What are you working on today?

Thread Rating: 9 votes, 5.00 average.
(9 votes)
Results 51 to 75 of 6318

  1. Post
    #51
    Running first Azure deployment

    ** porges crosses his fingers

  2. Post
    #52
    porges wrote:
    Running first Azure deployment

    ** porges crosses his fingers
    Remember, just substitute the words "some server somewhere" for "in the cloud" when you're reading the documentation, and it'll all be straight forward.

  3. Post
    #53
    Phew, A5 out of the way, finally. I intend to clean up the code we used a bit and open-source it, and then begin refactoring it later in the year.

    At work, getting ready to deploy our Puppetmaster so we can start building new servers against Puppet. Still working on the Java middleware server.

    Personal project wise, I need to start a list and set some priorities, but in general, the shit I want to work on includes:

    * blog port to Hakyll
    * Start a website that indexes all the NZ LAN parties
    * completing my Arch Linux based liveUSB for my NAS (may port this to Debian actually...)
    * reformat main PC with properly aligned sectors (herp derp)
    * rig up and write up my first foray into Arduino projects
    * write a simple network server in Haskell or Scala (I'm thinking something that speaks DHCP)

    They also had a "hacking competition" at A5, which was really, really cool. I'm probably going to take the time to replace Virtualbox with VMware, setup Backtrack, and have a go at penetrating a few of my servers .

  4. Post
    #54
    I hate Fitnesse / Slim / whatever it is, so much. I'm not sure how writing integration tests in a table based pseudocode is better than writing it in actual code. Consider that the developers have to write the code that does the actual test, then the plumbing to connect the table based pseudocode to the actual code. ....*twitch*

  5. Post
    #55
    frio wrote:
    * reformat main PC with properly aligned sectors (herp derp)
    ouch.

  6. Post
    #56
    Not too much of a big deal (copying everything to the NAS at the moment), but a pain in the ass nonetheless. Still, it means I've got a window to update the firmware on the hard drives, update the BIOS and overclock it mercilessly without worrying too much about ****ing the system over.

  7. Post
    #57
    I did it on a fairly empty drive and it took hours :F

  8. Post
    #58
    Really? It should take no time at all; I did it to all the 2TB drives in the NAS just a week or so ago. I'm not going to bother trying to keep my current FS, so basically, all that needs to be done is scorching the partition table and creating new partitions on 4KiB boundaries (64 sectors is usually right for the first one). Should take all of 5 minutes + the reinstall :E.

    If you were preserving the partition table and filesystems on the drive though, yeah, that'd definitely take time.

  9. Post
    #59
    Ah yep. I didn't have a big enough drive to move everything onto :P

  10. Post
    #60
    I'm feeling a bit stink because I haven't touched my project in a few days. But I was loaned a copy of 'Rework' by 37Signals and am working my way through that.

  11. Post
    #61
    Hacking contest? Damn. Knew I should have gone to A5. Sounds interesting.

    Good stuff re: Arduino. Once you understand the basics of it, stop using it as an arduino and programme the micro directly over ISP. It's pretty easy to write a basic AVR programme using C with AVR-GCC.

    Once you've got that under your belt give ARM development a shot - I highly recommend it.

  12. Post
    #62
    Edward Diego wrote:
    Remember, just substitute the words "some server somewhere" for "in the cloud" when you're reading the documentation, and it'll all be straight forward.
    If the cloud didn't solve your problem you didn't use enough of it.


    As for what I'm doing, putting the finishing touches on a presentation I'm giving on Thursday, part of which I'm still running experiments tomorrow (at this rate). Managed to uncover something new and fairly well performing this afternoon just before going home, so that's good. Has to do with partitioning data based on the set of responses made so far, and then using greedily induced lists on that partition to find the next thing to ask about ... that's probably the best I'm going to do without writing the thesis .

