diff options
| author | Katharina Fey <kookie@spacekookie.de> | 2017-10-12 15:13:13 +0200 |
|---|---|---|
| committer | Katharina Fey <kookie@spacekookie.de> | 2017-10-12 15:13:13 +0200 |
| commit | 964d22f599ab7f300663e4ef3dac0f8749b1076d (patch) | |
| tree | 5ab4ded280c12e230fdb34bef1839ec683e854c0 /content/blog | |
| parent | b004cfd65b993a92a24767a9b022fb4e04554e7d (diff) | |
Moving all of the website into the public repo. This is a mess
Diffstat (limited to 'content/blog')
19 files changed, 1079 insertions, 0 deletions
diff --git a/content/blog/00_introduction.md b/content/blog/00_introduction.md new file mode 100644 index 0000000..dabfe55 --- /dev/null +++ b/content/blog/00_introduction.md @@ -0,0 +1,14 @@ +Title: An open source GPU +Category: Blog +Tags: OpenMPU, matrices, opengl, hardware + + +For as long as I can remember I've wanted to build my own computer. Sorta from scratch. + +And there is something I want to do even more: build a GPU. Something that can process vectors and matrecies in an efficient manner, build a graphics rendering pipeline in hardware. Sorta from scratch. + +And that's what this project is about. OpenMCU stands for "Open Matrix Computation Unit" and will be a part of the OpenGPU (working title). + +## The basic architecture + +`<insert picture>`
\ No newline at end of file diff --git a/content/blog/hardware/001-omnitool-introduction.md b/content/blog/hardware/001-omnitool-introduction.md new file mode 100755 index 0000000..dd7b5e2 --- /dev/null +++ b/content/blog/hardware/001-omnitool-introduction.md @@ -0,0 +1,57 @@ +Title: 01. Omnitool - Introduction +Slug: 01-omnitool-introduction +Category: Blog +Tags: Dev Diary, Hardware +Date: 17-09-2015 15:45 +Status: published +Illustration: omnitool_background2.jpg + +Any good fan of the game, book, comic, stuff series will now jump up and down in excitement. The omnitool is a constant companion, helper and friend for Commander Shepard and a life saver in several situations. Whether it be hacking doors, turning into a plasma blade and cutting peoples faces open or just plain transferring a bribe credit to the slimy Vorcha in front of you. + +But please, settle down, I haven't invented holographic technology. Nor have I invented plasma tubings or even solved the financial crisis by coming up with a perfect currency (of couse generically called 'credits') that everybody wants to use. + +### Then why waste your time +That's a good question :) I would hope that this series doesn't turn into a waste of time for anyone. Because...well, while I haven't done any of those things. I am planning on building an omnitool. Just a bit more low-tech. + +I was actually inspired by something on Adafruit, called the Flora. It's a round gimmick with a ring of RGB LED's, GPS (I think) and an arduino to program it. + + + +I was only really inspired to do this project when a friend of mine showed off his Flora on the CCCamp2015. + +He did some minor modifications to it, including a wristband (IDE cable) and a battery on the underside and programmed a few modes for displaying time (as an analogue clock) and a flashlight by just dialing the LED's up to full power. + + + +And that's kinda what gave me the idea for an omnitool. The idea of circular rings of LED's as display elements are pretty cool. + +### Basic concept + +So the basic conceptis a simple. Create a wrist accessory with one or two LED rings (using shift register RGP LED's to display patterns, colours and different brightness settings), include a generic SOC to program, probably something single core ARM. Give it a bunch of RAM to run applications and a embedded systems linux. + +Include GPS, blutooth, a sensor package such as temperature, preasure, accelerometer, etc. + +Include extention slots where, with a simple click, the tool can be expanded to include speakers, a microphone, a bigger screen, a bigger battery, etc. + +And all this in the form factor of the so beloved omnitool from Mass Effect. + +I know this is a bit of a crazy project. And it will take months, if not years to complete. + +Because this is the thing: I want to do it all as custom cut PCB and maybe some custom cut plastic for casings. + +I've been getting into KiCad recently, with my first project the Christmas Bauble ([Click here for details](/dev-diary/jolly-christmas-decoration/)) and have fallen in love with the tool – Don't worry Ashley, not *that* much :) + +But it is pretty awesome and I urge everybody who wants to get into that sort of electronics stuff to checkout it out! [KiCad](http://www.kicad-pcb.com) + +### What to expect + +So what will this series be? (Hopefully) regular status reports about what I've been doing, writing about my experimences with the project, letting you guys know what I'm learning and generally just let people follow the project. + +All stuff about the project will be in a Github repo. From the KiCad files to the C firmware I'll have to write. Everything you would potentially need to make your own, study it and learn from it is in there. + +**[Omnitool Repository](https://github.com/spacekookie/omnitool)** + + +I hope that you follow along. And I'm looking forwards to comments from all of you. Have a lovely day and read you soon. + +~Kate diff --git a/content/blog/hardware/002-christmas-bauble.md b/content/blog/hardware/002-christmas-bauble.md new file mode 100755 index 0000000..b0cc62c --- /dev/null +++ b/content/blog/hardware/002-christmas-bauble.md @@ -0,0 +1,33 @@ +Title: Jolly Christmas Decoration +Category: Blog +Date: 2015-09-17 15:30 +Tags: Dev Diary, Hardware + +Christmas is getting closer (not really but let's just roll with it) and I wanted to learn [KiCad](www.kicad-pcb.com) a software that let's you create circuits and design PCB for manufacture. + +I found a tutorial series online by a guy named [Ashley Mills](https://www.youtube.com/channel/UCaBNA-lmg35Wfx2eh2oDkWg) (with quite a legendary beard) who showed off a simple circuit using a 555-timer, a shift register and an XOR gate made from NPN transistors and resistors to display and repeat a pattern on several LED's. + +The series focused on getting to know KiCad and all it's features. And while I did that in the first revision of my board, I've diverged from it since. I can however recommend his videos on KiCad to anyone who wants to dive into PCB design, has no clue about the software and could use a little chuckle while also learning some really awesome software (youtube channel link above). + +### My Christmas Bauble + +So this is what I've got. + + + +As you can see it's a round PCB with simple 5mm LED's around the edges. It no longer uses NPN transistors but rather a single SMD XOR gate. Much easier to wire up, cheaper and less prone to errors as well. + +In general I've switched the entire design over to primarily use SMD components as they're smaller and more elegant. And it theoretically allowed me to get the footprint of the board down to something that isn't too excruciatingly expensive to produce. + +It took me two more revisions to get the board to a state where it's not too complex and actually fit on a single layer (!) with no vias except for the holes for the LED's obviously. + +It uses a round cell battery on the back of the board to hide it away and has a hole at the top to actually hang off a christmas tree. Theoretically the battery should lasta few days, so maybe have a few spare ones around in the christmas season. + +### What now? + +I haven't manufactured this yet. I am still thinking about refining the design slightly. I have the **entire** back to work with and add things. I was thinking about adding a simple bluetooth chip so that patterns could be pushed to the device via an android app. But that's the future. For now it should actually be functional and maybe I'll order some `Revision 3` boards just to see that everything worked. + +Here is a dynamic render from KiCad as well. + + +And be sure to checkout my Github repo for the project if you want the KiCad files. Either to play around with them or to manufacture some baubles yourself. If you do, I'd be interested in pictures of the decorations on your christmas trees so I can add them to this article as a slideshow :) diff --git a/content/blog/hardware/003-christmas-bauble-update.md b/content/blog/hardware/003-christmas-bauble-update.md new file mode 100755 index 0000000..ee4800d --- /dev/null +++ b/content/blog/hardware/003-christmas-bauble-update.md @@ -0,0 +1,38 @@ +Title: [Update] Jolly Christmas Decoration +Category: Blog +Date: 2015-11-27 15:30 +Tags: Dev Diary, Hardware + +You might remember I played around with Kicad a few months ago and made this [tacky little thing](/hardware/jolly-christmas-decoration/). Just about 2 1/2 weeks ago I went onto [DirtyPCB](http://dirtypcbs.com/) to get them actually made. I wanted to have gone through the production process and get something built before I started doing more complicated projects. + +Unfortunately I discovered a little mistake with the design in the layout that ended up at the manufacturer (Rev 3.1). I tried to fix them but Rev 3.2 didn't make it in time, which means my boards will be a bit more complicated to power. However not too complicated as the power-in are just throughholes so I can actually strap anything behind it to power it. + +But without further a due, here is the result from DirtyPCB (which I am actually quite impressed with). + + + +Now, I'm new to all of this so I started doing beep-tests on the pads to make sure things were properly connected and all the boards passed them. The production quality is pretty good. Unfortunately I can't start assembling them yet just because none of the parts I ordered for them have arrived yet. The manufacture and shipping of the boards actually beat the shipping of off-the-shelf parts! + +Anyways, I'm kinda excited. First time making an electronics project. I might post another update on when the parts arive and post a few gifs of the finished products. If I don't I'll probably tweet about it though. + +Now, I have another smaller electronics project in the making where I am, again, waiting for parts to arrive to do some testing. And already designing a modular PCB board. (Limited a bit with the 10x10cm limitations on DirtyPCB I need to design my project in a way that I can take a bunch of smaller panels and stick them together into a large one, which would take hundreds of dollars to make elsewhere). + +[But realistically for the production quality I saw with these, I'd be happy to give them my money again for future projects. Especially at that price, just unbeatable.](https://www.youtube.com/watch?v=d36wUmJGzvA) + +:) + +Anyways, enough