clux«4,519

View All

• 

  2

  bumps

   

  textadept 4.0 with CoffeeScript

  Posted on Sep 4th 2011, 17:17 CEST :: 1 Comment

  Having professionally(!) been programming for a while now, I've run into this situation where I'm actually putting quite precise requirements on the tools I use to develop, both professionally (C++) and at home (CoffeeScript with NodeJS). Thus, after having seen some of the amazing benefits of big clunky IDE's for C++ (keypress to go to declaration/implementation) as well as some of its atrocious drawbacks (what slow, badly designed monolithic beasts of programs they are - but you have to use them with big C++ codebases because of the benefits), I have been hunting for good features for my home web development setup.I have tried the clunky IDE plugin route through NetBeans, and a whole lot of simpler ones, because lets face it, CoffeeScript probably isn't going to get full IDE type jump-to-function support soon, so your aim should be the best possible editor with syntax highlighting - and it should be correct syntax highlighting (the gedit plugin fails on regexes for instance). NetBeans does have a CoffeeScript plugin. In addition to syntax highlighting it does regular compiles to help you quickly spot errors, and it tells you intelligently. The program on the other hand is terribly slow, and if you exceed 8 tabs, you have to scroll through them in the slowest possible way ever.
  
  After realizing that the big beast of a program route may not be the best I went through an assortment of simple editors. None of them really impressed me, so I'll just skip right to the point of this post; the editor that actually works well.
  
  textadept. In fact, it not only just works well, it is amazing. I can work with it completely without a mouse, which is an essential when you want to do laptop work with a choice of either the useless mouse replacement, or the replacement that's too rude to mention in polite company.
  
  [uimg]
  
  

  Good features

  
  1. Buffers instead of tabs. Instead of finding the file you want in a huge list of tabs (that you might have to additionally scroll), you press ctrl-b to open the buffer list and you start typing its name. It autocompletes between the list of file you have open and it's super quick - I mean, you already have your hands on the keyboard anyway since you are coding!
  
  2. Easily extensible. There are modules you can hook in for most popular languages. I added my own shortcut key using a few lines of Lua, and I customized the coffeescript module to be able to build and run inside the editor easily with a simple keystroke.
  
  3. Snippets. For instance in CoffeeScript, a default snippet is the shortcut for console.log; c. It doesn't intrusively popup, but it fills in what you didn't write when you press tab. A very sensible autocomplete. There are also some even more amazing ones to do quick for loops: forr + tab will give you for %1(name) in [%2(start)..%3(finish)]%4( by %5(step)). And you simply write name start finish and optionally step by tabbing between the numbered areas. Super, super quick.
  
  4. Autocompletable recent file list. [i.e. open recent essentially becomes another type of ctrl-b]
  
  5. Settings saved in a backupable folder.
  
  6. Super fast. But really, if your text editor isn't why are you even attempting to use it?
  
  
  

  Install Guide

  
  1. Get the latest textadept version (the .tgz one, not source, it is pre-built) + the modules pack that is also in that download list.
  
  2. Get the javascript module
  
  3. Get the coffeescript module
  
  4. Extract the textadept folder somewhere you feel comfortable with and put the modules archive on top of it.
  
  5. Create a 'javascript' folder under this module folder and put the contents of the javascript module in there.
  
  6. Create a shortcut on wherever you put shortcuts / or symlink it to /usr/bin. (Note that you should always start textadept without a second (file) parameter as you want it to remember your session). Start it once to auto-generate a .textadept folder in you /home/username folder.
  
  7. Create a file without an extension inside aforementioned .textadept folder called 'theme'. Inside it, simply put the text 'dark'.
  
  8. Create the directory structure .textadept/modules/coffeescript and put the contents of the coffeescript module in there. Think it has to be in here rather than under the application path because it relies on the javascript one.
  
  9. Edit your init.lua file under .textadept/ insert the following at the bottom:
  

  keys.cw = { function()
    if buffer._type then
      buffer:close()
      gui.goto_view(-1, false)
      view:unsplit()
    else
      buffer:close()
    end
  end }

  This is because when you try to run or compile your code it changes to split view mode to show the output from stdout. You want close (with ctrl-w) to take you back to unsplit view, and for the program to keep you in the right file. Unless you like working with split views, you will find this annoying, and you will need this step.
  
