wolfwings

Borrowed pistolpup's suit again, or at least parts of it. Going partial-suiting at Zona Rosa (outdoor mall), walk around a'while until I'm too hot/tired to keep going, then head home. Partialing first to feel out the place for fullsuiting later, see how some of the stores react, and also since I lack a proper spotter I'd rather have something I can don/doff inside my car or a bathroom instead of needing a full changing room, Just In Case.

Does anyone local out there want to tag along? I'd still only partial, but it'd be more fun as a group excursion and to have someone else to take photo's/video's. I even have the camera you can use! =^.^=

Anyways... laundry now, then bed before I head out.

This gryph is BORED.

I want an excuse to get out of the house. Anyone out there in the Kansas City area (or maybe 200 miles radius of there) want some company? Or want to meet up and just hang out or something?

I'm serious here. I just want to get out of the apartment if I can this weekend, and do instead of sit and think. Anyone need their brake pads replaced? Gutters cleaned? Some non-tedius physical task would be great. I'd say let's go suiting... but then I keep remembering, wait, no suit of my own yet. *laughs* Happy to be a handler if someone wants to go suiting out in public or something.

Hrm... there's always the Macy's mall, that place is so many leap-years beyond dead seems ideal for something to go on there... *ponderponder*

Yeah... lazy-bird had the lazy trimmed off of him, wants to be active-bird now between the experiences at RCFM and it turning into summer finally. =^.^=

Gotta head in to work soonish... finally coming down off the con-high I've been riding for the last... gah, damn near half a week now. It's hard to have PCD when you're still floating, but it was definately there.

And now trying to decide... $1k to get into suiting Right Now or not... mostly at a toss-up because I'm not sure if I want something on the cute-side like Beef Jerky or not.

Thoughts?

...will fill in FAR more details later.

• Didn't suit anywhere near as much as I wanted to.
• But did suit FAR too much. Thanks, pistolpup, for the loan of the suit!
• Got bit by the bug so much I actually went back upstairs, unpacked enough of the suit to partial, hand-washed and dried the undersuit, and romped back downstairs for 4 extra hours instead of sleeping those 4 hours. =^.^=
• Got to see likeshine and painteddog briefly at least, which was good. But nowhere near enough.
• Pizza Feed went well, helped out a bit there too.
• Missed the Furp Toast on actual Memorial Day since I hadn't heard about it. =-.-=
• Abused my CB far too much on the lightning-fast drive home to get here in enough time to sleep at all before work, to avoid calling off from work.

I've officially had it with how wonky the SixAxis is. It's returning multi-byte values in big-endian format it looks like, no wonder they didn't reveal the accelerometer data over the USB interface... they can't, since USB specifies little-endian data order. (Basically, all numbers are byte-swapped. 1 becomes 256, 256 becomes 1, 257 stays the same, etc.)

Previously, I was trying to 'play nice' and just modify Sony's HID Report Descriptor and continue using the existing parsing infrastructure in the HID stack, as all the other devices in the kernel currently do. At this point, screw that, I'm re-writing the HID Report Descriptor to something ultra compact and very sane to read instead of the 148-byte monolith of standards-breakage it is now, and just use the HID Report pre-filter hook to re-write the raw bytestream.

Now... to decide on an axis-ordering... comments, anyone? I'm trying to decide if I should just shove things in like this or not since from what I've seen the R3 stick is the least-used outside of L3+R3 specific games, so this layout would make it easiest to use the PS3 controller as a 'generic gamepad' in most games:

+ pad becomes X/Y
L3 stick becomes Rx/Ry
R3 stick becomes Z/Rz

In effect I'm trying to match up the results to the stereotypical Gravis USB Gamepad layout as far as possible.

And oddly, while the Linux Input stack supports buttons that return analog values ...well, many games don't like those because they're highly unexpected. =^.^= Makes for 'sticky button' symptoms, as many games just check 'are you 0?' for is a button pressed or not, and a button can 'last report' a 1 or 2 instead of a 0. I'll crowbar the values to 0 when the 'button' sensor says 0, as it's a proper way to reveal the functionality and will work for all instances as far as I can tell.

And I have to say, all this HID-hacking really gives me a lot more respect for the robust control-binding interface the MAME developers built. I used to wonder why they supported such complexity, now I'm realizing it's actually very capable, since it can support far more advanced control interfaces easilly than the typical '16 switch' gamepads most of us grew up with.

Not of THAT much use except perhaps to illustrate a simple PThread system with rudimentary triple-buffering, since I'll be rolling this information into some work I'm doing modifying the existing 'Sony Quirks' HID Driver for the Linux kernel to use a similair reporting metric instead of the '28 axis and 19 buttons, many of which either duplicate information or are ALWAYS zero' morass the stock controller returns.

Mind you, many games don't like applying an analog button to a digital action, but I find that to be their problem, not mine. Now to figure out how to report these axes mappings using the HID spec...

Last note, this is also an attempt at trying to apply both Literate Programming and the Linux Kernel Specs (which are remarkably similair to my own, and I understand the reasons for the differences and think they're better so I'll adopt them) so if anyone wants to glance over the code, feel free. The 'test.c' file was quite a bit more slapdash than the rest that I actually tried to polish up as much as I could, and the Makefile is so short and simple as to be ignorable.

PThreaded SixAxis HIDRAW interface kit

...since I'm working on retrofitting a vector-based movement system to an old FASA game (Interceptor) I've realized a few traits I didn't realize before until I tested them via brute-force to verify things to myself.

When you sum things up in a vector-movement system as applied to a hex grid, there's two basic mergers that let you quickly simplify things down to a single pair of vectors that will always be on neighboring sides of the movement hex. I.E. A single fixed line on a 360-degree circle, common sense but neat on how things simplify down.

First, a hex grid to work with:

    __
 __/#W\__
/#A\__/#E\
\__/  \__/
/#Z\__/#D\
\__/#X\__/
   \__/

Yes, I'm being a bit lazy on the labelling, deal. Long story short:

• 1 #A + 1 #E = 1 #W.
• 1 #W + 1 #X = 0.

Yes, it's minor, but I felt like sharing. Oh, and interceptor critical-hit sheets if you want good-quality scans, and the live WIP for the by-hand re-building I'm making of the sheets into pure PostScript to be made into PDF's eventually.