run -d -h plex --name="plex" -v /etc/docker/plex:/config -v /mnt/nas:/data -p 32400:32400 timhaak/plex

Once the container is up and running, access http://yourserver:32400/web/ to confirm that Plex Media Server is running. You can check the version number by clicking the gear icon next to your server in the left navigation panel, then selecting Settings.

Hacking the container: commit it and manually update Plex from upstream

If you’re more interested in hacking the current setup, there’s a way to commit your existing Plex image, manually perform the upgrade, and restart the container.

First, make sure the plex container is running (docker start plex) and then commit the container to your local filesystem (replacing username with your preferred username):

docker commit plex username/plex:latest

Then we can stop the container, and start a new instance where bash is the first process:

docker stop plex

docker rm plex

# Replace username with the username you selected above.
docker run -t -i --name="plex" -h plex username/plex:latest /bin/bash

Once inside the new plex container, let’s grab the latest Plex Media Server package and force installation:

curl -O https://downloads.plex.tv/plex-media-server/0.9.11.4.739-a4e710f/plexmediaserver_0.9.11.4.739-a4e710f_amd64.deb

dpkg -i plexmediaserver_0.9.11.4.739-a4e710f_amd64.deb

# When prompted, select Y to install the package maintainer's versions of files. In my instance, this updated the init script as well as the upstream repository.

Now, we can re-commit the image with the new Plex package. Hit Ctrl+D to exit the bash process, then run:

docker commit plex username/plex:latest

docker rm plex

# Customize this command with your config and data directories.
docker run -d -h plex --name="plex" -v /etc/docker/plex:/config -v /mnt/nas:/data -p 32400:32400 username/plex /start.sh

# Commit the image again so it will run start.sh if ever relaunched:
docker commit plex username/plex:latest

You’ll also need to adjust your /etc/init/plex.conf upstart script to point to username/plex.

The downside of this method is now that you’ve forked the original Plex image locally and will have to do this again for updates. But hey, wasn’t playing around with Docker interesting?




I am currently running Ubuntu 14.04 LTS for a home server, with a mix of Windows, OS X and Linux clients for both work and personal use.
I prefer Ubuntu LTS releases without Unity - XFCE is much more my style of desktop interface.
Check out my profile for more information.
Categories: Docker, Jake B, Plex, Ubuntu Tags:

Running a containerized media server with Ubuntu 14.04, Docker, and Plex

November 23rd, 2014 No comments

I recently took it upon myself to rebuild a general-purpose home server – installing a new Intel 530 240GB solid-state drive to replace a “spinning rust” drive, and installing a fresh copy of Ubuntu 14.04 now that 14.04.1 has released and there is much less complaining online.

The “new hotness” that I’d like to discuss has been the use of Docker to containerize various processes. Docker gets a lot of press these days, but the way I see it is a way to ensure that your special snowflake applications and services don’t get the opportunity to conflict with one another. In my setup, I have four containers running:

I like the following things about Docker:

  • Since it’s new, there are a lot of repositories and configuration instructions online for reference.
  • I can make sure that applications like Sonarr/NZBDrone get the right version of Mono that won’t conflict with my base system.
  • As a network administrator, I can ensure that only the necessary ports for a service get forwarded outside the container.
  • If an application state gets messed up, it won’t impact the rest of the system as much – I can destroy and recreate the individual container by itself.

There are some drawbacks though:

  • Because a lot of the images and Dockerfiles out there are community-based, there are some that don’t follow best practices or fall out of an update cycle.
  • Software updates can become trickier if the application is unable to upgrade itself in-place; you may have to pull a new Dockerfile and hope that your existing configuration works with a new image.
  • From a security standpoint, it’s best to verify exactly what an image or Dockerfile does before running it – for example, that it pulls content from official repositories (the docker-plex configuration is guilty of using a third-party repo, for example.)

To get started, on Ubuntu 14.04 you can install a stable version of Docker following these instructions, although the latest version has some additional features like docker exec that make “getting inside” containers to troubleshoot much easier. I was able to get all these containers running properly with the current stable version (1.0.1~dfsg1-0ubuntu1~ubuntu0.14.04.1). Once Docker is installed, you can grab each of the containers above with a combination of docker search and docker pull, then list the downloaded containers with docker images.