  10. Edit the init.lua file under ./textadept/modules/coffeescript and search for m_run. There is only one command at the moment and it is wrong. Remove that line and put these two there instead:
  

  m_run.run_command.coffee = 'coffee %(filename)'
  m_run.compile_command.coffee = 'coffee -pb %(filename)'

  
  You want -b because that's how it's used before it's passed to node, so you will be able to more accurately determine what line numbers in the corresponding javascript that is being talked about in the stacktraces.
  
  Addendum: In above step; take out the lua function that binds a shortcut to ctrl-j for inserting raw javascript (if messes up goto line shortcut)
  
  
  

  That's all!

  
  A bit complicated, but it sure is worth it! 4.0 just came out, so this felt like a worthy celebration of an editor that gets surprisingly little attention. Check it out. It's worth it.
  
  poopiter
  
  
  Show comments for this entry
• 

  0

  bumps

   

  Good Times With Spam

  Posted on Jul 22nd 2011, 9:32 CEST :: 2 Comments

  Got a spam bot set up to hammer the comment button here on clux.org it seems.
  Hot windows serialz etc. Sorry if people's notification tab gets spammed.
  
  Originally, I set in place an anonymous posting limit so that individuals can't spam anonymously (doing so even required a captcha), so this led to someone registering manually and bot-hammering from there. This wasn't a big problem, as I can still one click to delete all the comments of a user - and user registration is meant to be a more tedious process than that.
  
  The annoying thing, is that both the actions that can be done anonymously requires breaking a captcha, and at this point mine is definitely broken as spam comes in anonymously as well. Which means, well, stronger captcha soon :/. In the mean time, user registration is temporarily down.
  
  
  Show comments for this entry
• 

  1

  bumps

   

  Working With C++

  Posted on Jun 22nd 2011, 18:22 CEST :: 1 Comment

  static pthread_t clux_thread;
  int main() {
    //...ignored 5,000,000 lines here
    pthread_create(&clux;_bg_thread, NULL, Worker::clux, (void*)new Workspace());
    //...ignored 50,000 lines here
  }
  
  //worker_clux.cpp
  static pthread_t clux_bg_thread;
  Changeset * clux_work(BACON * bacon, METAL * metal) {
    metal->listen();
    
    Code * code = new Code();
    while (true) {
      (stash.push(*bacon)) ? ++code_quality : code_quality--;
      Changeset * chng = code->commit(new CIMsg(code));
      if (code->merge_and_build(&clux;_bg_thread)) { //transactional. switches to bg thread while build blocks
        code->push();
        return chng;
      }
      code->rollback();
      code->append(new Code(code_quality));
    }
  }
  
  void * Worker::clux(void * workspace) {
    Workspace place = workspace;
    workspace->init('ubuntu');
    while(workspace->activate()) {
      METAL metal = (conservative) ? clementine('/home/clux/Music', 'Opeth') : chrome('grooveshark.com', metal);
      BACON * bacon = (BACON*)falloc(2*sizeof(BaconContainer));
      
      while(bacon->in_fridge() && metal->find())
        myProduct->development_branch->document(clux_work(bacon, &metal;));
  
      if (!leaving_time) //above loop exited early => no network or food
        marioKart->advertise()->play();
  
      int hours = (day_num%7 >= 5) ? 56 : 16;
      workspace->deactivate();
      sleep(hours*3600*1000);
    }
    exit(-1); //require manual startup
  }
  
  void * Slacker::clux(void * workspace) {
    while(true) {
      (internet && !num_bugs) ? internet->browse() : clux::ponder();
      if (irc->uptime()) (mood=='meh') ? irc->troll() : irc->chat();
      if (marioKart->advertise()) marioKart->play();
    }
  }
  

  
  
  
  Show comments for this entry
• 

  3

  bumps

   

  Achievement Inflation

  Posted on May 30th 2011, 20:14 CEST :: 10 Comments

  As a person who has replaced all sports with competitive gaming vods, it displeases me that most of my gaming time is spent playing casual, mostly graphics oriented games with low gameplay value. As compensation, I have speedrunned N, Trine, Minesweeper, TrackMania and done miscellaneous Quake trickery over the years. This has given me enjoyment and some vague sense of accomplishment for doing something few other have done well. Innovating in a competitive environment feels particularly good when you finally pull it off.
  