ramblings. Read you later. + +### Update...update + +Right...so after tinkering with the bauble a bit I found out a few things. The most important one being that I made some mistakes. Some big ones :) + + - Pin 9 of the shift register was connected to both input A and input B of the XOR gate. Which meant that both inputs were always the same...which also meant that the output was always 0. + - The 555-timer clock ran at several hundred kilohertz. I had to change the capacitor down to ~12µF and the resistors to ~4.7 ohms. + - The coin-cell battery didn't have enough juice to run it. Two had to be put in parallel. Even then, two batteries would not be able to run for very long. + + To make the bauble work I bridget the xor gate completely, so just feeding back the shift register end to the beginning. + + In addition to those things some of the LED's sometimes didn't work. I'm not sure if that is due to broken shift registers, traces or LEDs. All in all I do consider it to have tought me quite a lot about electronics, going through the process of producing a PCB and debugging electronics once it arrived and inevidably goes wrong :) + +I am currently in the process of redesigning the entire circuite from scratch. And making it easier to solder. I want to make it into a beginner soldering kit that people can both learn how to solder with and also have something to hang off their christmas tree in the jolly season. diff --git a/content/blog/hardware/004-ws2812b-breakouts.md b/content/blog/hardware/004-ws2812b-breakouts.md new file mode 100755 index 0000000..190d100 --- /dev/null +++ b/content/blog/hardware/004-ws2812b-breakouts.md @@ -0,0 +1,21 @@ +Title: WS2812b LED breakouts +Category: Blog +Date: 2016-03-16 12:08 +Tags: Dev Diary, Hardware + +You all know I have a fetish for [ws2812b RGB-LEDs](http://rgb-123.com/wp-content/uploads/2013/08/PinLayout.jpg). I admit that. They're just awesome. And recently I've found myself wanting to do some projects with them (*cough* programmable alarm clock *cough*). But I dislike the strips, although they're pretty cheap they are usually just very messy and horrible. And making a PCB for every project can be weird. Especially if gigantic PCBs would be required. + +So I designed this little doodad. + + + +The idea is the following: Sometimes you just need a few ws2812b (I'm saying that word too often in this post) somewhere. But you don't want to lay a strip. Or make a big PCB for it. +So here is an alternative. Easily make ~800 of these for 25$, screw them to a surface, connect **PRETTY** wires between them, such as [Ribbon cables](http://cdn.usdigital.com/assets/images/galleries/ca-c10-f-c10_0.jpg) and boom. You're done. + +### One to rule many + +Now...I mentioned that programmable alarm clock earlier. And while I'm not quite done designing what will go into it all, I do know that I want to have a ~ 21x9 Pixel display, each individually addressable. And instead of building a way too big PCB that will be insanely expensive to manufacture...why not split them up? Then I have these "tiles" of LEDs that I screw to a backplate, wire everything together and from the outside you can't tell the difference. + +With 21:9 (Aspect ratio and pixels) in mind, this is my prototype: + +<TO BE CONTINUED>
\ No newline at end of file diff --git a/content/blog/hardware/005-open-plantb0t.md b/content/blog/hardware/005-open-plantb0t.md new file mode 100755 index 0000000..647ddf3 --- /dev/null +++ b/content/blog/hardware/005-open-plantb0t.md @@ -0,0 +1,51 @@ +Title: Open Plantb0t - Rev A +Category: Blog +Date: 2016-03-16 12:08 +Tags: Dev Diary, Hardware +Illustration: banners/plantb0t_revA.png + +Howdy everybody, + +Spring is coming in Berlin and thus my thoughts - as every year - are with plants. And growing shit. I live in an appartment with a tiny tiny balcony so I don't have much space but that has never stopped me from wanting to cram as many plants into the space as possible to the point of starting nuclear fusion. + +In addition to that I have a few house-plants and very water-sensitive trees in my appartment. My current approach is to go around with a jug of water every couple of days and water them individually - making sure the soil has a certain moisture and doesn't exceed a certain limit - but I've always had the dream of being able to automate away as much as possible. That's where the idea of `Plantb0t` started. And I want to tell you a little bit about it. + +The basic idea is to have a little controller in each plant-pot that measure the moisture of the soil and reports that back to me via an ESP-12 SOM (System on a Module). The ESP has WiFi capabilities and would log to an MQTT server on my home media server. This way (when I'm at home - none of that IoT shit) I can see how my plans are doing. + +### Current state + +So that's what Revision A of Plantb0t is. I also added a second sensor slot which is meant to be populated by a temperature sensor but could theoretically house a second moisture sensor. In the end the probes are only sticks in the ground that have a resistance between them. + +Here is a dynamic render of the board (that went into prototype production on the 29th of march, 2016). + + + +As you can see it's powered by an ESP-12 and comes with it's own programmer (The lovely CP2102) and micro-USB header. The USB-Port is currently the only way to power the board. + +In the future it is planned to bypass the USB power and only use it for the programmer and otherwise drive everything off an externla powerboard which provides 3.3V for the Plantb0t. + +In the bottom you see two constant current sources that can power two analogue sensors that get multiplexed into the ADC of the ESP-12. + +GPIO pin headers are included for external gismoz such as a pump to act on the moisture data as well as screwholes to mount the whole thing in a 3D printed case. + +In total the board is only 5x5cm big! + +### Future plans + +A few things I want to realise with this project in the next coming weeks: + + - Primarily the Rev A board needs to be tested to make sure that the programmer works + - Figure out a good way to calibrate the sensors. Maybe drive a button via GPIO? + - Design a power board that generates 3.3V for the board (but not the programmer!) from a solar panel and a battery to decouple the entire sensor-board from all power-sockets. + +For the next revision of the board (Rev B) I want to include more sensor slots. Maybe work on the part spacing a bit and increase footprint sizes. It should be easier to solder and someof the parts are ridiculously small. I mean...I have the whole back to work with? + +I also have some crazy ideas for a "Plantb0t+" Version with even MOAR SENSORS (Including a pH-value sensor!). But that's all faaaaar in the future. + +Either way...I'm excited for my boards to get here (hopefully in the next 7-8 days) as well as all the parts I need for the prototypes. + +I leave you with a screenshot from KiCad where you get to see under the hood of the board. Cheers o/ + + + +(The project has a [Github](https://github.com/spacekookie/open_plantb0t) repo where I will try to populate the wiki with as much info as possible)
\ No newline at end of file diff --git a/content/blog/miscellanious/000-newdawn.md b/content/blog/miscellanious/000-newdawn.md new file mode 100755 index 0000000..4acadb7 --- /dev/null +++ b/content/blog/miscellanious/000-newdawn.md @@ -0,0 +1,89 @@ +Title: Static sites vs Wordpress +Category: Blog +Date: 2015-08-14 20:59 +Tags: Dev Diary, Meta + +Regular readers will notice slight differences in my blog from the last time they were here. A lot has changed in the last two years since I started this website. Back then I decided to use Wordpress to host the stuff I wanted to write about because it offered lots of plugins and an easy to use CMS + editor to write articles. But over time I noticed that wordpress, while nice, can be more trouble than it's worth. + +So for the last couple of months I've had a look at different web frameworks. Django, Ruby on Rails and some static site generators like Nikola and now Pelican. The last one is now powering this website. + +The advantages of a static site generator is significantly less server load, simpler design and less overhead with updates and security issues. With no database to hold any kind of data and only static `html` documents being generated at "compile time" and dealt via a very simple html server using exploits or manipulating the website to gain access to my server is almost impossible. + +So...while you can read a lot about why static site generators are cool and what their limitations are, I wanted to use this article to write a little bit about the challenges that I had to deal with porting my old website to a new framework. Because I didn't want to start from scratch. I wanted to keep all my stuff. + +### Creating basic site layout. + +Most of my website are blog posts of different categories and streams. For example, I had one page that displayed all blog posts while I then had other pages that displayed only certain categories. Most static site generators don't really support that. On Pelican I am now using one category for Blogs (Blog) and then have other categories displayed on pages that I statically link in my config: + +```python +MENUITEMS = ( + ('Home', '/'), + ('About Me', '/about-me/'), + ('Blog', '/blog/'), + ('Projects', '/projects/'), + ('LibGDX', '/libgdx-game-of-codes/'), + ('Linux', '/linux/'), + ('Teaching', '/teaching/') +) +``` + +So essentially the menu items link to the category pages that then display articles. By default these pages only displayed a list of posts so I had to modify the category template in my theme. But more about that later. + +Right now I still don't have a work-around for having different categories on the "Blog" page. But my current idea is to not link to a category page but rather a "tag" page. Then give every article that should show up on that page the "blog" tag (or use some voodoo setting that automatically adds tags to articles). + +To get pretty links you can set some options to save pages and posts under different URLs in the config: + +```pytho +nARTICLE_URL = '{category}/{slug}' +ARTICLE_SAVE_AS = '{category}/{slug}/index.html' + +PAGE_URL = '{slug}' +PAGE_SAVE_AS = '{slug}/index.html' + +CATEGORY_URL = '{slug}' +CATEGORY_SAVE_AS = '{slug}/index.html' +``` + +### Theming with [Pelican Themes](http://www.pelicanthemes.com) + +Obviously a website should be something personal. And on my old wordpress site I spent several weeks tweaking the Wordpress Theme to my liking. Static site generators like Nikola or Pelican do offer theming where people much