ArchiveTeam Warrior on Hyper-V

I’ve gotten the ArchiveTeam Warrior working rather nicely on Windows 8 Professional’s built in Hyper-V Virtual machine. You can get it here or I’ve mirrored it on Archive.org here

This was driven by the desire to not have to install VM software when Windows 8 Professional has a perfectly functional VM built right in as a first class citizen. Continue reading

Moronix: Why AMD won’t take Michael Larabel seriously, and you shouldn’t either

This was written as a response to a post I made on the Phoronix forum relating to an article which consists entirely of Michael complaining about a lack of a review sample. I post there only because I keep hoping if I say it enough Michael will eventually get his act together. Its a stupid thought sure but I figure its worth a shot. I plan to update this at a future date to add links to examples but I can’t be bothered currently. This is all my opinion so if you don’t like it so be it.

If Michael wants a review sample he should start acting like a responsible journalist and news site. Did he ever think the reason AMD doesn’t send him review samples could be related to his reputation and how he manages his site? Perhaps they realize associating with his site actually makes them look worse due to how little he bothers with his reviews and articles.

Sure AMD gets publicity from his reviews posts, but they would get a lot more if he actually bothered to even scratch the surface of why or how things preform the way they do in comparison to other cards. Instead he only posts graphs and single sentence blurbs because its quick and easy. If four out of five articles posted here weren’t just link bate or link spam to past articles Michael may actually gain him some credibility. Articles that do nothing but link to other articles or sites controlled by Michael, which of course are filled to the brim with ads, do nothing to make him look good.

What’s even more sad is he takes the time to write these articles complaining rather than one articles with actual content that is useful to Linux users. Saying AMD doesn’t care about you because they didn’t send me free stuff, isn’t how you gain readers and people who will pay for your content, writing well thought out and researched posts and articles will. We all know how much he loves automation, how about he adds a feature to PTS to fill out bug reports for him. If every article about a regression or bug included a link to the bug report for it, where possible, he’d actually be doing something good for the community rather than just leaching off of them for ad revenue. There is only so many people who do the whole Linux news thing so if AMD or Nvidia were to send out review units there isn’t large selection of site, but when it comes to enthusiast computing Linux users are in the single digits percentage wise, there is no surprise then don’t bother there.

The Linux market AMD and Nvidia Care about is enterprise, people who buy Quadro, Tesla, and FireGL/FirePro cards. Consumer desktop Linux support is just a nice side effect of supporting Compute workloads and software like Siemens NX, Real3d, Maya and others. Gaming and fancy Compiz effects are nice side effects but they aren’t where the money is to be made on linux. That said, SteamOS may change that but until then I don’t expect AMD to spend a cent to support Linux desktop users, I’m glad they do and will continue to buy their hardware because they do, but I know its a drop in the bucket when it comes to their revenue.

Now I’m glad there is at least somewhere that takes the time to post these benchmarks and test the hardware but I would be an idiot to not demand more. Michael does nothing but half ass his articles and then complain when he isn’t taken seriously. If he wants to be taken seriously by the Linux community and companies like AMD and Nvidia he needs actually put some effort into his articles and take pride in them, not just post 5 links to past articles and his anzwix crap. The ratio of links to sentences approaches one in far to many of the articles here and it makes it damn clear all Michael cares about is page rank and ad revenue, not the content of the post itself. Compare his posts to articles posted on LWN or TomsHardware and see how many you can find that have more 7 out of 8 sentences linking to other articles on the site.

NOTE: This was written mostly as a rant and may be updated, deleted or modified if I feel like it.

Supybot WikiSearch plugin

TL;DR New Supybot WikiSearch plugin found here.

So a while back I wanted to add a simple Wiki Search to my Supybot IRC bot(well actually gribble based) so that users could easily search for project documentation on our wiki. I knew there were multiple options here. The first would be to just use an aliased search using Google site:insertwikihere.com but that couldn’t take advantage of the sites own search function and depends on Google’s indexing of the site. As the wiki is constantly being updated and has many technical subject that are not always easily googled I wasn’t sure I liked that idea.