There are some quirks to remember. On the first run, you’ll need to docker run most of these containers and provide a hostname, box name, ports to forward and shared directories (known as volumes). On all subsequent runs, you can just use docker start $container_name – but I’ll describe a cheap and easy way of turning that command into an upstart service later. I generally save the start commands as shell scripts in /usr/local/bin/docker-start/*.sh so that I can reference them or adjust them later. The start commands I’ve used look like:

Plex
docker run -d -h plex --name="plex" -v /etc/docker/plex:/config -v /mnt/nas:/data -p 32400:32400 timhaak/plex
SABnzbd+
docker run -d -h sabnzbd --name="sabnzbd" -v /etc/docker/sabnzbd:/config -v /mnt/nas:/data -p 8080:8080 -p 9090:9090 timhaak/sabnzbd
Sonarr
docker run -d -h sonarr --name="sonarr" -v /etc/docker/sonarr:/config -v /mnt/nas:/data -p 8989:8989 tuxeh/sonarr
CouchPotato
docker run -d -h couchpotato --name="couchpotato" -e EDGE=1 -v /etc/docker/couchpotato:/config -v /mnt/nas:/data -v /etc/localtime:/etc/localtime:ro -p 5050:5050 needo/couchpotato
These applications have a “/config” and a “/data” shared volume defined. /data points to “/mnt/nas”, which is a CIFS share to a network attached storage appliance mounted on the host. /config points to a directory structure I created for each application on the host in /etc/docker/$container_name. I generally apply “chmod 777” permissions to each configuration directory until I find out what user ID the container is writing as, then lock it down from there.

For each initial start command, I choose to run the service as a daemon with -d. I also set a hostname with the “-h” parameter, as well as a friendly container name with “–name”; otherwise Docker likes to reference containers with wild adjectives combined with scientists, like “drunk_heisenberg”.

Each of these containers generally has a set of instructions to get up and running, whether it be on Github, the developer’s own site or the Docker Hub. Some, like SABnzbd+, just require that you go to http://yourserverip:8080/ and complete the setup wizard. Plex required an additional set of configuration steps described at the original repository:

  • Once Plex starts up on port 32400, access http://yourserverip:32400/web/ and confirm that the interface loads.
  • Switch back to your host machine, and find the place where the /config directory was mounted (in the example above, it’s /etc/docker/plex). Enter the Library/Application Support/Plex Media Server directory and edit the Preferences.xml file. In the <Preferences> tag, add the following attribute: allowedNetworks=”192.168.1.0/255.255.255.0″ where the IP address range matches that of your home network. In my case, the entire file looked like:

    <?xml version="1.0" encoding="utf-8"?>
    <Preferences MachineIdentifier="(guid)" ProcessedMachineIdentifier="(another_guid)" allowedNetworks="192.168.1.0/255.255.255.0" />

  • Run docker stop plex && docker start plex to restart the container, then load http://yourserverip:32400/web/ again. You should be prompted to accept the EULA and can now add library locations to the server.

Sonarr needed to be updated (from the NZBDrone branding) as well. From the GitHub README, you can enable in-container upgrades:

[C]onfigure Sonarr to use the update script in /etc/service/sonarr/update.sh. This is configured under Settings > (show advanced) > General > Updates > change Mechanism to Script.

To automatically ensure these containers start on reboot, you can either use restart policies (Docker 1.2+) or write an upstart script to start and stop the appropriate container. I’ve modified the example from the Docker website slightly to stop the container as well:

description "SABnzbd Docker container"
author "Jake"
start on filesystem and started docker
stop on runlevel [!2345]
respawn
script
/usr/bin/docker start -a sabnzbd
end script
pre-stop exec /usr/bin/docker stop sabnzbd

Copy this script to /etc/init/sabnzbd.conf; you can then copy it to plex, couchpotato, and sonarr.conf and change the name of the container and title in each. You can then test it by rebooting your system and running “docker ps -a” to ensure that all containers come up cleanly, or running “docker stop $container; service $container start”. If you run into trouble, the upstart logs are in /var/log/upstart/$container_name.conf.

Hopefully this introduction to a media server with Docker containers was thought-provoking; I hope to have further updates down the line for other applications, best practices and how this setup continues to operate in its lifetime.




I am currently running Ubuntu 14.04 LTS for a home server, with a mix of Windows, OS X and Linux clients for both work and personal use.
I prefer Ubuntu LTS releases without Unity - XFCE is much more my style of desktop interface.
Check out my profile for more information.
Categories: Docker, Jake B, Plex, Ubuntu Tags:

Cloud software for a Synology NAS and setting up OwnCloud

November 8th, 2014 No comments

Recently the Kitchener Waterloo Linux Users Group held a couple of presentations on setting up your own personally hosted cloud. With their permission we are pleased to also present it below:

Read more…




I am currently running a variety of distributions, primarily Linux Mint 18.
Previously I was running KDE 4.3.3 on top of Fedora 11 (for the first experiment) and KDE 4.6.5 on top of Gentoo (for the second experiment).