better at writing CSS and html put together something nice. I would recommend cloning a template that you want to use (most of them are on Github) and then modifying it however you like. For example, my template and the modifications I'm making on it can be found here: [github.com/spacekookie/nest](https://github.com/SpaceKookie/nest). + +The first thing I did was tweak the dynamic content filling a bit. I mentioned earlier that I had to modify some stuff to make sure that I didn't just get a list of entries in my blog page but rather either the content or a summary of the page. With a little plugin called "summary" that becomes even nicer but I'm getting ahead of myself. + +```html +{% for article in articles %} + <dt>{{ article.locale_date}}</dt> + <dd> + <a href="{{ SITEURL }}/{{ article.url }}"><h2>{{ article.title }}</h2></a> + <p>{{ article.summary }}</p> + </dd> +{% endfor %} +``` + +As you can see you can embed data from your pelican site into HTML via curly braces, aka Jinja. So the above snippet is obviously a loop that takes an article from articles (which is provided by Pelican in my case) and then renders a "h2" link to the article with the title (`article.title`) and then adds the `article.summary`. + +If you wanted to display the entire content of the post it's as trivial as changing the summary to article.content. That's the beauty of Jinja: it's ridiculously easy :) + +### Next up + +Another thing you might wanna have a look at is the static/css folder of your theme. In the one I use there is a `nest.css` file that contains a lot of modifications to the underlying bootstrap theme. Including changing paddings, colours as well as overriding headers to not have these weird dashes (that look pretty cool for some parts. But not so cool for others). + +But that's all details then. Other things you might want to consider if you move your Wordpress blog to a static site generator is that Wordpress sets up a lot of metadata that then ends up in your Markdown files (if you choose Markdown). Which means that you might want to go through all your articles cleaning out unwanted metadata that might just screw things up. + +I'm using Sublime Text for the multi-cursor/ multi-file edits and regex searching which made editing my article metadata less of a pain in the bum. + +As for this site: I still have a lot of things to work out. For example I still don't have a projects page. Markdown is nice for writing articles but I'm thinking about adding Restructured Text (`.rst`) files for static pages. It's a lot more powerful but also more annoying to write. + +**-EDIT-** + +A quick insertion a few weeks after having created but not yet published this site. + +Alternatively, something I've now started using for my front page is a dedicated template for certain pages. You can set a template via `Template: <template_name>` into a files metadata. Then create a corresponsing `<template_name>.html` in your themes template folder. + +That way I can have special settings for certain pages without having to work with embedded if statements in a sinle html template. + +**-EDIT-** + +But for today that shall be it. I hope you like my new website. Enjoy the new comment system as well (which I just moved to Disqus because that's pretty cool). Until another day. + +~ Kate diff --git a/content/blog/miscellanious/001-lonelyrobot.md b/content/blog/miscellanious/001-lonelyrobot.md new file mode 100755 index 0000000..0da001a --- /dev/null +++ b/content/blog/miscellanious/001-lonelyrobot.md @@ -0,0 +1,45 @@ +Title: Lonely Robot and the future +Category: Blog +Date: 2015-08-25 15:30 +Tags: Dev Diary, Lonely Robot + +Hey everybody, long time no read. + +As I returned from vacation on the Chaos Communication Camp 2015 (Not sure if I'll post more about that) and probably starting a new job next week (*pssst* not sure if I should talk about it ;) ) the rest of my summer is still ahead of me and I'm booming with ideas and inspiration to do stuff. + +I've started more intensively coding on the `newdawn` branch of Reedb, the C port of the database and planning some features for the old codebase via the `backports` branch. Because the new codebase will use a different crypto backend (from OpenSSL to gnu_crypt) a migration agent will be neccesary to migrate between 0.11.x to 0.12+ vaults. But as very few people currently use Reedb and most setups are for testing purposes only that isn't a very big priority right now. Depends on how the current version of reedb develops :) + +But that's talk for another day. What else has been going on? After the Chaos Communication Camp 2015 I've been playing around a bit with my rad1o badge. + + + +But not much has resulted from that yet. The distribution I'm using (Fedora 22) at this time unfortunately has a broken arm-gcc package which means that a linker for embedded systems isn't working properly. So hacking on that will have to wait a little bit. But I will very likely post more stuff about that in the future. + +Now, what was this post supposed to be about? Not Reedb. Or my new job. Or the Rad1o Badge or even the cccamp. It's supposed to be about a new software studio I created. + +### Lonely Robot + +So far we have a website at [lonelyrobot.io](https://www.lonelyrobot.io), an issue tracker at [bugs.lonelyrobot.io](https://bugs.lonelyrobot.io) and are expanding our web prescence but mostly working on projects. + +Two things that we currently have going on are an Android game called **Graviton** (which started out as a tech demo that got out of hand) and, more exciting, LRGE, the `Lonely Robot Game Engine`. + +The whole thing got started between my boyfriend and me who wanted to make video games together. And after a few months of day-dreaming, talking about ideas late at night and bitching about the current state of the gaming industry. But only after a few months of talking we actually started doing something. + +Over the last couple of months we've been working with the LibGDX framework, making some minior and other major modifications to it and writing the specification for an engine. + +We decided to use C++ for it and build it on top of SDL (the Simple DirectMedia Layer) with a very modular design which will allow for modules to be swapped in and out. +While we are still early in the planning phase of the engine it is what we want to focus our efforts on for the next couple of months, possibly the next year. + +While we do make (free) software most of our ideas are for games. Our vision is that with this game engine we will be in a position where we can create them. + +We decided to go for a self-written engine over something like Unreal or Unity because of a multitude of reasons. One of which is that neither of us are very great at blackbox development (that is the development of systems with another system that the developer doesn't fully understand). It is a problem I've always had with game modding but have also run into when playing around with the Unreal Development Kit. + +Even with LibGDX I've always wanted to understand the inner workings of the framework which in the end lead me to modifying large chunks of it. + +So that's that. I wanted to write about it here but will probably move other thoughts about the studio to my Lonely Robot dev blog on [lonelyrobot.io](https://www.lonelyrobot.io/blog). + +That'll be it for now. I have an idea for a different project brewing in my head but I don't want to talk about it for now. All you should know now is: `hardware` :D + +Until another day, + +Kate
\ No newline at end of file diff --git a/content/blog/miscellanious/002-reedb-c-port.md b/content/blog/miscellanious/002-reedb-c-port.md new file mode 100755 index 0000000..5fe627e --- /dev/null +++ b/content/blog/miscellanious/002-reedb-c-port.md @@ -0,0 +1,53 @@ +Title: All aboard, it's a C port +Category: Blog +Date: 2015-10-21 18:02 +Tags: Dev Diary, Reedb + +It's been a long time coming. This blog post, not what I am writing about. Though...what I am writing about has also been a long time coming. So in a way, yes, I guess I was right. + +### Anyways. + +Over a year ago I switched to Linux. I was a Mac fangirl before. I loved the ecosystem, the OS, loved the convenience and the idea of having a system that just *worked* and the power of a root terminal. +But something changed. The garden of bliss grew smaller and smaller and as I realised that I didn't own or understand my computer, that Apple was in charge of what icons I used on my Desktop, I was drawn towards Linux. + +I had used the OS on servers before and even my gaming computer ran Ubuntu to play the 64bit variant of *Kerbal Space Program*. But that last step...that took a bit longer. + +One of the issues I was faced with was compatibility of software. And while most of the things I used (Eclipse, Sublime Text, Spotify, etc.) were also available on Linux, one thing wasn't: my password manager. + +I shall not name it by name because I don't want to advertise a product I no longer want to use. But my password manager was a problem. + +So just over a year ago I set out to write my own password manager. > How hard could it be < I remember myself saying. Oh could I have been more wrong? Probably. But it wasn't easy. + +### The Origin + +Back in the day (and I know it's silly because it's only been a year. But still - a lot of things have changed since then) I knew Java, some Python and wanted to learn Ruby. My brother told me about a mobile framework with which I could easily make a mobile version of the password manager and I was convinced: + +I wanted to write it in Ruby. + +7 months of development, feedback, discussions in a variety of hackerspaces with an even larger variety of people later - Reedb 0.10, the first usable version of what was once a password manager and had migrated into something bigger, something much cooler than I ever thought it would be: a database. + +And as I added more and more features and this database became more and more intelligent I suffered from the limits of Ruby. Speed was terrible, packaging was practically impossible and it became obvious to me that Ruby was a language for the web (not because of the language. But rather because of the people that used it. The rails hipsters and web devs). + +After mucking around with it for at least a month, trying several build systems and desperately trying to get this application to work I had enough. I wanted to port it. + +Only around 3000 lines of code it wouldn't be too difficult to port Reedb into a different langauge. After all, most of the work had gone into the design process, not the actual coding. And while I wrestled with myself and tried a variety of languages, in