| --- Log opened Mon Aug 27 00:00:30 2012 |
| 00:12 | | Orthia [orthianz@3CF3A5.E1CD01.C6689C.33956A] has quit [Ping timeout: 121 seconds] |
| 00:14 | | himi-cat [fow035@Nightstar-5d05bada.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 00:17 | | PinkFreud [WhyNot@NetworkAdministrator.Nightstar.Net] has joined #code |
| 00:17 | | mode/#code [+o PinkFreud] by ChanServ |
| 00:50 | | Orthia [orthianz@3CF3A5.E1CD01.C6689C.33956A] has joined #code |
| 01:28 | <&McMartin> | gnolam: Excellent. I can't catch it cheating. |
| 01:28 | <&McMartin> | (You can do similar tricks in RL with tracing mobius strips) |
| 01:39 | | himi-cat [fow035@D741F1.243F35.CADC30.81D435] has joined #code |
| 02:15 | | * McMartin gets enough of Swing-Hex-Inverter working that he thinks shifting over to targeting Android is now sensible. |
| 02:22 | <&McMartin> | Should step 1 be the design document, or the UI mockup~ |
| 02:41 | <&McMartin> | Answer: The design document, which includes UI mockups. |
| 02:41 | | Kindamoody[zZz] is now known as Kindamoody |
| 02:54 | | Tarinaky [tarinaky@Nightstar-dd7e4a05.net] has joined #code |
| 03:39 | <&McMartin> | Ha |
| 03:39 | <&McMartin> | [inform] vaporware says, "But I have more severity levels, so that's now a notice instead of a warning. Also, in what may be a first for compilers, one of the levels is 'protip'." |
| 03:39 | <&McMartin> | [inform] zarf says, "boo hiss" |
| 03:39 | <&McMartin> | [inform] vaporware | warning RLR4502: '=' used in a condition context, did you mean '=='? |
| 03:39 | <&McMartin> | [inform] vaporware | protip RLR4503: explicitly compare the result to 0 if that's what you want to do |
| 03:44 | | PinkFreud [WhyNot@NetworkAdministrator.Nightstar.Net] has quit [Ping timeout: 121 seconds] |
| 04:12 | | PinkFreud [WhyNot@NetworkAdministrator.Nightstar.Net] has joined #code |
| 04:12 | | mode/#code [+o PinkFreud] by ChanServ |
| 04:56 | | iospace is now known as iospacedout |
| 06:01 | | celticminstrel [celticminst@Nightstar-05d23b97.cable.rogers.com] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 06:28 | | Kindamoody is now known as Kindamoody|out |
| 06:38 | | Derakon is now known as Derakon[AFK] |
| 07:31 | < froztbyte> | nice |
| 08:22 | | Attilla_ [Obsolete@Nightstar-8e1acbff.as43234.net] has quit [Ping timeout: 121 seconds] |
| 08:52 | | himi-cat [fow035@D741F1.243F35.CADC30.81D435] has quit [Ping timeout: 121 seconds] |
| 10:23 | <&McMartin> | Hooray |
| 10:23 | <&McMartin> | https://hkn.eecs.berkeley.edu/~mcmartin/HexDesign.html |
| 10:23 | <&McMartin> | Anyone who's actually messed with Android - if I'm missing something obvious here do let me know =D |
| 10:26 | | Kindamoody|out is now known as Kindamoody |
| 10:28 | | You're now known as TheWatcher |
| 10:30 | | * jerith takes a look. |
| 10:31 | <&jerith> | McMartin: "vversion" in the summary. :-P |
| 10:35 | <&jerith> | McMartin: Saving board state every time a move is made is probably easier than explicitly handling it in pause(). |
| 10:35 | <&jerith> | This also lets you recover from crashes nicely. |
| 10:36 | < Moltare> | I note in the second part of the Activities section that you can tell an app in its manifest to fix itself to a specific orientation |
| 10:37 | < Moltare> | which has the handy side-effect of not pausing and resuming the app whenever your handset changes orientation |
| 10:37 | <&jerith> | I prefer autoswitching orientation, but with the option to lock a specific orientation in the settings. |
| 10:37 | <&McMartin> | jerith: Can you do that? |
| 10:38 | <&McMartin> | (Save every time a move is made) |
| 10:38 | <&McMartin> | The thing is, if I want to save the AI's search state, that clearly needs to happen in, well, stop(), I guess. |
| 10:38 | <&McMartin> | Assuming stop() even exists in API level 7, which I'm not totally sure it does. |