Big distributions, little RAM 7

October 13th, 2014 4 comments

It’s been a while but once again here is the latest instalment of the series of posts where I install the major, full desktop, distributions into a limited hardware machine and report on how they perform. Once again, and like before, I’ve decided to re-run my previous tests this time using the following distributions:

  • Debian 7.6 (GNOME)
  • Elementary OS 0.2 (Luna)
  • Fedora 20 (GNOME)
  • Kubuntu 14.04 (KDE)
  • Linux Mint 17 (Cinnamon)
  • Linux Mint 17 (MATE)
  • Mageia 4.1 (GNOME)
  • Mageia 4.1 (KDE)
  • OpenSUSE 13.1 (GNOME)
  • OpenSUSE 13.1 (KDE)
  • Ubuntu 14.04 (Unity)
  • Xubuntu 14.04 (Xfce)

I also attempted to try and install Fedora 20 (KDE) but it just wouldn’t go.

All of the tests were done within VirtualBox on ‘machines’ with the following specifications:

  • Total RAM: 512MB
  • Hard drive: 8GB
  • CPU type: x86 with PAE/NX
  • Graphics: 3D Acceleration enabled

The tests were all done using VirtualBox 4.3.12, and I did not install VirtualBox tools (although some distributions may have shipped with them). I also left the screen resolution at the default (whatever the distribution chose) and accepted the installation defaults. All tests were run between October 6th, 2014 and October 13th, 2014 so your results may not be identical.

Results

Just as before I have compiled a series of bar graphs to show you how each installation stacks up against one another. Measurements were taken using the free -m command for memory and the df -h command for disk usage.

Like before I have provided the results file as a download so you can see exactly what the numbers were or create your own custom comparisons (see below for link).

Things to know before looking at the graphs

First off if your distribution of choice didn’t appear in the list above its probably not reasonably possible to be installed (i.e. I don’t have hours to compile Gentoo) or I didn’t feel it was mainstream enough (pretty much anything with LXDE). As always feel free to run your own tests and link them in the comments for everyone to see.

First boot memory (RAM) usage

This test was measured on the first startup after finishing a fresh install.

 

All Data Points

All Data Points

RAM

RAM

Buffers/Cache

Buffers/Cache

RAM - Buffers/Cache

RAM – Buffers/Cache

Swap Usage

Swap Usage

RAM - Buffers/Cache + Swap

RAM – Buffers/Cache + Swap

Memory (RAM) usage after updates

This test was performed after all updates were installed and a reboot was performed.

All Data Points

All Data Points

RAM

RAM

Buffers/Cache

Buffers/Cache

RAM - Buffers/Cache

RAM – Buffers/Cache

Swap Usage

Swap Usage

RAM - Buffers/Cache + Swap

RAM – Buffers/Cache + Swap

Memory (RAM) usage change after updates

The net growth or decline in RAM usage after applying all of the updates.

All Data Points

All Data Points

RAM

RAM

Buffers/Cache

Buffers/Cache

RAM - Buffers/Cache

RAM – Buffers/Cache

Swap Usage

Swap Usage

RAM - Buffers/Cache + Swap

RAM – Buffers/Cache + Swap

Install size after updates

The hard drive space used by the distribution after applying all of the updates.

Install Size

Install Size

Conclusion

Once again I will leave the conclusions to you. Source data provided below.

Source Data




I am currently running a variety of distributions, primarily Linux Mint 18.
Previously I was running KDE 4.3.3 on top of Fedora 11 (for the first experiment) and KDE 4.6.5 on top of Gentoo (for the second experiment).

CoreGTK 2.24.0 Released!

August 4th, 2014 No comments

The initial version of CoreGTK, version 2.24.0, has been tagged for release today.

Features include:

  • Targets GTK+ 2.24
  • Support for GtkBuilder
  • Can be used on Linux, Mac and Windows

CoreGTK is an Objective-C language binding for the GTK+ widget toolkit. Like other “core” Objective-C libraries, CoreGTK is designed to be a thin wrapper. CoreGTK is free software, licensed under the GNU LGPL.

You can find more information about the project here and the release itself here.

This post originally appeared on my personal website here.




I am currently running a variety of distributions, primarily Linux Mint 18.
Previously I was running KDE 4.3.3 on top of Fedora 11 (for the first experiment) and KDE 4.6.5 on top of Gentoo (for the second experiment).

Linux alternatives: Mp3tag → EasyTAG

August 4th, 2014 1 comment

A big part of my move from Windows to Linux has been finding replacements for the applications that I had previously used day-to-day that are not available on Linux. For the major applications like my web browser (Firefox), e-mail client (Thunderbird), password manager (KeePass2) this hasn’t been a problem because they are all available on Linux as well. Heck you can even install Microsoft Office with the latest version of wine if you wanted to.

Unfortunately there still remains some programs that will simply not run under Linux. Thankfully this isn’t a huge deal because Linux has plenty of alternative applications that fill in all of the gaps – the trick is just finding the one that is right for you.

Mp3tag is an excellent Windows application that lets you edit the meta data (i.e. artist, album, track, etc.) inside of an MP3, OGG or similar file.

Mp3tag on Windows

Mp3tag on Windows

As a Linux alternative to this excellent program I’ve found a very similar application called EasyTAG that offers at least all of the features that I used to use in Mp3tag (and possibly even more).