the end I settled for C. + +With it came a wave of problems. Conventient datastructures that just existed in Ruby such as the Hash (`{}`) or dynamic types made it easy to prototype something and quickly work with large amounts of data. All of that was different in C. And it took me a few months to really start to understand the C ways. + +In the end I made a breakthrough with the design process when I finally discovered unions. And it's been a few weeks since. For now the C port of the project lives on the `newdawn` branch of the [github repo](https://github.com/reepass/reedb/tree/newdawn). The issue tracking has since been migrated to the [Lonely Robot Redmine](https://bugs.lonelyrobot.io/projects/reedb/issues) where you're welcome to fly by and check out the progress. + +### The Future + +I don't know how long it will take for me to finish this port. I'm making good progress, wind is in my favour. But there are still questions to be answered. Especially when it comes to the encryption of things. + +But overall I'm happy with my descision. C is definately different. But it's a good kind of different where you can feel the control you have over your code. The performance is brilliant and a pure C binding makes integrating it into other languages or writing extentions for it as easy as pie. + +I'll wrap this article up for now, it's gotten rather long. I hope that I can post updates about Reedb soon. Plus, I also have some other cool stuff lining up in the hardware section of my mad projects. I still want to do the Omnitool project and I'll definately keep that series alive. But it's a rather large undertaking. And I want to get some experience with smaller projects before I try to do something as mad as that. + it isn't exactly cheap to fuck up a prototype ;) + +--- + +Anyways, I have to sail back into the C... + +Kate + +P.S. Sorry for the bad sailing puns. I promise there won't be any more, for shore. diff --git a/content/blog/miscellanious/003-recovering-a-luks-container.md b/content/blog/miscellanious/003-recovering-a-luks-container.md new file mode 100755 index 0000000..5cb6b29 --- /dev/null +++ b/content/blog/miscellanious/003-recovering-a-luks-container.md @@ -0,0 +1,59 @@ +Title: Recovering a destroyed LUKs container +Category: Blog +Date: 2015-11-19 11:41 +Tags: Dev Diary + +So...funny thing happened to me the other day. And by funny I mean not funny. Actually I mean quite the oposite of funny. I booted my computer after a normal night of sleep (after I had actually shut down my laptop after several weeks of activity) and...nothing. + +I stared at my plymouth boot screen while nothing prompted me to type in my passphrase to decrypt my harddrive and the first thought through my mind was: + +> Fuck...I don't have backup. + +### How to debug + +Now...not to worry, after some time I was dropped into a recovery console where I could ask very simple questions like what kernel modules were present and what Systemd had been up to. And at first I thought the problem was clear: `Module failed to load: vboxdrv` and other messages populated my screen – all about VirtualBox kernel modules. + +So the problem was clear. I had fucked up something when installing a new kernel or VirtualBox or anything else. So I blacklisted the modules and moved on...just...that it didn't. The problem persisted. Thinking that I had fucked something up when dealing with the GRUB config or the GRUB recovery console I got my trusty Fedora 22 live-USB out and booted off that. + +### How not to panic + +To the realise that my 256GB SSD was only 500MB full (which was rightfully detected as an `ext4` formatted volume. The rest of my drive was marked as `unpartitioned space`. \jawdrop. + +Now...here is where things get and got interesting. But first let's have a look at my setup. + +``` +sda (the actual drive) +├── sda1 (ext4, mounted as /boot, contains my kernel) +└── sda2 (LUKS Encrypted Volume, contains subvolumes) + ├── vc-root (RootFS) + ├── vc-home (HomeFS) + └── vc-swap (guess c:) +``` + +So as you can see my boot drive is outside the LUKS container and unencrypted which was why I even got the chance to enter a recovery console. The rest of my system is encrypted. And seeing that only sda1 was being picked up it meant that the partition table on my disk must have had been destroyed to the point that it no longer knew sda2. + +Knowing this didn't help very much though and it took me a few hours to fix this. + +### Restoring the Partition Table + +So the main problem was that my partition table was broken. I don't want to start speculating as to why this happened. Maybe my SSD just lost a few blocks, maybe it was bombarded by solar radiation or maybe (just maybe) I was obducted by aliens in the night, refused to give out my master passphrase in my sleep and because of frustration of not being able to get to my data they deleted some junks from my partition table just to spite me. + +Either way, a combination of two applications saved my life and hopefully will save yours. + +`testdisk` and `cfdisk` + +At first, make sure you have backups ;) And don't blame me if you fuck it up. Also you need to know EXACTLY what your layout is to restore this. Otherwise BAD THINGS WILL HAPPEN *waves hand around warning-ly* + +Run `testdisk` on your drive, enter through the screens, let it do a deep search and just say yes to everything it wants to do. This restored the LUKS header for me again at which point my computer at least started seeing the encryption container again. Didn't mean I could log in because keyfiles couldn't be found (they're not in the header apparently). + +After that, I ran `cfdisk`. What this program does (or can do) is rebuild your partition table. After letting testdisk have it's go it found my LUKS header and completely destroyed my ext4 bootpartition. So in my case this is what it looked like. + + +What you will want to do is hit NEW, select the correct size of your partitions. Depending on how running testdisk went for you it might have found different parititions, all of them or none. Please! For the love of god, make sure you get your sectors right. Becase if you don't it will seriously damage your system and might make it completely unusable. +In my caseit was easy, I filled in my boot partition, marked it as bootable and set it's type correctly, also fixed the type error where before sda2 was being picked up as an LVM and not a LUKS container (this is obviously from my running system). And this is what I ended up with. + + +Make sure you write your changes, exit and reboot. And if you did everything right, you will have a working system again. +And that's that. I hope this article will be of use to someone at some point. And remember: make backups! + +Cheers o/
\ No newline at end of file diff --git a/content/blog/miscellanious/004-getting-started-with-xmmpp.md b/content/blog/miscellanious/004-getting-started-with-xmmpp.md new file mode 100755 index 0000000..edbc829 --- /dev/null +++ b/content/blog/miscellanious/004-getting-started-with-xmmpp.md @@ -0,0 +1,91 @@ +Title: Getting started with XMPP/ Jabber +Category: Blog +Date: 2015-11-24 12:14 +Tags: Dev Diary + +So after having spoken to a friend the other day and trying to get him to start using XMPP (aka as Jabber) instead of facebook messenger I realised that even while I thought it was trivial to set up other people might disagree with me. So here a little guide :) + +Now...this isn't just a step-by-step instruction of what to do. In fact this article is more about getting you to understand XMPP than registered with a specific server. XMPP (formerly known as Jabber, just so you know why some people use the terms interchangably) is a chat protocol, not a chat service. It is based on XML and was originally created for near instantanious message delivery (chats). However since then it has been extended to also be able to VoIP and video (more or less good, that all depends on your client). +So XMPP is a protocol that anyone can just use. However...when I say "get started with XMPP" I don't mean set up your own servers and chat system. I mean Jabber, which is still the name of a chat service. + + +### A network instead of a server + +The current open Jabber infrastructure is built around a bunch of servers (actually quite a **lot** of servers) that can communicate with each other. While inside the network it doesn't matter what server someone is registered with. As long as both people are part of the network they can chat with each other. And that's that. (Cryptographically and version wise it becomes a bit more complicated than that. But for the end user, that's what it boils down to). + +There is a list of servers in the Jabber network (also called directory) available [here](https://xmpp.net/directory.php). As you can see there are quite a few servers out there that will allow you to register. Now...when picking a server please make note of a few things. + + - Check the software has had updates in at least the last year. You don't want to trust your private chats to outdated software, especially because that will usually mean that the server admin can't be bothered to update to newer versions of plugins and protocols. + - Check that the server passes both "Server to Server" and "Client to Client" security tests. (Both in the green). + - You like the domain. You don't want to have a domain "@kinkymotherfucker.com" if you don't like it :D + + +### Register with a server + +So let's asume you've found a server you like. Scolling through the list I would probably register at [blah.im](https://blah.im). Note that you will have to import an SSL certificate. You will have to if you don't have the CA [Cert Root certificate installed](https://www.google.com/search?q=What+is+CA+Cert&ie=utf-8&oe=utf-8#q=What+is+CACert). + +But you will also notice that the server doesn't actually have anything on it's website. To register on servers in a lot of cases you will need a jabber client to do it for you. I am using Pidgin and will thus also demonstrate it with that client. Pidgin is free open source software and runs on almost any platform. But feel free to use a different client if you find one you like more. The features should all just be transferable. + +To install Pidgin please go to their website and follow download and installation instructions for your platform. For mine (Fedora) it's as simple as typing `dnf install pidgin`. I will assume you managed to install it and we +move on :) + +In Pidgin navigate to accounts and Manage your accounts. + + + +In the opening window click on **add** and then select XMPP from the list of possible accounts to add. +Fill in your desired