  It's therefore interesting to note that few of these games even include a web based hiscore or achievement system - or sometimes even a proper timing system. But the one thing they all have in common, is smooth, responsive, and mostly fair gameplay. It seems that these days, however, this kind of achievement based gaming is mostly being falsely fabricated through steam/xbox achievements (of usually very easy games). TF2, for instance, has almost 400 achievements ranging from very easy, to simply tedious or requires several people. This really isn't a substitute, but writing simple console scripts to automate parts of getting some of them was.
  
  Interestingly, however, most people responded very angrily when I posted about this, lamenting that I had somehow tricked my way into getting absolutely meaningless achievements. Seriously. Achievements are usually not hard, often do not feel like achievements, whereas exploiting the system might. You may disagree with this, and that's fine too. The point is that what resonates with you with some sense of achievement is relative to your personal context.
  
  With respect to gaming I have two defining properties to point out; an unusually long history in scientific academia, and much competitive Quake 3. These two factors define on some over-generalised level what I'm proud of: clever trickery, or exhibits of pure skill.
  
  It's therefore with joy that I actually found hard achievements that rewards both: The Super Meat Boy no-death run achievements. They require excellent micro, but they also reward tricksters by allowing situational character switching. It's a first for me to see developers take advantage of this so well with built in achievements (though, I should criticize them for also destroying speedrunning in a very well-suited for speedrun game by including a broken hiscore system, but that's another matter).
  [uimg]
  While these can become great personal goals to beat, they appear widely unpopular elsewhere; only one no-death run made it into X360 because they were considered too hard for instance. There is clearly an element of danger here. Built in achievements are addictive and the difference between what I sought out to what achievements represent is that they present you with a goal, not the other way around. You feel like you should be able to do many of them, and so you try. This is very natural with the right mindset.
  
  Regardless, I advocate that people should just have to learn how to moderate their gaming time (vs. claims about addictiveness), and accept that there will always be players better than you (vs. too hard claims), but I know how persuasive such incentives can be. On one hand, I did all the meat boy no death runs mentioned above, but I turned down the opportunity to do the two last bandage girl no-death achievements (cotton alley) because they felt too extreme. I would call it self-control, but frankly I just got bored. Good games are few and far between these days though, don't shit down other people's necks just because their addictive. We keep coming back because they are fun, not because we require our fix.
  
  zombie boy took 10+ hours, that was the end of my patience
  
  
  Show comments for this entry
• 

  2

  bumps

   

  End Brain Dump V27

  Posted on May 26th 2011, 21:32 CEST :: 0 Comments

  This is apparently what happens when I experiment with new technology for months and don't post anything. I can't say much more than that I am thankful I at least got one good dump out on the internet. Heh.
  
  Anyone stopping by here are exempt of reading any of this of course. The following 7 posts (which is the likely number of unread updates given there's been, what, 4 months) are a short version of my experiences with certain bits of web technology. Essentially, my road from
  1. PHP.CodeIgniter.SQL
  2. Python.Django.SQL
  3. Python.Django.MongoDB
  4. JavaScript.NodeJS.MongoDB.
  + My experience randomly changing to a the Colemak keyboard layout. I am very happy to see that sufficient amounts of my rebellious spirit remains to do something so unnecessary. Something that ultimately just annoys the fuck out of everyone.
  
  Anyway. According to the age old Irenicus quote on my profile; this is, and always has been, my manual brain organizer. It's likely not useful for you, but you are allowed to peek in for.. I don't know, 2050s historical reasons, perhaps you even need to document a hitherto undiscovered neural condition!
  
  It's interesting to imagine what your legacy will be given the current evolution of the internet. I bet that at some point, there will be some sort of steam-like achievement to do have to read through certain internet journals to unlock a particular secret. I hope for your sake that journal is not mine.
  
  but a man can dream..
  
  
  Comment on this entry
• 

  1

  bumps

   

  Brain Dump V27 Part 7. Ubuntu.