EasyTAG on Linux

EasyTAG on Linux

For anyone looking for a good meta data editor I would highly recommend trying this one out.




I am currently running a variety of distributions, primarily Linux Mint 18.
Previously I was running KDE 4.3.3 on top of Fedora 11 (for the first experiment) and KDE 4.6.5 on top of Gentoo (for the second experiment).

Force Thunderbird/Enigmail to use a specific signing (hash) algorithm

June 8th, 2014 No comments

If you’ve had issues trying to get Thunderbird to send your PGP signed e-mail using anything other than SHA-1 there is a quick and easy fix that will let you pick whichever hash you prefer.

1) Open up Thunderbird’s preferences

2) On the Advanced Tab, under General click Config Editor

3) In the about:config window search for “extensions.enigmail.mimeHashAlgorithm” without quotes. Double click on this and enter a value. The value will determine which hash algorithm is used for signing.

The values are as follows:

0: Automatic selection, let GnuPG choose (note that while this may be the default it may also be the one that doesn’t work depending on your configuration).
1: SHA-1
2: RIPEMD-160
3: SHA-256
4: SHA-384
5: SHA-512
6: SHA-224

This post originally appeared on my personal website here.




I am currently running a variety of distributions, primarily Linux Mint 18.
Previously I was running KDE 4.3.3 on top of Fedora 11 (for the first experiment) and KDE 4.6.5 on top of Gentoo (for the second experiment).

Change the default sort order in Nautilus

February 9th, 2014 1 comment

The default sort order in Nautilus has been changed to sorting alphabetically by name and the option to change this seems to be broken. For example I prefer my files to be sorted by type so I ran

dconf-editor

and browsed to org/gnome/nautilus/preferences. From there you should be able to change the value by using the drop down:

 

Seems easy enough

Seems easy enough

Unfortunately the only option available is modification time. Once you change it to that you can’t even go back to name. This also appears to be a problem when trying to set the value using the command line interface like this:

dconf write /org/gnome/nautilus/preferences/default-sort-order type

I received an “error: 0-4:unknown keyword” message when I tried to run that.

Thanks to the folks over on the Ask Ubuntu forum I was finally able to get it to change by issuing this command instead:

gsettings set org.gnome.nautilus.preferences default-sort-order type

where type could be swapped out for whatever you prefer it to be ordered by.

Great Success!

Great Success!

CoreGTK

January 28th, 2014 2 comments

A while back I made it my goal to put together an open source project as my way of contributing back to the community. Well fast forward a couple of months and my hobby project is finally ready to be shown the light of day. I give you… CoreGTK

CoreGTK is an Objective-C binding for the GTK+ library which wraps all objects descending from GtkWidget (plus a few others here and there). Like other “core” Objective-C libraries it is designed to be a very thin wrapper, so that anyone familiar with the C version of GTK+ should be able to pick it up easily.

However the real goal of CoreGTK is not to replace the C implementation for every day use but instead to allow developers to more easily code GTK+ interfaces using Objective-C. This could be especially useful if a developer already has a program, say one they are developing for the Mac, and they want to port it to Linux or Windows. With a little bit of MVC a savvy developer would only need to re-write the GUI portion of their application in CoreGTK.

So what does a CoreGTK application look like? Pretty much like a normal Objective-C program:

/*
 * Objective-C imports
 */
#import <Foundation/Foundation.h>
#import "CGTK.h"
#import "CGTKButton.h"
#import "CGTKSignalConnector.h"
#import "CGTKWindow.h"

/*
 * C imports
 */
#import <gtk/gtk.h>

@interface HelloWorld : NSObject
/* This is a callback function. The data arguments are ignored
 * in this example. More callbacks below. */
+(void)hello;

/* Another callback */
+(void)destroy;
@end

@implementation HelloWorld
int main(int argc, char *argv[])
{
    NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init];

    /* We could use also CGTKWidget here instead */
    CGTKWindow *window;
    CGTKButton *button;

    /* This is called in all GTK applications. Arguments are parsed
    * from the command line and are returned to the application. */
    [CGTK autoInitWithArgc:argc andArgv:argv];

    /* Create a new window */
    window = [[CGTKWindow alloc] initWithGtkWindowType:GTK_WINDOW_TOPLEVEL];

    /* Here we connect the "destroy" event to a signal handler in 
     * the HelloWorld class */
    [CGTKSignalConnector connectGpointer:[window WIDGET] 
        withSignal:@"destroy" toTarget:[HelloWorld class] 
        withSelector:@selector(destroy) andData:NULL];

    /* Sets the border width of the window */
    [window setBorderWidth: [NSNumber numberWithInt:10]];

    /* Creates a new button with the label "Hello World" */
    button = [[CGTKButton alloc] initWithLabel:@"Hello World"];