username, the server you want to register with and a passphrase. You can leave the resource blank. Also make sure you tick the box "Create this new Account on the Server". + +Servers provide different ways to register. Some just have a registration webpage, some have an API that pidgin can talk to. Some make pidgin open a browser window and guide you to their registration site. This is something unique to the server you choose to register on. + +In the case of the *blah.im* server pidgin opens a new website where I can register my nickname (which I will not do because I already have an account I like to use). Check your input with the example picture below. + + + +And that's that. You should be registered and ready to log in and chat with other people who also use Jabber/ XMPP, no matter what server they're on. + + +### Encryption via Jabber (OTR) + +Jabber by itself can be secured via SSL and several transport layer security measures but that makes it no more secure than any other service. The server provider can still read all messages and log them without you ever knowing it. + +Because of that a lot of people use separate encryption with Jabber called "Off the Record", short "OTR". + +What OTR does is encrypt messages on your computer, sends them to your friend and then locally on their computer decrypts them again. This has however two drawbacks. + + 1. Both you and your friend need to be online to chat with each other over OTR. You can't send them an offline message and let the server cache it until they come back online to read it. + 2. OTR does not support multiple devices. That means you can't start chatting on your PC, have to leave and pick up the conversation on your phone. You will need to start a new conversation. And a lot of mobile clients don't properly support OTR because they shut down the session when you lock your screen. + +To address both these issues there is a new crypto protocol called "Axolotl" which fixes both of these issues. Axolotl is however a generic protocol and can be used with literally anything. To adapt it to XMPP and integrate it into the already existing infrastructure of servers there is a second protocol called "OMEMO" which implements Axolotl for XMPP. It is however still very new and *very* few clients support it at this time. In fact, the only Jabber client I know of is **Conversations** on Android. + +But let's assume the downsides of OTR don't bother you (they don't bother most people). How would you go about using it? OTR in Pidgin is a plugin that needs to be enabled. Depending on what platform you install it to you might have to install the plugin yourself which can be more or less work. (On Fedora it's just `dnf install pidgin-otr`) + +But [you can figure that out yourself, I hope.](http://lmgtfy.com/?q=Install+Pidgin+OTR+plugin). + +*Waits for you to install the plugin* + +Good, now what needs to happen. First you need to active the plugin and generate yourself a key. Go to **Tools** and then **Plugins**, search for the OTR plugin and enable it. Then go to it's configuration page. + +<img class="dual" src="/images/jabber/pidgin3.png" align="left"><img class="dual" src="/images/jabber/pidgin4.png" align="right"> + +You will need to generate a key. A key in this case means a private key, if you're already somewhat familiar with cryptographic concepts. It's a key that is unique to you, should be protected, private and *never* shared with anyone. It is thus called your **private key**. + +When clicking the button to generate a key Pidgin will make one for you and save it somewhere on your filesystem. It allows you to encrypt and decrypt data (chats, files, etc.) that people send to you. +Afterwards a key fingerprint will show up and the generate button will be greyed out. + +A key fingerprint is sort of a signature. It can identify you as you. So if someone sees your fingerprint they can be sure they're talking to the right person (if they've verified the fingerprint via another medium, e.g. meeting in person). But the fingerprint doesn't expose any secure information about your key. + +You can also change some basic information about how OTR should work on your system. I won't go over these for now. + + +And that's it. You're done. You can initiate new private conversations with people via the **OTR** submenu in the chat screen. And know that everything you say to the person in that session is secure. And here is the best thing: OTR provides something known as "forward secrecy". That means that if at some point someone steals your laptop or phone and you loose your private key that doesn't mean that, even if someone logged every single piece of encryted text going over the network between you and a certain person and has all the information needed to theoretically decrypt your messages, they can't! + +Because while you chat with OTR the key continously changes. So if you ever loose your key, you don't have to worry about old chats becoming public or visible for others to see. + +(As long as you don't log them in cleartext of course). + + + +#### Happy chattin' + +~ Kate
\ No newline at end of file diff --git a/content/blog/miscellanious/005-getting-started-with-reedb.md b/content/blog/miscellanious/005-getting-started-with-reedb.md new file mode 100755 index 0000000..ac7824c --- /dev/null +++ b/content/blog/miscellanious/005-getting-started-with-reedb.md @@ -0,0 +1,127 @@ +Title: Hacking on Reedb +Category: Blog +Date: 2016-03-14 10:43 +Tags: Dev Diary, Reedb + +So...it's been a while :) Exams are over, code has been written, bugs have been fixed, frustrations have been had. Terrible christmas gifts have been sold off on ebay and found a new owner with a misguided sense of style. I've also gone a bit mad with one of my other projects: The Christmas bauble. As you might recall it started as a harmless joke to learn KiCAD, ended up actually being manufactured (I never saw that bonus from dirtypcb for mentioning them a lot :c) and has now gone into planning phase for Revision B. + + + +But more on that later. I'm back, in the spring with new energy and drive. To talk to you about Reedb (yet again). + +With `0.12` coming closer and closer to being a reality I wanted to quickly draft up something how to interact with Reedb. The API is basically stable at this point and while the C-binding isn't quite done the C++ interface is ready to be used (and almost actually hooked up :) ). + + +### Initialising a context + +Getting started with reedb requires a context which holds a bunch of information about what vaults exist, what tokens haven been scoped, users active, watchdogs, etc. etc. + +In addition to that there are vault interfaces that get attached to a context that can then actually interact with vaults. This way different vaults can get handled by different interfaces that are completely separated from each other. + +```C++ +reedb *rdb = new reedb(); +rdb->set_os(LINUX); +rdb->set_distro(SYSTEM_D); +rdb->set_verbose(true); +rdb->finalise(); +``` + +The OS and sitro flags determine what configuration paths and formats are specified as well as how to launch reedb at system startup (if such a behaviour is wanted/ set up). + +After defining all the parameters required or wanted to initialise the Reedb context call finalise to make it official and make the context usable. Before `finalise()` is called, trying to access other functions via the context will result in an error being thrown. + +### Vault interfaces + +So after having a Reedb context you have to register a vaults interface to it. Multiple interfaces can be registered and separated which means that certain vaults can be accessed that require different settings (for example a minimum passphrase length). Generally it just offers more flexibility to the developers. + +```C++ +rdb_vaults *v = new rdb_vaults(); +rdb->register_vinterface(v); + +vault_meta meta = v->create("fancy_vault", "~/Documents/", "MyD0gisnot!mypassword!"); // P.S. I don't have a dog :) +``` + +The create function will generate a key, encrypt the Master key with the provided passphrase and dump it to disk. In addition a folder structure and configuration is written. The config is mostly future proofing - none of the values are actually currently used. But it will hold information about zones, users and cipher modes in the future. + +--- + +After creating a vault you still need to authenticate on it. The unencrypted key might still be held in RAM (in secure memory that is) but just because you created a vault doesn't automaticaly mean you have access to it. So after calling `create` you need to call: + +```C++ +// A token is malloced for you in secmem. Do not free it yourself. Let Reedb do it for you! +rdb_token *token = v->authenticate(meta.id, "MyD0gisnot!mypassword!"); +``` + +You need the UUID from the interface we are addressing the vault via - we can find the UUID in the vault_meta we were handed during creation. Alternatively we can ask the vaults interface. + +From the docs: +> A UUID is provided from the management wrapper and isn't stored in the vault itself. A vault doesn't care about its own ID, nor does it even know it has one. +> +> Do not try to hard-code UUIDs into your program as they might be non-persistent across runtimes. + +Authentication only takes the ID and passphrase at the moment. However a user-auth function will be added in at least the next version. Both return a token that will be required for **every** operation that follows. + +And that's it...you can now interact with your shiny new Reedb Vault :) + +```C++ +std::string file_id = "Reedb.org"; +map<std::string*, std::string*> content(); +content["Username"] = "Peter Pan"; +content["Passphrase"] = "flower123"; + +/* Then take all that data and insert it */ +v->insert(meta.id, token, &file_id, &content); // Takes the pointer to a content map to save memory during inserts. +``` + +As you can see you need a vault-id and a token to even be allowed to the next step. Then to insert a piece of data you need to give it a name. Reedb is object-oriented which means that every dataset has a name and is an "object" on the FS ( Blockdevice mode is in planning :) ). So from that day on your piece of data will be available if you query for "Reedb.org". + +```C++ +map<std::string*, file_meta*> data; +data = query_file(meta.id, token, "Reedb.org"); +``` + +That will put a query return into your map. A query return isn't quite data. It's basically a name of a data-set mapped to it's head. A file head contains a bunch of metadata that isn't exactly deemed "important". Like it's name, a category, some tags and whatever else you might want to save in there. +In fact you can extend header fields at will. + +```C++ +map<std::string*, std::string*> meta_delta(); +// ... +v->migrate_headers(&meta_delta); +``` + +From the docs (again): + +> A meta_delta is the name of a meta-field that should be inserted mapped to its type in a std::map<?,?