Categories: Top ::
About
Codejunkie
Monologues of a mobile retro coder.
skeezix[at]codejedi.com
www.codejedi.com
Subscribe
Subscribe to a syndicated RSS feed. I've
also made a Livejournal version and Ben whipped up an auto-RSS Livejournal
Blogs
Michael Mace
JoelOnSoftware
Bruce Schneier
Wil Wheaton
I, Cringely
WritingOnYourPalm
Dan Gillmor
GrandTextAuto
Freedom to Tinker
Mark's SysInternals Blog
A List Apart
Tam's Palm
Bytecellar retro goodness
DadHacker; epic rants.
Lost Garden
Bill Ing
Ben Combee
PocketGoddess
PocketFactory
ModApex
Random Links
PalmInfoCenter
Zodiac Gamer
GP32x
Little Green Desktop
Atari Age
Penny Arcade
Hack-a-Day
Retro Remakes
SHMUPS!
Podcasts
1SRC
RetroGamingRadio
Recent Entries
| September 2008 | ||||||
|---|---|---|---|---|---|---|
| Sun | Mon | Tue | Wed | Thu | Fri | Sat |
| 1 | 2 | 3 | 4 | 5 | 6 | |
| 7 | 8 | 9 | 10 | 11 | 12 | 13 |
| 14 | 15 | 16 | 17 | 18 | 19 | 20 |
| 21 | 22 | 23 | 24 | 25 | 26 | 27 |
| 28 | 29 | 30 | ||||
Archives
This is a vent. A rant. A "why the gorramned hell do I put up with this shit?" Time is tight. I have a little baby upstairs who eats up every ounce of time, so I cut into sleep to get anything done... and yet, in this decade, developers are being slapped in the face at every instant. Don't get me started about Palm right now, but let me focus on Pocket PC right now.
Alright, so my latest app works on about 6 platforms pretty much flawlessly.. but on Pocket PC its a little spotty. Let me check things off..
GAPI - A very low level API for essentially making games; ie: It grabs the screen and inputs for you, so that when someone pushes a physical device button or taps on the screen, your application receives the event instead of any other application. Player hits a fire button, you get a button press. Great! On my particular device, the stylus is in fact all captured by my application -- tap in the top where the Start bar normally would be and my app catches it. Great. An alpha tester tells me this isn't the case for them.. they can pop right out of the game by tapping in the wrong area and the OS catches it. I have a similar problem though -- some of the buttons on the device go to my application.. but some shoot right out to other apps in the OS, causing application switches when pressed. The entire point of GAPI for us developers is so shit like this doesn't happen. And its written by device vendors and included in the device.. and apparently is totally unreliable. Thanks. Perfect. PEACHY.
Memory allocation - OKay, on a Pocket PC device (2003SE say), you have storage versus program (runtime) memory sharing one pool; you play with a slider to adjust what you want -- more data storage, or more runtime memory. So my game here needs a lot of memory (for now, until tweaked more), so I slide the widget over and give myself some RAM. Exit the memory application and watch as the RAM automatically adjusts back where it was.. or worse, some other random place. OKay, so the OS decided it needed to update some file somewhere and that needed space so it readjusts the slider. Well, why the gorramned hell provide a slider if it randomly moves? I need to set it, and leave it the hell alone. I'm mad because I can set it, drpo the device into the cradle for debugging and that event triggers a memory re-slide and voosh, I lose half my RAM. Unusable.
Environment - Of course, all this depends on the application running, the debugger connecting.. vanilla things like that. Basics. Well, when you drop the device into the cradle it immediately runs at about 10% of its speed while it presumably jabbers with the machine its plugged into.. peachy but fine, I can deal with it, though my application requires speed and reacts differently when there isn't any. Fine. More to point is that half the time when you launch the debugger, the handheld crashes.. or Active Sync crashes, or the environment (eVC 4 say) hangs. It even tells you 'the device crashes, you need to exit and restart eVC now' -- holy shit, you guys cannot handle when the remote dies? More to point.. out of about 8 tries (each taking a few minutes, due to all the slowdowns and hiccups) only about 1 o2 2 will succeed.. the handheld will just go whitescreen, or RAM adjust will zorch out the ability for it to run at the last second, or the debugger connection will crash....
I started 4 hours ago trying to actually debug something. A trivial task in any modern environment, but anyone whose done it knows handheld development is on the other of 10 years or so behind (really, I mean that.) After 4 hours, I've not gotten _anywhere_ -- I've debugged RAM problems the OS causes, and debugged USB oddities to try and keep the connection up more, and I've fought with the handheld OS screwing with me over and over about its RAM allocations, and fought with GAPI letting events slip through, and with the memory allocator returning memory when its actually run out, and all sorts of Voodoo.
The only question now is -- with what implement will I smash this device?
I'm seriously going to reconsider this application; my time is so very tight.. whose isn't? Who can waste 4-5 hours just fucking around with shit that has nothing to do with what you actually meant to do?
[ Category: / technology / codejedi ] [link] [Comments]>