    /* When the button receives the "clicked" signal, it will call the
     * function hello() in the HelloWorld class (below) */
    [CGTKSignalConnector connectGpointer:[button WIDGET] 
        withSignal:@"clicked" toTarget:[HelloWorld class] 
        withSelector:@selector(hello) andData:NULL];

    /* This packs the button into the window (a gtk container) */
    [window add:button];

    /* The final step is to display this newly created widget */
    [button show];

    /* and the window */
    [window show];

    /* All GTK applications must have a [CGTK main] call. Control ends here
     * and waits for an event to occur (like a key press or
     * mouse event). */
    [CGTK main];

    [pool release];

    return 0;
}

+(void)hello
{
    NSLog(@"Hello World");
}

+(void)destroy
{
    [CGTK mainQuit];
}
@end
Hello World in action

Hello World in action

And because Objective-C is completely compatible with regular old C code there is nothing stopping you from simply extracting the GTK+ objects and using them like normal.

// Use it as an Objective-C CoreGTK object!
CGTKWindow *cWindow = [[CGTKWindow alloc] 
    initWithGtkWindowType:GTK_WINDOW_TOPLEVEL];

// Or as a C GTK+ window!
GtkWindow *gWindow = [cWindow WINDOW];

// Or even as a C GtkWidget!
GtkWidget *gWidget = [cWindow WIDGET];

// This...
[cWindow show];

// ...is the same as this:
gtk_widget_show([cWindow WIDGET]);

You can even use a UI builder like GLADE, import the XML and wire up the signals to Objective-C instance and class methods.

CGTKBuilder *builder = [[CGTKBuilder alloc] init];
if(![builder addFromFile:@"test.glade"])
{
    NSLog(@"Error loading GUI file");
    return 1;
}

[CGTKBuilder setDebug:YES];

NSDictionary *dic = [[NSDictionary alloc] initWithObjectsAndKeys:
                 [CGTKCallbackData withObject:[CGTK class] 
                     andSEL:@selector(mainQuit)], @"endMainLoop",
                 [CGTKCallbackData withObject:[HelloWorld class] 
                     andSEL:@selector(hello)], @"on_button2_clicked",
                 [CGTKCallbackData withObject:[HelloWorld class] 
                     andSEL:@selector(hello)], @"on_button1_activate",
                 nil];

[builder connectSignalsToObjects:dic];

CGTKWidget *w = [builder getWidgetWithName:@"window1"];
if(w != nil)
{
    [w showAll];
}

[builder release];

So there you have it that’s CoreGTK in a nutshell.

There are a variety of ways to help me out with this project if you are so inclined to do so. The first task is probably just to get familiar with it. Download CoreGTK from the GitHub project page and play around with it. If you find a bug (very likely) please create an issue for it.

Another easy way to get familiar with CoreGTK is to help write/fix documentation – a lot of which is written in the source code itself. Sadly most of the current documentation simply states which underlying GTK+ function is called and so it could be cleaned up quite a bit.

At the moment there really isn’t anything more formal than that in place but of course code contributions would also be welcome!

Update: added some pictures of the same program running on all three operating systems.

Hello World on Windows

Hello World on Windows

Hello World on Mac

Hello World on Mac

Hello World on Linux

Hello World on Linux

This post originally appeared on my personal website here.




I am currently running a variety of distributions, primarily Linux Mint 18.
Previously I was running KDE 4.3.3 on top of Fedora 11 (for the first experiment) and KDE 4.6.5 on top of Gentoo (for the second experiment).

Open source project hosting options

September 8th, 2013 2 comments

So you want to host an open source project using one of the many free services available but can’t decide which one to use? If only someone would put together a quick summary of each of the major offerings…

Hosting providers covered in this post:

  • Bitbucket
  • CodePlex
  • GitHub
  • Gitorious
  • Google Code
  • Launchpad
  • SourceForge

Bitbucket

Bitbucket is a hosting site for the distributed version control systems (DVCS) Git and Mercurial. The service offering includes an issue tracker and wiki, as well as integration with a number of popular services such as Basecamp, Flowdock, and Twitter.

Features:

  • Supports both Git and Mercurial
  • Allows private repositories for free, up to 5 users
  • Unlimited repositories
  • Has JIRA integration for issue tracking
  • Has its own REST API

Downsides:

  • Only allows up to 5 users for free (a user defined as someone with read or write access)

CodePlex

CodePlex is Microsoft’s free open source project hosting site. You can create projects to share with the world, collaborate with others on their projects, and download open source software.

Features:

  • Supports both Git & Mercurial
  • Integrated Wiki that allows to add rich documentation and nice looking pages
  • Bug Tracker and Discussion Forums included

Downsides:

  • Often times feels more like a code publishing platform than a collaboration site
  • Primarily geared toward .NET projects

GitHub

Build software better, together. Powerful collaboration, code review, and code management for open source and private projects.