>. +> If a meta should be removed set the type to "-1". +> +> When removing meta fields from active vaults data needs to be migrated via rdb_meta_migr(...). Also be aware that removing active meta fields can cause terrible memory corruption. Be warned! + +A file_meta is exactly that: a vault header. It can be further searched and filtered with RQL (Reedb Query Language) that we will not go into further in this blog post. Just know that it exists :) + +*hint hint* `"$CATEGORY: [Social | Website | Online] $TAGS:[Private & Friends] $NAME: ~[Face]"` :) + +Deleting, updating files and updating vaults is analogue to what we already saw. Basically you always keep your vault ID and token on you, then provide the interface with some data. + +Some of the steps might seem a bit verbose but that's just so that the user (aka developer) gets maximum control over what she is doing with her code. It also allows for more precice error handling - narrowing down the source of the error further for the end-user. + +### A tiny last thing + +There are two interfaces for Reedb. A C++ and a C one. And you pick which one you want to use by either doing + +```C++ +#include<reedb/core.hpp> +``` + +or +```C +#include<reedb/core.h> +``` + +The C Interface is pretty much analogue to the C++ one (with obvious slight differences). + +```C +vault_meta *meta; +rdb_vaults *vaults = rdb->create(&meta, ...); +``` + +That's it for today. I hope this article gave you a quick introduction to the native interface and makes you at least a little curious or excited to work with it :)
\ No newline at end of file diff --git a/content/blog/miscellanious/006-another-blog-update.md b/content/blog/miscellanious/006-another-blog-update.md new file mode 100644 index 0000000..d1e75a4 --- /dev/null +++ b/content/blog/miscellanious/006-another-blog-update.md @@ -0,0 +1,27 @@ +Title: Winter update +Category: Blog +Tags: Dev Diary, Meta +Date: 2016-12-2 10:43 + +Howdy everybody! + +As the year is winding down and we're all getting ready for the jump to take us out of what has (in my opinion) been a *very* shitty year, I looked at my blog and could only shake my head. + +I had moved this over from Wordpress to Pelican and basically replicated all of the layouts to the extent that some of Pelicans own functionality had to be abused to make it work. But as I kept publishing things on here I realised that most of the features I had implemented went unused. + +And so, for the last few days I have tweaked the layout (and design - as some might notice) to be a bit more traditional again. + +I was also considering to change theme but after not finding anything I liked I decided to hack the fuck out of my current one instead. You can check out all of my horrible changes [here](https://github.com/spacekookie/nest). + +I've also finally done some stuff that I've wanted to do for ages - such as pimping up the front page, adding a proper projects page and go through some of my old tutorial series, fix their formatting (yea right "perfect wordpress import...") and update them to newer API's of libraries. Some articles have just been dropped because I would have had to re-work their formatting and they were no longer relevant. Stuff will slowly be introduced again, with proper formatting :) + + +### Everything else + +In the terms of literally everything except my blog: I'm looking forward to the **33C3**. I'll be joining with the c-base assembly. My first congress in almost a decade! Expect maybe an update from that. And maybe there might be some christmas hacking. It's always more fun to do silly RGB LED stuff if it ends up annoying people on the tree! + +Also, with the blog now in a bit better shape I will try to keep a closer journal of what I'm doing. But hey...no promises, right? ;) + +I shall leave you with this piece of relaxing GIF. + +<img class="padded" src="http://i.imgur.com/KZquOZM.gif" align="center">
\ No newline at end of file diff --git a/content/blog/miscellanious/007-post-congress-2017.md b/content/blog/miscellanious/007-post-congress-2017.md new file mode 100644 index 0000000..fd951ff --- /dev/null +++ b/content/blog/miscellanious/007-post-congress-2017.md @@ -0,0 +1,67 @@ +Title: Post 33C3, what next? +Category: Blog +Tags: Dev Diary, Congress, CCC, Hamburg +Date: 2017-01-03 12:28 + +Howdy everybody, + +I just came back from the annual hacker conference in Hamburg, Germany known as the "Chaos Communication Congress" (or CCC for short). It was the first time I was there for the entire venue and the first time I was able to go at all since *2008*. So yay! + +It was a lot of fun and I have a lot of nice memories to hold onto now. I talked to a lot of interesting people, learned new things, got inspired to do new things and continue on old things. + +More importantly, I loved the chance to get in touch with some other women in the tech industry (via Haecksen & Queer Feminist Geeks), talk about problems, attempt to come up with solutions and just generally rant about things :) + +I also found out that I am in no way, shape or form a dancing person. Although electronic club music is fun! + +## Some talks I went to + +Following is a non-comprehensive list of the talks I went to. I am filling this from memory, so some talks might have been missed or dropped. And maybe I'll just edit them in later without anyone ever knowing. + +<br/> + +**[How Do I Crack Satellite and Cable Pay TV?](https://fahrplan.events.ccc.de/congress/2016/Fahrplan/events/8127.html)** + +A really quite epic lecture about using glitching to extract keys from a very dated security layout. Not that anyone should do this (it's not worth doing it anyways...never anything good on) but it will teach you a lot of stuff about hardware security + +**[Bootstraping a slightly more secure laptop](https://fahrplan.events.ccc.de/congress/2016/Fahrplan/events/8314.html)** + +A talk about the flip-side of TAILS which aims to introduce trusted computing into a world where the machine can't be trusted. HEADS on the other hand uses coreboot and cleverness to create a verifiable machine environment to build an OS on top of. Made me want to get an old thinkpad on ebay to play with :) + +**[The Nibbletronic](https://fahrplan.events.ccc.de/congress/2016/Fahrplan/events/7925.html)** + +A relatively short talk about the creation of a musical instrument. Learning by doing and failing. Quite interesting for me as a hardware designer (as a hobbyist) but also a musician. + +**[Shut Up and Take My Money!](https://fahrplan.events.ccc.de/congress/2016/Fahrplan/events/7969.html)** + +If you have a bank account with N26...stop having a bank account with N26. Their security is absolutely horrible. And while, yes, all of these security issues have been fixed, it shows a rather lacking attitude towards security from their engineering team. Best demonstration of client-side security gone wrong. And why ReST APIs are fucking aweful! + +**[Untrusting the CPU](https://fahrplan.events.ccc.de/congress/2016/Fahrplan/events/8014.html)** + +This was a great talk given by a close friend of mine about one of his super crazy projects. The idea being to construct an FPGA powered PCI-E device for laptops and/ or desktop computers that intercepts messages to the display, encodes and decodes text into them to provide an interface for encrypted messages without using the CPU. It's really quite interesting and I can't wait to see what he does with it. + +**[Making Technology Inclusive Through Papercraft and Sound](https://fahrplan.events.ccc.de/congress/2016/Fahrplan/events/7975.html)** + +One of my favourite talks was about an engineering toy kit that was aiming to be more inclusive. The problem it attempts to tackle are the incredible low numbers of women in computer science and engineering (significantly lower than in other scientific fields). There are a lot of reasons why women aren't well represented in the fields and they are all cultural. This talk was about trying to change the culture around teaching people about electronics and code to be more inclusive towards groups of people (mostly girls/ women) who would otherwise be missed. + +I really enjoyed the talk on a lot of different levels. One was the technical aspect of creating a childrens toy on the cheap that is inclusive and universally programmable through audio encoding. Quite worth a watch. + +I don't think that just with projects like this the culture around women in tech will change. But it's a start. What we realistically need is a change in culture throughout all layers of society. I think the problems around women in tech are quite complicated. And unfortunately usually result in a bunch of assholes starting to shout either about how feminism is evil or how diversity isn't important. And biases aren't actually thaaaaat bad, right? */s* + +I could rant here forever and it's questionable how many people would actually care :) I can recommend this talk. Let's leave it at that :) + +**[The Moon and European Space Exploration](https://fahrplan.events.ccc.de/congress/2016/Fahrplan/events/8406.html) and [Interplanetary Colonization](https://fahrplan.events.ccc.de/congress/2016/Fahrplan/events/7942.html)** + +Those were just the first two talks from a series of space talks. The first one was from one of the heads of ESA about their plans to colonise the moon for profit! And science of course... It was quite funny and definately worth watching. + +The second one I almost liked more, though mostly the first part of it. Liz George manages to explain incredibly well in a very short amount of time what challenges exist when discovering exo-planets. The second part (by somebody else) is a bit more vague about how to actually get there and is less science, more fiction. But hey :P + + +## Going into 2017 + +So in short: 33C3 was pretty epic! And I honestly can't wait for next year. It's not clear yet where it will be held but it will be epic non-the-less. And who knows, maybe I have a talk to hold by then :) + +Which brings me to this year. Last year was fucking shitty. Politically...On a personal level it actually went quite well. And I got a lot of shit done. I did Google Summer of Code, I made *huge* progress on my game project (yes, I will post about that at some point). And especially in the last months of the year, I redesigned and rerouted the Open Plantb0t board. On january 1st, 2017 the revision A2 design went into production. + +I hope to get all my parts together soon and build up a second prototype series which (hopefully) works better than the last ;) I will keep y'all updated on that. + +Until then, I hope you've had a happy new years eve and not an all too terrible year...yet ;)