  Posted on May 26th 2011, 21:16 CEST :: 3 Comments

  A requirement of the current NodeJS versions, is essentially, that you develop on Linux. I say Essentially, because it's technically possible to build with Cygwin, but it also technically possible to get your dick trapped in a box opener. My point is, don't do it.
  [uimg]
  I went for the latest Ubuntu, mostly influenced by brothar's success with it. I was so happy with it, that it immediately overtook my daily windows usage. In fact, my windows usage is now limited to photo editing and gaming time. Things just take so much less time here. I can:
  
  1. Open up a terminal, get the programs I want immediately by typing sudo apt-get install X, Y, Z. No browsing kat.ph, manual unraring, obligatory virus check. It just works. Thought Chrome handled updates well on Windows? Try Linux. Also: fuck everything about windows update, and it needing to boot all the time.
  2. For a calculator; open up a terminal. Use the python interpreter.
  3. Need another terminal? Install Guake. Handles multiple tabs; Ctrl-T for new shell, Ctrl-W for close this shell, Ctrl-Left/Right for moving between tabs. It's persistent. Pops down (with last active tab shown) with the press of a button. If you're a Quake fan you'll use the console button, because that's exactly what it looks like, but with adjustable opacity.
  4. Want a mIRC replacement? install XChat, no need to register (as with everything).
  5. Replacement Music player / Video player? No need. Clementine / VLC. The best simple stuff that just works for Windows work here as well.
  6. Task manager? The ps command can list. ps -A | grep chrome will give you the list of running chrome processes. sudo kill is consequently fun. "Are you sure [, windows user,] that you would like to end this process?" GTFO, KILL. Enter. Done.
  7. Copy an entire folder with permissions? Without dragging the mouse? tar cf . - | tar xf - -C /target/dir. This compresses everything (with tar) in the current directory, pipes it to stdout, from which tar extracts it to the target directory. Might seem convoluted, but it reflects many of the fantastically helpful things you can do in the terminal. Send info from and to different programs in one command.
  
  This is my discoveries from less than a month of use. It looks and feels great, and there's replacements for everything. Except games. So dual boot is perfect. It creates isolated environments where one's productive, and the other is good for fragging. My development time + enjoyment is up, and gaming time is down / more concentrated.
  
  hax++
  
  
  Show comments for this entry
• 

  0

  bumps

   

  Brain Dump V27 Part 6. Client-Side JavaScript Madness.

  Posted on May 26th 2011, 10:29 CEST :: 0 Comments

  Much of my previous JavaScript experience was highly coupled with the PrototypeJS framework. This operates by extends many of the built in prototypes so that you can code beautifully in an object oriented way by chaining. Unfortunately, extending built in prototypes is very bad behaviour, so I have since jumped the jQuery bandwagon, but applying UnderscoreJS for getting chain like helper methods. So far it is less error prone than what PrototypeJS was, and animations actually queue up in a sensible way (something I could never get right for all cases with Prototype's animation framework Scriptaculous - as can be seen by hovering/unhovering the clux.org logo intensely).
  
  Additionally, UnderscoreJS includes a template helper function (source looks identical to this). Which means that I could theoretically use it both client side and server side. I have, however, decided to take it to the next level. I will only render templates client side.
  [uimg]
  It's remarkable that I can even consider this, but browser progression has indeed now reached levels that mean developers can simply discount users without JavaScript. But while it can be dangerous (one missing comma, and your page could be unusable), with proper testing the benefits far outweigh the dangers.
  
  1. Everything except the first response can be done in low-bandwidth JSON. The first response requires just a little more. Prepare a 10 line, one-size-fits-all template with a script tag that dumps a stringified object right inside. With proper handling this can be used to render it on the client for the first response as well.
  2. With a REST interface, you can simplify your code so much that creating a mobile version of your site is a matter of just changing some templates/JavaScript. I use Spine for interacting seamlessly with my backend (because it is essentially a lighter version of BackboneJS).
  3. With good use of script loader/minifiers/compilers, you could theoretically serve only one highly compressed file, and load in others pieces of JavaScript as you would need it. Minimizing data transferred and number of HTTP requests. RequireJS combined with UglifyJS seems the way to go.
  4. With HTML5 history, you can make function like a normal website where urls are bookmarkable as well. The implication is of course that every url you create must map onto the one you are getting JSON requests from. But this should be trivial. Simply serve the short template described in point 1, and embed the normal JSON response in the script tag. Catching it is as easy as (req.xhr) ? res.send(c) : res.render('layout',{c:c}); with Node.
  
  a few years ago, this would have been horrible, now it's amazing.
  
  
  Comment on this entry