Features:

  • Supports Git
  • Powerful and easy to use graphical tools
  • Easy team management
  • Integrated wiki, issue tracker and code review

Downsides:

  • Only supports Git
  • Quite a few ‘dead’ projects on the site

Gitorious

The Git hosting software that you can install yourself. Gitorious.org provides free hosting for open source projects that use Git.

Features:

  • Supports Git
  • Free project hosting
  • Integrated wiki
  • Can download the software and install it on your own server

Downsides:

  • Only supports Git

Google Code

Project Hosting on Google Code provides a free collaborative development environment for open source projects.

Features:

  • Supports Subversion, Mercurial Git
  • Integrated wiki

Downsides:

  • Not very pretty

Launchpad

Launchpad is a software collaboration platform.

Features:

  • Supports Bazaar
  • Integrated bug tracking and code reviews
  • Ubuntu package building and hosting
  • Mailing lists

Downsides:

  • Only supports Bazaar
  • Geared toward Ubuntu (which can be a downside depending on your project)

SourceForge

Find, Create, and Publish Open Source software for free.

Features:

  • Supports Git, Mercurial, Subversion
  • Integrated issue tracking, wiki, discussion forums
  • Stat tracking

Downsides:

  • Ads
  • A lot of ‘dead’ projects

 

Now obviously I’ve missed some things and glossed over others but my goal here was to provide a quick ‘at a glance’ summary of each. Check the individual websites for more. Thanks to the people over at Stack Exchange for doing a lot of the legwork.




I am currently running a variety of distributions, primarily Linux Mint 18.
Previously I was running KDE 4.3.3 on top of Fedora 11 (for the first experiment) and KDE 4.6.5 on top of Gentoo (for the second experiment).

One license to rule them all? Noooooooooope!

August 20th, 2013 No comments

Lately I’ve been taking a look at the various open source software licenses in an attempt to better understand the differences between them. Here is my five minute summary of the most popular licenses:

GNU Public License (GPL)

Requires that any project using a GPL-licensed component must also be made available under the GPL. Basically once you go GPL you can’t go back.

Lesser GNU Public License (LGPL)

Basically the same as the GPL except that if something uses software licensed as LGPL it also doesn’t need to be licensed the same. So if you write a program that uses an LGPL library, say a program with a GTK+ user interface, it doesn’t need to be licensed LGPL. This is useful for commercial applications that rely on open source technology.

v2 vs v3

There are a number of differences between version 2 and version 3 of the GPL and LGPL licenses. Version 3 attempts to clarify a number of issues in version 2 including how patents, DRM, etc. are handled but a number of developers don’t seem to like the differences so version 2 is still quite popular.

MIT

This license allows for almost anything as long as a copy of the license and copyright are included in any distribution of the code. It can be used in commercial software without issue.

BSD3

Similar to the MIT, this license basically only requires that a copy of the license and copyright are included in any distribution of the code. The major difference between this and the MIT is that the BSD3 prohibits the use of the copyright holder’s name in any promotion of derivative work.

Apache

Apache is similar to the BSD license in that you have to provide a copy of the license in any derivative works. In addition there are a number of extra safeguards, such as patent grants, that set it apart from BSD.




I am currently running a variety of distributions, primarily Linux Mint 18.
Previously I was running KDE 4.3.3 on top of Fedora 11 (for the first experiment) and KDE 4.6.5 on top of Gentoo (for the second experiment).

Listen up, Kubuntu: the enraging tale of sound over HDMI

August 4th, 2013 2 comments

Full disclosure: I live with Kayla, and had to jump in to help resolve an enraging problem we ran into on the Kubuntu installation with KDE, PulseAudio and the undesirable experience of not having sound in applications. It involved a fair bit of terminal work and investigation, plus a minimal understanding of how sound works on Linux. TuxRadar has a good article that tries to explain things. When there are problems, though, the diagram looks much more like the (admittedly outdated) 2007 version:

The traditional spiderweb of complexity involved in Linux audio.

The traditional spiderweb of complexity involved in Linux audio.

To give you some background, the sound solution for the projection system is more complicated than “audio out from PC, into amplifier”. I’ve had a large amount of success in the past with optical out (S/PDIF) from Linux, with only a single trip to alsamixer required to unmute the relevant output. No, of course the audio path from this environment has to be more complicated, and looks something like:

Approximate diagram of display and audio output involved from Kubuntu machine

As a result, the video card actually acts as the sound output device, and the amplifier takes care of both passing the video signal to the projector and decoding/outputting the audio signal to the speakers and subwoofer. Under Windows, this works very well: in Control Panel > Sound, you right-click on the nVidia HDMI audio output and set it as the default device, then restart whatever application plays audio.