\ No newline at end of file diff --git a/content/blog/miscellanious/008-libgdx-super-ui.md b/content/blog/miscellanious/008-libgdx-super-ui.md new file mode 100644 index 0000000..c86499d --- /dev/null +++ b/content/blog/miscellanious/008-libgdx-super-ui.md @@ -0,0 +1,116 @@ +Title: LibGDX UI utility: Super UI +Category: Blog +Tags: Dev Diary, LibGDX, Game Dev, Lonely Robot +Date: 2017-01-24 00:14 + +**Let me tell you a factual statement** + +*UI programming is terrible* + +**Let me tell you an even more factual statement** + +*UI programming in LibGDX is even more terrible* + +I am a big fan of LibGDX. It's a really nifty library/ framework to get started with game development if you're more comfortable inside a code editor than a full blown game engine that is more targeted towards designers and artists. And I put my money where my mouth is: I have a series about LibGDX development for beginners on this blog and work almost exclusively with it when it comes to my own projects. + +Yet, there is something that bothers me and there didn't seem to be a great solution to fix it. UI code structure. In this post I want to highlight a utility I have written for LibGDX which is very easily embeddable into your existing projects which will you help structure UI code more efficiently. + +## The root problem + +The reason I dislike UI programming with LibGDX is that it usually results in very long code files or passing dozens of parameters into sub-classes that are needed to update the UI for button presses, etc. + +This goes so far that I have written an editor for game assets before just to realise that (once the development was complete) it had become completely unmaintainable and I had to start from scratch with better structure. It is incredibly easy to just throw out a UI design with Scene2D and LibGDX but unfortunately it is equally easy to produce very bad code which will turn into a big spaghetti mess. + +Let's look at an example problem that I wanted to solve. + + + +Looking at this structure we have three main components that interact with each other. We have a class that handles UI logic (setting up actors in tables, adding listeners, etc), we have a window state which in the particular case which made me write an alternative was a "Lobby Handle" which coordinated what players were going to enter a match, the map, game mode and if everybody in the multiplayer match was set to "Ready". Lastly we have the actual network signal handlers that listen to TCP/ UDP packets and execute code to write/ read from the window state as well as update UI elements. + +Implementing this structure with Scene2D and LibGDX will result in a lot of very ugly code. Because the network signals need to know everything about the UI (how it is structured, etc). And our window state can be written to by two different sources which means that we need to mutex it to avoid race conditions. + +So, what was I trying to solve? First a bit of limitation of scope. Because a lot of UI problems have been solved over and over again and usually at the cost of runtime performance or with a *lot* of extra code. + +1. UI code doesn't have to be embedded in a screen +2. All UI code can access the shared context of the screen +3. UI elements can update each other +4. Clean API that can be called on from anywhere (with a reference to the handle) that triggers range of functions. + +So with that in mind, this is what I did. + +```java + +class MyUIHandle extands UIHandle { + public static enum UI implements UI_BASE { + PLAYER_LIST; + } + + { /** Initialiser block for new objects */ + + registerHandle(new PlayerList(), UI.PLAYER_LIST); + // ... more handles + } + + @Override + public void initialise(Stage s, Object ... var) { ... } + + public class PlayerList extends UIContainer { + + @Override + public void initialise(Stage s) { ... } + + // Define more API here ... + } + +} + +``` + +When we initialise a new `UIHandle` the initialiser block will create our `PlayerLists` and register them with the `UIHandle`. That code is hidden away from you. You can see that we're implementing a different enum type that we overload with values so that we can address submodules via a compile-time checkable value (such as enums). From inside (and outside) this class `UIContainer's` are available via `handle.get(UI.SUB_HANDLE)`. Obviously keeping your enum labels short will make your function calls snappier :) + +The following graphic will sort-of explain the layout in more detail. + + + +What you might also notice is that the `UIHandle` has an initialise function with variadic parameters while the `UIContainer` class only takes a stage. That is because window context is stored once in the `UIHandle` and then accessable from all `UIContainer` classes. This way we only need to do the inversion of control pattern once instead of for every sub-component. + +You can keep the `UIContainer` classes outside this code-file. Then you might however want to provide a construct that does another inversion of control so that an external `UIContainer` can access the context provided via initialise! + +```java + +public class PlayerList extends UIContainer { + + private MyUIHandle parent; + public PlayerList(MyUIHandle parent) { this.parent = parent; } + + // ... +} + +``` + +Now let's talk about that public API. In our original example we wanted to have networking code update some UI elements. And we want UI elements to update other UI elements. So first of all, we keep context in each `UIContainer` about what UI elements are accessable to it. So what we can do in every of our submodules is this: + +```java + parent.get(UI.PLAYER_LIST).updatePlayers(playerList); +``` + +It also means that if we get new data from – say – a network socket or AI simulation, we can very easily update data in some random UI element. + +```java + handle.get(UI.PLAYER_LIST).populate(playerList); +``` + +So all in all, we have solved the following problems: + +1. We have access to all game state in the UI code without passing too many parameters into lots of sub-classes +2. UI code can be moved into lots of files for easier understandability +3. Context isn't duplicated +4. UI code can update other UI code without needing a direct reference to it. + +The individual `UIContainer` instances are essentially independant of each other via dependency injection. + +This library isn't done yet. Most of this is kinda hacked together to fit into **my** game. But I'm interested in making it more generic and putting it on Github. Especially because I can see myself using it again in the future. + +Hope this might be useful to somebody out there. If you have questions, comments, hatemail... + +[Twitter](https://twitter.com/spacekookie) or [E-Mail](mailto:kookie@spacekookie.de)
\ No newline at end of file diff --git a/content/blog/miscellanious/009-moonscript-adventures.md b/content/blog/miscellanious/009-moonscript-adventures.md new file mode 100644 index 0000000..d991a89 --- /dev/null +++ b/content/blog/miscellanious/009-moonscript-adventures.md @@ -0,0 +1,60 @@ +Title: Dabbling with Moonscript +Category: Blog +Tags: Dev Diary, moonscript, programming +Date: 2017-05-06 11:55 + + + +Recently I've started learning/ using Moonscript. It's a language that compiles to [lua](https://www.lua.org/) and as such can run in the LuaJIT, an alternative lua engine which allows very easy and *fast* ffi calls into native code. This makes lua code capable of writing very performant applications and games that use native rendering, window creation or general libraries. + +But in my opinion lua has always felt a bit cumbersome. I use awesomewm so I had to write it occasionally to customise my UI layout. And this is where Moonscript comes in. It's a lot of syntactic sugar on top of lua as well as some other concepts such as object orientation which lua just plain out doesn't have. And while yes, you can write good code without OO (*cough* **C** *cough*) it is a nice tool to have in your pocket, especially when writing GUI applications or games. + + +## The language + +```Moonscript +class Thing + name: "unknown" + +class Person extends Thing + say_name: => print "Hello, I am #{@name}!" + +with Person! + .name = "MoonScript" + \say_name! +``` + +As you can see Moonscript is an indentation based language which (in my opinion) combines syntactic elements from lua and ruby together. In the snippet above (which is from the [moonscript website](http://moonscript.org/)) you can see classes, inheritance as well as the `with` keyword which allows you to initialise/ work with objects without typing it's variable name over and over again. + +If you want to learn more about the language, I can only recommend you have a look at the [Moonscript in 15 minutes guide](https://github.com/leafo/moonscript/wiki/Learn-MoonScript-in-15-Minutes) + + +## How to use it + +You can just write Moonscript files, add `#!/usr/bin/env moon` to them and get going. Obviously that's pretty cool for little scripts that you just want to get going. But not so great for larger applications because a) you don't have access to `ffi` via luaJIT and b) it adds additional startup cost. + +So instead for my projects so far (which so far are a [game](https://github.com/spacekookie/dinodino) and a desktop app) I use a `Makefile` to build and run the Moonscript compiler and then execute the `init.lua` with luajit. + +```Makefile +SOURCES := $(wildcard *.moon) $(wildcard **/*.moon) +LUAOUT := $(SOURCES:.moon=.lua)0 + +.PHONY: all run build + +all: run +build: $(LUAOUT) +%.lua: %.moon + moonc $< +run: build + luajit init.lua +``` + +## Wrapping up + +So...I'm kinda excited about this. Most of the code I write is either in C or Java (depending on what exactly I'm doing). And those two strongly typed and compiled languages have served me well and will continue to be my go-to solutions for a lot of problems. + +But I've long been looking for a dynamicly typed, interpreted/ just-in-time compiled language that I can use for anything from little scripts to medium-sized desktop applications. I used to use python for this but have recently (over the last 6-9 months) fallen out of love and developed a rather passionate dislike of it and it's ecosystem. + +My current project will get it's own little article at some point but I don't mind teasing the progress here. I'm writing a new UI for redshift which works with X11 linux backends and is heavily inspired by f.lux on MacOS. It's written in moonscript, with my own forked version of redshift (which I call [libredshift](https://github.com/spacekookie/libredshift)). It's on [github](https://github.com/spacekookie/redshift_ctrl) and licensed under MIT. + +Hope I've made you a little curious about Moonscript. And stay tuned for updates on future projects with it :)