| 10:38 | <&jerith> | McMartin: Sure. Just put the save code in the ends of make_move() or whatever instead of pause(). |
| 10:39 | <&jerith> | Bear in mind that my experience with this stuff is strictly pre-Gingerbread, but you want to support that anyway. |
| 10:39 | <&McMartin> | OK, that sounds like a good plan, then. |
| 10:39 | <&McMartin> | Yeah, Gingerbread is in fact the level I intend to support in the first place, since it's what my personal phone runs. =P |
| 10:39 | <&jerith> | There may be performance implications if you're storing a lot of state, but HI's state seems to be small. |
| 10:39 | <&McMartin> | It's 121 integers. |
| 10:40 | < Moltare> | Oh god, I've been working on Runescape too long |
| 10:40 | < Moltare> | My initial reaction was "Fuck me, that's /huge/" |
| 10:41 | <&jerith> | I'm reasonably sure that stop() or whatever isn't guaranteed to run if the app is killed while paused. Or something. |
| 10:41 | <&McMartin> | (I'm targeting Froyo programmatically both because there's still nontrivial numbers of Froyo devices out there, and because nothing G-specific does anything I care about |
| 10:41 | < Moltare> | Yeah, you can guarantee that pause() runs but not stop() |
| 10:42 | < Moltare> | You can tell if a pause() is due to a quit request with isFinishing |
| 10:42 | <&jerith> | Moltare: |
| 10:43 | <&jerith> | Moltare: You can't tell if there's going to be a quit-without-stop() later, though. |
| 10:43 | <&McMartin> | OK, I don't see it here now |
| 10:43 | < Moltare> | True |
| 10:43 | <&McMartin> | But I swear I ran into something in the tutorials that said that starting at a certain API level you got a guarantee of onStop(). |
| 10:43 | <&McMartin> | That level was, however, past Gingerbread. |
| 10:43 | <&jerith> | Yeah. |
| 10:44 | <&jerith> | If you have any network connections, you probably want those to happen in a Service rather than an Activity. |
| 10:45 | <&McMartin> | Netplay is not in the cards, at least not yet. |
| 10:45 | <&McMartin> | I'm still not totally sure how workerthreads fit into this. |
| 10:45 | <&jerith> | It would be Good Practice to put all the gameplay stuff in a Service and only do UI in the Activity, except this has nontrivial performance implications. |
| 10:45 | <&jerith> | Lots of packing and unpacking in the IPC. |
| 10:46 | <&jerith> | I don't recall any Android-specific worker thread stuff. |
| 10:47 | <&McMartin> | "Do I kill them in pause() or in stop()" |
| 10:47 | <&jerith> | So you get all the power and flexibility of Java threads. |
| 10:47 | < Moltare> | I tried to put an ntp request in my UI thread when I first tried Android |
| 10:47 | < Moltare> | that was a poor move |
| 10:47 | <&McMartin> | Yeah |
| 10:47 | <&jerith> | And all the gnashing of teeth that implies. |
| 10:48 | <&jerith> | Also, avoid allocating memory at runtime if possible. |
| 10:48 | <&jerith> | GC runs can produce noticeable stutter. |
| 10:49 | <&McMartin> | Mmm |
| 10:49 | <&McMartin> | That's for a refinement pass |
| 10:51 | <&McMartin> | A number of bits of the AI code use ArrayLists. |
| 10:54 | <&jerith> | It shouldn't /really/ be a problem here. |
| 10:55 | <&jerith> | But if you know how big they're likely to grow you can statically allocate one and reuse it or something. |
| 10:55 | <&jerith> | Definitely at optimisation-time, though. |
| 10:56 | <&jerith> | Write clean code, then mess it up in the interests of performance. |
| 10:56 | <&McMartin> | This is particularly the case given that the program only does real-time operations while animating moves. |
| 10:56 | <&jerith> | Yeah, but GC can kick in whenever it likes. |
| 10:57 | <&McMartin> | Right, but this isn't Canabalt, so one move looking jerky isn't the kind of issue it'd be for an action game. |
| 10:58 | <&McMartin> | Oh, actually |