In the KDE environment, sound is managed by a utility called Phonon in the System Settings > Multimedia panel, which has multiple backends for ALSA and PulseAudio. It will essentially communicate with the highest-level sound output system installed that it has support for. When you make a change in a default Kubuntu install in Phonon it appears to be talking to PulseAudio, which in turn changes necessary ALSA settings. Sort of complicated, but I guess it handles the idea that multiple applications can play audio and not tie up the sound card at the same time – which has not always been the case with Linux.

In my traditional experience with the GNOME and Unity interfaces, it always seems like KDE took its own path with audio that wasn’t exactly standard. Here’s the problem I ran into: KDE listed the two audio devices (Intel HDA and nVidia HDA), with the nVidia interface containing four possible outputs – two stereo and two listed as 5.1. In the Phonon control panel, only one of these four was selectable at a time, and not necessarily corresponding to multiple channel output. Testing the output did not play audio, and it was apparent that none of it was making it to the amplifier to be decoded or output to the speakers.

Using some documentation from the ArchLinux wiki on ALSA, I was able to use the aplay -l command to find out the list of detected devices – there were four provided by the video card:

**** List of PLAYBACK Hardware Devices ****
card 0: PCH [HDA Intel PCH], device 0: ALC892 Analog [ALC892 Analog]
Subdevices: 1/1
Subdevice #0: subdevice #0
card 0: PCH [HDA Intel PCH], device 1: ALC892 Digital [ALC892 Digital]
Subdevices: 1/1
Subdevice #0: subdevice #0
card 1: NVidia [HDA NVidia], device 3: HDMI 0 [HDMI 0]
Subdevices: 1/1
Subdevice #0: subdevice #0
card 1: NVidia [HDA NVidia], device 7: HDMI 0 [HDMI 0]
Subdevices: 1/1
Subdevice #0: subdevice #0
card 1: NVidia [HDA NVidia], device 8: HDMI 0 [HDMI 0]
Subdevices: 1/1
Subdevice #0: subdevice #0
card 1: NVidia [HDA NVidia], device 9: HDMI 0 [HDMI 0]
Subdevices: 1/1
Subdevice #0: subdevice #0

and then use aplay -D plughw:1,N /usr/share/sounds/alsa/Front_Center.wav repeatedly where N is the number of one of the nVidia detected devices. Trial and error let me discover that card 1, device 7 was the desired output – but there was still no sound from the speakers in any KDE applications or the Netflix Desktop client. Using the ALSA output directly in VLC, I was able to get an MP3 file to play properly when selecting the second nVidia HDMI output in the list. This corresponds to the position in the aplay output, but VLC is opaque about the exact card/device that is selected.

At this point my patience was wearing pretty thin. Examining the audio listing further – and I don’t exactly remember how I got to this point – the “active” HDMI output presented in Phonon was actually presented as card 1, device 3. PulseAudio essentially grabbed the first available output and wouldn’t let me select any others. There were some additional PulseAudio tools provided that showed the only possible “sink” was card 1,3.

The brute-force, ham-handed solution was to remove PulseAudio from a terminal (sudo apt-get remove pulseaudio) and restart KDE, presenting me with the following list of possible devices read directly from ALSA. I bumped the “hw:1,7” card to the top and also quit the system tray version of Amarok.

A list of all the raw ALSA devices detected by KDE/Phonon after removing PulseAudio.

A list of all the raw ALSA devices detected by KDE/Phonon after removing PulseAudio.

Result: Bliss! By forcing KDE to output to the correct device through ALSA, all applications started playing sounds and harmony was restored to the household.

At some point after the experiment I will see if I can get PulseAudio to work properly with this configuration, but both Kayla and I are OK with the limitations of this setup. And hey – audio works wonderfully now.




I am currently running Ubuntu 14.04 LTS for a home server, with a mix of Windows, OS X and Linux clients for both work and personal use.
I prefer Ubuntu LTS releases without Unity - XFCE is much more my style of desktop interface.
Check out my profile for more information.

An Ambitious Goal

August 1st, 2013 3 comments

Every since we announced the start of the third Linux Experiment I’ve been trying to think of a way in which I could contribute that would be different from the excellent ideas the others have come up with so far. After batting around some ideas over the past week I think I’ve finally come up with how I want to contribute back to the community. But first a little back story.

A large project now, GNOME was started because there wasn't a good open source alternative at the time

A large project now, GNOME was started because there wasn’t a good open source alternative at the time

During the day I develop commercial software. An unfortunate result of this is that my personal hobby projects often get put on the back burner because in all honesty when I get home I’d rather be doing something else. As a result I’ve developed, pun intended, quite a catalogue of projects which are currently on hold until I can find time/motivation to actually make something of them. These projects run the gamut of little helper scripts, written to make my life more convenient, all the way up to desktop applications, designed to take on bigger tasks. The sad thing is that while a lot of these projects have potential I simply haven’t been able to finish them, and I know that if I just could they would be of use to others as well. So for this Experiment I’ve decided to finally do something with them.