\ No newline at end of file diff --git a/content/blog/summerofcode/gsoc-000-i-got-accepted.md b/content/blog/summerofcode/gsoc-000-i-got-accepted.md new file mode 100755 index 0000000..894bd85 --- /dev/null +++ b/content/blog/summerofcode/gsoc-000-i-got-accepted.md @@ -0,0 +1,25 @@ +Title: I got accepted to GSoC 2016 +Category: Blog +Date: 2016-04-27 18:47 +Tags: GSoC2016 + + + +The title should be self explanatory about that one :) + +But let me go back a little bit. A couple of weeks ago I sat in the basement of my local hackerspace talking to a friend about crypto when somebody joined the conversation, asking if I was a student and if I might be interested in Google Summer of Code. + +After I looked up the project and familiarised myself with what had to be done, I thought it would be interesting to try to apply. And so I did. I wrote a long-ish proposal of what I wanted to do, how I would do it and when exactly I would acomplish my goals. (You can read my original proposal [here](https://storage.googleapis.com/summerofcode-prod.appspot.com/gsoc/core_project/doc/1458924075_GSOCProposal-KatharinaSabel.pdf?Expires=1461863360&GoogleAccessId=summerofcode-prod%40appspot.gserviceaccount.com&Signature=h0y5Nzi7llFNWKzt9%2BLGLvxcAPZ%2FaO7ni1ZyRDA3uFi6PD%2BDBmtIB6RJAr4Ulhv6fe64IFyB%2FI9iuVIYWIInYTmN7pZ9aUxw6TgxgFYguIywfcE2yUZ4o5UKb0PUbwI0Pu7o6mq%2BzSDXqlegpVOgujQ9k2QuTg1T1CqGzSi%2FnC4u6H0mB%2BxzWGGpoBC6rFwkKM1S70gE7hJ0EZpgYWr9H9zKPcwrfPtx99zqb488sH6STGYJf4tFrDRnnr57k2zbSN%2BO17chZtVBjGUYrKoyU6B%2FGB8MexFE6rmYaTCr5AjgqGWm97VCCwZkpHbRiTtFH5yT825G9%2FkRPYHkxsPnCw%3D%3D)) + +In the meantime I actually had a sit-down with my mentor (the person joining the conversation in that basement) and made further plans how to implement things. + +And so this is it. The next month or so I will have time to get to know the code base of the project (although I partially already have), meet more people from the community and generally get into the rythm of what GSoC is. + +I will be posting three blog posts on the official [Freifunk Blog](http://blog.freifunk.net/), one in a couple of days/ weeks, one in the height of the project and one that will go into the aftermath of the project. + +But in the meantime I will be keeping my blog up to date about what I am doing, how things are going, my challenges and things I learn. +In the hopes that people might find it useful and lean things from it. Or just to save my insane ramblings in some narcissistic pleasure...to think that I am relevant in the world :P + +Read you soon, + +Kate
\ No newline at end of file diff --git a/content/blog/summerofcode/gsoc-001-prepping-the-patient.md b/content/blog/summerofcode/gsoc-001-prepping-the-patient.md new file mode 100755 index 0000000..46fd034 --- /dev/null +++ b/content/blog/summerofcode/gsoc-001-prepping-the-patient.md @@ -0,0 +1,37 @@ +Title: First steps...baby steps +Category: Blog +Date: 2016-06-02 19:56 +Tags: GSoC2016 + +So it's been almost two months, the community bonding period has passed, blog posts were written, talks held and slowly but surely I'm working myself into the qaul.net codebase. + +It's always weird joining a larger project and seeing established build setups, code conventions or generally things where your first thought is "I would have done that differently...". But it's really fun. + +I'm currently working myself into [mbed.tls](https://tls.mbed.org/) which is the crypto library which was chosen to power the cryptographic backend for libqaul (which powers qaul.net). + +That includes some code that will probably not make it into a later version of my branch: the debugger. + +### The De-bugger?! + + + +Well...debuger might be a bit of a strong word, it's basically a way to develop core functions of qaul.net without having to start a GUI, going through NetworkManager dialup or oslr bootup. + +There I am currently busy writing a wrapper around a new namespace added to libqaul: `qcry` (short for qaul crypto) and properly integrate all the mbed.tls sources into the library so they can be accessed by libqaul. The idea being that I don't have to leave vim and the terminal to develop on the core cryptographic components such as: + + - Key generation + - Identify generation (with private key fingerprints) + - Identity verification + - ??? + +Only in the last step of the last bulletin do I actually have to involve the GUI of qaul.net. And until that point I wish to not come in contact with it (if avoidable). + +So most of next week will be getting to know mbed-tls as I have never worked with it before. But hey...can't be worse than the gcrypt documentation¹ :') + +Hope to read you soon with more updates (probably rants). + +Kate o/ + +--- + +¹I am sure I will eat my words in 4 weeks diff --git a/content/blog/summerofcode/gsoc-002-final-evaluation.md b/content/blog/summerofcode/gsoc-002-final-evaluation.md new file mode 100644 index 0000000..6c8a4e4 --- /dev/null +++ b/content/blog/summerofcode/gsoc-002-final-evaluation.md @@ -0,0 +1,69 @@ +Title: What I have done in GSoC 2016 +Category: Blog +Date: 2016-08-19 18:13 +Tags: GSoC2016 + +Google Summer of Code is coming to an end. And as the final bugs are getting squashed and more code is being prepared for the big merge, I am sitting here, trying to think of how to represent my work. + +I thought I would write up a little blog post, explaining what I've done and what still remains to be done. + +### The TLDR + +My main contributions are all available [here](https://github.com/spacekookie/qaul.net/commits/qaul_crypto?author=spacekookie) (spacekookie/qaul.net on the `qaul_crypto` branch). I did a lot of small commits. Most of my code can be found in this [sub-directory](https://github.com/spacekookie/qaul.net/tree/qaul_crypto/src/libqaul/crypto). + +In addition to that I ported an existing project (from python) to C to be relevant for future front-end endevours of the client. It's called [librobohash](https://github.com/spacekookie/librobohash). I didn't end up finishing the port because there were more pressing issues in qaul.net and the UI was delayed. + +While most of my work has been in hidden backend systems there is a demo you can run. The source compiles and has been tested under Linux (Ubuntu 16.04 and Fedora 24) and is located under the `src/client/dbg/` directory. The demo creates two new users (to simulate communication between two nodes), adds the public keys to the keystore and then continues to sign and verify messages. If the demo returns lots of "0" and "OK" it went okay :) + +Feel free to play with the demo; for example, switch out `message` for `fakemessage` during verification :) The source for the demo can be found under `src/libqaul/qcry_wrapper.c` + +### The good (aka what I have done) + +<img class="dual" src="/images/gsoc/02_cryptoui.png" align="left"> + +The two main components that I've written during GSoC2016 are internally referenced as `qcry_arbit` and `qcry_context`. They are two modules that make up the new crypto module in qaul.net. + +As I explained in my first blog post on the [Freifunk blog](http://blog.freifunk.net/2016/gsoc2016-wrapping-crypto-module-qaulnet) the Arbiter provides a static API for the rest of the library (libqaul) to interact with the crypto module. + +The context holds the actual magic of holding user keys, signing and verifying messages and (theoretically) encrypting messages as well. + +Possible with this API at this time is to create users, to sign messages with a users private key and to verify messages that are sent to you from other users. Originally it was planned to split the arbiter into the actual API and a dispatcher which would allow for concurrent access to the inner functions. However it was established through tests that the design was overkill and was thus scrapped. + +A keystore was added in addition to the user store already existing in qaul.net to provide an easy way to store public keys (mapped against fingerprints) that are received from flood events on the network. + +In total the crypto submodule adds another ~2.2k lines of code to the project. + + +### The bad (aka what I haven't yet done) + +So far completely un-implemented is encryption. Unfortunately working with the crypto library selected for the task turned out to be more challenging than expected. With almost no documentation and a few very niche examples I basically went through the library line-by-line to understand how it worked. + +As such, my focus was set on signature exchanges at first because the verifiability of messages and the change to address users by their fingerprints was deemed more important. + +My contributions to qaul.net won't end with the end of Summer of Code. The function stubs are already provided and I plan on implementing the encryption features in the coming weeks. + + +### The ugly (aka what I can't do yet) + +Signatures (and also encryption) of private messages (so messages that aren't flooded to everybody) is currently impossible. This is due to the way that the communication system in qaul.net works. + +I have talked to my mentor and he said that they were currently in the process of re-writing the communication sub-system in libqaul. This means two things: + + 1. I need to wait for those changes to be done until I can finish what I set out to do + 2. Some of the code I wrote (hooking into the current communication system) is being made obsolete :( + + +### In conclusion + +What I can say is this: qaul.net has gotten a very big step closer to becoming a more secure network of communication. The crypto submodule is tested and easy to use. What might happen is that parts of the code get merged (the crypto submodule itself) without merging any of the code that hooks into the communication stack. + +I had a lot of fun working on this project and I am looking forward to more contributions. I have a few cool ideas that I want to discuss with the rest of the team and I am glad that I participated in the Google Summer of Code. + +I was interested in open source before and I contributed to my own projects on github. But the experience I gained this summer will be helpful for me, not just for my own work, but to be less reluctant to join other developer communities. + +And I look forward to seeing my code get merged into qaul.net :) + +Read you soon, + + +~Kate
\ No newline at end of file |