The other option I had thought of was to find an existing Wikipedia search plugin and modify that for use with our wiki. This lead me to quantumlemur‘s github and his Wikipedia plugin. At first glance I could see where to modify the Wikiurl and that it would work for the task fairly well. I set to work modifying it and got it into a working state. I modified the search parts to point to our wiki’s url and removed the part where it pastes a snippit from the page so that it just returned the page title and url. This was to prevent the bot from being too spammy and prevent searches from hiding real channel conversation.

And it didn’t work. It turns out that his plugin actually parses the HTML of the search page and doesn’t use the nice MediaWiki API. At the time I was lazy and modified it to work with the theme of our wiki and the mediawiki version used and all was good… ish.

Then recently we wanted to add a search for another projects wiki so I had to go in and modify things again, or I could just rewrite it from scratch. And that is what I did.

The end result was 50 lines of code known as the Supybot WikiSearch plugin. If I get unlazy and ambitious I’ll add some way to have one plugin manage searches for multiple wiki’s and have it auto-detect the api.php path from just the wiki’s home page. I also plan on allowing it to display the snippits as quantumlemur‘s version does though it will stay disabled by default.

Edit: Further work on this has lead me to take back my comment about the MediaWiki being “nice” Oh god is it a pain in the arse and totally in consistent on what it returns wiki to wiki installation. Whoever designed that needs a good slap in the face. Perhaps one day I’ll do a full rant on that…

Casio Prizm GCC SDK v0.3

So after a long overdue major overhaul of the makefiles and some additions to libfxcg I am ready to release PrizmSDK v0.3. The biggest improvement is to the makefiles and build system and it puts me one step closer to a much cleaner setup. Most of the changes are based on the makefiles used by DevkitPro and I plan to move to an installer based system similar to theirs in the future as well.

The most obvious changes visible to the user is the new project directory layout. It is now much cleaner and uses a convenient project layout with specific folders for source files and temp object and such. All the configuration is still in the Makefile in the project directory, which has been cleaned up and commented. All of the dirty work was moved to a separate Makefile in the common folder of the SDK and the project directory shall now house only projects. The end goal is to allow the project directory to lie anyway and have environment variables for where the SDK lies. In the process I also made the makefiles much more cross platform safe though it might still need more tweaking to get just right.

I still wish to clean up the headers quite a bit more and move to just one header for all the syscalls and have that header include the extra headers with the core typedefs. There is no reason to have the headers split up as they are and I intend to clean that up with the next SDK update, if I don’t sneak it into this one. (if you notice filemodtimes change, this would be why)

The next big step will be to include an installer which can optionally grab msys instead of me having to include chunks of msys and cygwin for make and friends to function properly for people without them installed.

Please respond with any bugs or issues and I will address them as soon as possible. With how long it took me to do this release I’m sure I forgot something or made a typo somewhere and the only way I can fix it is if I know about it.

Download links:
xz: http://jonimoose.net/calcstuff/prizm/PrizmSDK-0.3.tar.xz (11M)
gz: http://jonimoose.net/calcstuff/prizm/PrizmSDK-0.3.tar.gz (42M)
zip: http://jonimoose.net/calcstuff/prizm/PrizmSDK-0.3.zip (44M)

Stuff and Things and Archii

F1rst P0st!!!1!!!!1!

Ok now that that is over with, I have officially installed wordpress on my site… now what. Oh yeah remember to use it.

On that note, I spent the night working on updating Archii but I have found a bit of an issue with a few things. One of which is that the latest udev does not seem to like linux 2.6.32.y on PPC. Right now it “works” but it spews errors to the console until minilogd OOM’s. I’m not sure what my options are here since the Archlinux PPC repo’s are going to continue to update no matter what and so I’ll either have to find a way to rebase the Wii patches or tell everyone to hold on to the older udev package. The issue being that newer initscripts will rely on a newer udev so that would only be a temporary solution. I know the zen kernel had the Wii patches merged at one point so it may be possible to use that but then you run into the issue of it not being a stock kernel which is working against the Arch way. I can upload newer rootfs tars if anyone is interested but I would rather wait until I see if I can work this out.

In other kernel related news its seems the Farter, author of GeeXboX for Wii, has made some patches to the gcnfb driver to make the gpu do the color space conversion. This is a big deal and will hopefully mean a speed boost at the cost of some ram. It would also mean that you can use the stock fbdev Xorg driver in place of xf86-video-cube. You can find more info at his blog Farter’s Mess.