  13. Post
    #63
    Ooh man, I do enjoy a refactoring that works well. Feels good.

  14. Post
    #64
    Got really annoyed at UTF-16 strings last night. Not related to job, yay!

  15. Post
    #65
    Jihad Jeep wrote:
    If the cloud didn't solve your problem you didn't use enough of it.
    Coworker currently demoing running with 360 cores

  16. Post
    #66
    porges wrote:
    Coworker currently demoing running with 360 cores
    I want

    Woke up at 3am thinking there was something horribly wrong with everything I'd been doing for the past month. There was, but it didn't change any of the results, so I went back to sleep happy.

  17. Post
    #67
    working @ launching a new clothing store! INTENSE DEADLINES!

  18. Post
    #68
    Suddenly got a whole stack of tenant profiles for my pro-bono project and have been putting them into WordPress. Ran into a few problems with contact details from some of them, so fell back onto using ULs with no bullets to list them coherently.

  19. Post
    #69
    porges wrote:
    Got really annoyed at UTF-16 strings last night. Not related to job, yay!
    Muahahaha, feel my pain. Python 2.x really sucks for unicode work.

  20. Post
    #70
    Edward Diego wrote:
    Muahahaha, feel my pain. Python 2.x really sucks for unicode work.
    The worst thing is that Java/C#'s built-in char type isn't a Unicode codepoint, it's a 16-bit UTF-16 unit... and anything outside the BMP requires two chars. So if you want to do any string processing which handles all Unicode characters properly you have to process the string, check for surrogates, then pass everything around as ints.

    I ended up with this:

    Code:
            static IEnumerable<int> EnumerateChars(string s)
            {
                for (int i = 0; i < s.Length; i++)
                {
                    if (char.IsSurrogate(s,i))
                    {
                        yield return char.ConvertToUtf32(s, i);
                        i++;
                    }
                    else
                    {
                        yield return s[i];
                    }
                }
            }
    Of course, it fails if the string is invalid UTF-16.

  21. Post
    #71
    A couple more tenant profiles added. Brings it to 8 out of 12 so far.

  22. Post
    #72
    Ah, I love architecture archaeology...

    The kneebone is connected to the HTTPStreamingIndexHandlerBase is connected to the HTTPNetStream which is returned in HTTPStreamingNetLoader.createNetStream...

    Still, at least it's consistent and coherent. Currently implementing pseuodstreaming in OSMF, FlowPlayer's feeling hacky as hell, RTMP plugin refuses to stream FLV, etc.

  23. Post
    #73
    Gave a presentation for which I had to be in at uni at an ungodly hour (9.20 ) .. still, went well, and now I can have a long weekend cos I'm a bum.

    The real work starts back up next week.

  24. Post
    #74
    If any of you remember my HDStream project, I've started coding the new backend services. It's going well so far, I'm writing it in C# .NET 3.5 as a console app and will run on Windows and Linux with Mono. It's a suite of in-built 'servers' that are the key services of the project that can be or should be distributed and scalable in this fashion. They all talk to a central master tracker server and can all be on the same machine if desired. They talk to each other via UDP. When time comes to add more storage or add more video encoding power, I just put the software on another server and enter the tracker details, enable the encoding module and it will be part of the eco-system, likewise for storage/mirror. It's been fun.

    The screenshot is running the master tracker on my local pc and all services on local pc via windows, 192.168.1.201 is a local CentOS VMWare machine running all services via mono using my windows server as the tracker and the 210.48.xx ip is my work machine using my local pc with the tracker port forwarded via the internet.

    Attached Images


  25. Post
    #75
    Now I have my master tracker issuing remote encode commands and the encoding server successfully spawning and utilising the StandardOutput of ffmpeg to create progress callbacks to the master tracker server. This is going to be handy. I have created a benchmark video with half animated video and half life-action (animated encodes faster) along with a speedtest feature to benchmark the capabilties of any particular encoding server in the cluster. Every progress update will be logged for further data analysis.

    Attached Images