Thanks to OpenOffice, LibreOffice and others there are actual viable open source alternatives to Microsoft Office

Thanks to OpenOffice.org, LibreOffice and others there are actual viable open source alternatives to Microsoft Office

Open source software is made up of many different components. It is simultaneously one part idea, perhaps a different way to accomplish X would be better, one part ideal, belief that sometimes it is best to give code away for free, one part execution, often times a developer just “scratching an itch” or trying a new technology, and one part delivery, someone enthusiastically giving it away and building a community around it. In fact that’s the wonderful thing about all of the projects we all know and love; they all started because someone somewhere thought they had something to share with the world. And that’s what I plan to do. For this Linux Experiment I plan on giving back by setting one of my hobby projects free.

Before this open source web browser we were all stuck with Internet Explorer 6

Before this open source web browser we were all stuck with Internet Explorer 6

Now obviously this is not only ambitious but perhaps quite naive as well especially given the framework of The Linux Experiment – I fully recognize that I have quite a bit of work ahead of me before any of my hobby code is ready to be viewed, let alone be used, by anyone else. I also understand that, given my own personal commitments and available time, it may be quite a while before anything actually comes of this plan. All of this isn’t exactly well suited for something like The Linux Experiment, which thrives on fresh content; there’s no point in me taking part in the Experiment if I won’t be ready to make a new post until months from now. That is why for my Experiment contributions I won’t be only relying on the open sourcing of my code, but rather I will be posting about the thought process and research that I am doing in order to start an open source project.

Topics that I intend to cover are things relevant to people wishing to free their own creations and will include things such as:

  • weighing the pros and cons as well as discussing the differences between the various open source licenses
  • the best place to host code
  • how to structure the project in order to (hopefully) get good community involvement
  • etc.

An interesting side effect of this approach will be somewhat of a new look into the process of open sourcing a project as it is written piece by piece, step by step, rather than in retrospect.

The first billion dollar company built on open source software

The first billion dollar company built on open source software

Coincidentally as I write this post the excellent website tuxmachines.org has put together a group of links discussing the pros of starting open source projects. I’ll be sure to read up on those after I first commit to this 😉

Linux: a hobby project initially created and open sourced by one 21 year old developer

Linux: a hobby project initially created and open sourced by one 21 year old developer

I hope that by the end of this Experiment I’ll have at least provided enough information for others to take their own back burner projects to the point where they too can share their ideas and creations with the world… even if I never actually get to that point myself.

P.S. If anyone out there has experience in starting an open source project from scratch or has any helpful insights or suggestions please post in the comments below, I would really love to hear them.




I am currently running a variety of distributions, primarily Linux Mint 18.
Previously I was running KDE 4.3.3 on top of Fedora 11 (for the first experiment) and KDE 4.6.5 on top of Gentoo (for the second experiment).

My Initial Thoughts/Experiences with ArchLinux

July 29th, 2013 2 comments

Hello again everyone! By this point, I have successfully installed ArchLinux, as well as KDE, and various other everyday applications necessary for my desktop.

Aside from the issues with the bootloader I experienced, the installation was relatively straight forward. Since I have never used ArchLinux before, I decided to follow the Beginner’s Guide in order to make sure I wasn’t screwing anything up. The really nice thing about this guide is that it only gives you the information that you need to get up and running. From here, you can add any packages you want, and do any necessary customization.

Overall, the install was fairly uneventful. I also managed to install KDE, Firefox, Flash, and Netflix (more below) without any issues.

Some time ago, there was a package created for Ubuntu that allows you to watch Netflix on Linux. Since then, someone has created a package for ArchLinux called netflix-desktop. What this does, is creates an instance of Firefox in WINE that runs Silverlight so that the Netflix video can be loaded. The only issue that I’m running into with this package is that when I full-screen the Netflix video, my taskbar in KDE still appears. For the time being, I’ve just set the taskbar to allow windows to go over top. If anyone has any suggestions on how to resolve this, please let me know.

netflix

This isn’t my screenshot. I found it on the interweb. I just wanted to give you a good idea of how netflix-desktop looked. I’d like to thank Richard in advance for the screenshot.

Back to a little more about ArchLinux specifically. I’ve really been enjoying their package management system. From my understanding so far, there are two main ways to obtain packages. The official repositories are backed by “pacman” which is the main package manager. Therefore, if you wanted to install kde, you would do “pacman -S kde”. This is similar to the package managers on other distributions such as apt-get. The Arch User Repository is a repository of build scripts created by ArchLinux users that allow you to compile and configure other packages not contained within the official repositories. The really neat thing about this is that it can also download and install and dependencies contained in the official repositories using pacman automatically.

As I go forward, I am also thinking of ways I can contribute to the ArchLinux community, but for now, I will continue to explore and experiment.


I am currently running ArchLinux (x86_64).
Check out my profile for more information.