| 10:58 | <&McMartin> | " Allocate objects during initialization, or between animations. Never make an allocation while an animation is running." |
| 10:58 | <&McMartin> | Done and done |
| 11:00 | <&McMartin> | Anyway, thanks. |
| 11:07 | <&jerith> | :-) |
| 11:11 | | Attilla [Obsolete@Nightstar-8e1acbff.as43234.net] has joined #code |
| 11:14 | | Attilla [Obsolete@Nightstar-8e1acbff.as43234.net] has quit [[NS] Quit: ] |
| 11:15 | | Attilla [Obsolete@Nightstar-8e1acbff.as43234.net] has joined #code |
| 11:27 | | Kindamoody is now known as Kindamoody|out |
| 12:01 | | himi-cat [fow035@Nightstar-5d05bada.internode.on.net] has joined #code |
| 12:41 | | iospacedout is now known as iospace |
| 13:47 | | himi-cat [fow035@Nightstar-5d05bada.internode.on.net] has quit [Connection closed] |
| 13:48 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has joined #code |
| 13:48 | | mode/#code [+o himi] by ChanServ |
| 14:59 | | celticminstrel [celticminst@Nightstar-05d23b97.cable.rogers.com] has joined #code |
| 16:25 | | Kindamoody|out is now known as Kindamoody |
| 17:46 | | Vasi is now known as rms |
| 18:54 | | jeroid [jerith@687AAB.1BBF0C.0D2839.6E6D48] has joined #code |
| 18:58 | | * McMartin works out he actually needs 123 ints and three booleans. Probably four to make the loading logic simpler. |
| 19:03 | <~Vornicus> | 123 differently named ints? |
| 19:04 | < AnnoDomini> | Why not 127 ints, all named using one letter only? :P |
| 19:05 | | * Vornicus , um, patpats anno |
| 19:05 | <&McMartin> | Vornicus: No, there's an 11x11 array and then two enumerations and then two booleans per player, of which one is always provably irrelevant |
| 19:20 | | Kindamoody is now known as Kindamoody[zZz] |
| 19:21 | | Attilla [Obsolete@Nightstar-8e1acbff.as43234.net] has quit [Ping timeout: 121 seconds] |
| 19:22 | | Attilla [Obsolete@Nightstar-8e1acbff.as43234.net] has joined #code |
| 20:23 | <&ToxicFrog> | eeeeeeeeeeeee |
| 20:23 | <@Tamber> | aaaaaaaaaaaaa? |
| 20:24 | <~Vornicus> | ooooooooooooooooo |
| 20:25 | <&ToxicFrog> | Just got a phone call from my recruiter at Google |
| 20:25 | <~Vornicus> | iiiiiiiiiiiiiiii |
| 20:25 | <&ToxicFrog> | Apologizing for the fact that it might be as long as two weeks before I hear anything back |
| 20:25 | <&ToxicFrog> | I replied that from what my friends at Google had said I was expecting to wait a month or more, so this is actually great news |
| 20:26 | < jeroid> | TF: Do you want to work for Google? |
| 20:26 | <&ToxicFrog> | She replied "I like to stay on top of the recruiting process, especially for recruitees who are doing well" |
| 20:26 | <&ToxicFrog> | jeroid: yes. |
| 20:26 | < jeroid> | Ah, cool. |
| 20:26 | < jeroid> | Which office? |
| 20:27 | < jeroid> | Also, dev or SRE? |
| 20:27 | <&ToxicFrog> | CA-WAT, the Kitchener/Waterloo office (ie, the one right next to me), and dev. |
| 20:30 | < jeroid> | Ah, nice. |
| 20:31 | < jeroid> | I didn't know they had an office in Canadia. |
| 20:31 | <&ToxicFrog> | They have several, actually, but most of them are sales/marketing |
| 20:32 | < jeroid> | They have sales offices in ZA. |
| 20:33 | < jeroid> | But no tech. |
| 20:35 | <&ToxicFrog> | Toronto is sales, Montreal is sales and development, Kitchener is pure development, and I have no idea what Ottawa does. |
| 20:39 | | jeroid [jerith@687AAB.1BBF0C.0D2839.6E6D48] has quit [[NS] Quit: Bye] |
| 22:21 | | RichyB [richardb@Nightstar-428e5766.bb.sky.com] has joined #code |
| 22:21 | | RichyB [richardb@Nightstar-428e5766.bb.sky.com] has quit [Client closed the connection] |
| 22:21 | | RichyB [richardb@Nightstar-428e5766.bb.sky.com] has joined #code |
| 23:56 | | Adolf_Hitler [ForEverAuti@4CA975.06D306.F95019.3FD4B0] has joined #code |
| 23:56 | < Adolf_Hitler> | :) |
| 23:56 | <@Tamber> | o.? |
| --- Log closed Tue Aug 28 00:00:45 2012 |