| --- Log opened Tue Jan 11 00:00:46 2011 |
| 00:33 | | Tarinaky [Tarinaky@Nightstar-f349ca6d.plus.com] has quit [Client closed the connection] |
| 00:41 | | Attilla [Some.Dude@Nightstar-87c66023.threembb.co.uk] has quit [[NS] Quit: ] |
| 00:41 | | celmin|supper is now known as celticminstrel |
| 00:45 | | gnolam [lenin@Nightstar-38637aa0.priv.bahnhof.se] has quit [[NS] Quit: Z?] |
| 00:51 | | Kindamoody is now known as Kindamoody[zZz] |
| 02:31 | | AnnoDomini [annodomini@Nightstar-10f8b8dc.adsl.tpnet.pl] has quit [[NS] Quit: Enough of this. I need sleep.] |
| 04:35 | | Thaqui [Thaqui@27B34E.D54D49.F53FA1.6A113C] has joined #code |
| 05:16 | | celticminstrel [celticminst@Nightstar-f8b608eb.cable.rogers.com] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 05:30 | | cpux is now known as shade_of_cpux |
| 05:43 | | Rhamphoryncus [rhamph@Nightstar-473f8685.abhsia.telus.net] has quit [Client exited] |
| 06:09 | | Ior [NSwebIRC@Nightstar-4764665d.tbcn.telia.com] has joined #code |
| 07:14 | | You're now known as TheWatcher |
| 08:05 | | Serah [Stalker@3A600C.A966FF.5BF32D.8E7ABA] has joined #code |
| 09:00 | | Serah [Stalker@3A600C.A966FF.5BF32D.8E7ABA] has quit [Ping timeout: 121 seconds] |
| 09:14 | | Kazriko [kaz@Nightstar-e09690fa.client.bresnan.net] has quit [Connection closed] |
| 09:17 | | Kazriko [kaz@Nightstar-e09690fa.client.bresnan.net] has joined #code |
| 09:17 | | mode/#code [+o Kazriko] by Reiver |
| 09:21 | | NSJavaGuest-6392 [NSjavaIRC@Nightstar-283cd59d.clmboh.sbcglobal.net] has joined #code |
| 09:21 | | NSJavaGuest-6392 is now known as Janus |
| 09:22 | | * Janus waits the required minute before launching into an inspired monologue |
| 09:23 | < Vornicus> | There's a required minute? |
| 09:23 | < Janus> | I don't wanna interrupt nothing! |
| 09:23 | | ServerMode/#code [+o Vornicus] by *.Nightstar.Net |
| 09:23 | < froztbyte> | it's quiet, you may proceed |
| 09:24 | | Vornicus changed the topic of #code to: Welcome to #Code! || Rants and monologues are encouraged; many cores, no waiting || Pastebin: http://pastebin.starforge.co.uk/ (Note antispam question, answer 'Yes') || ? x, ich werde x Wissenschaft tun |
| 09:25 | < Janus> | Okay! Then off I go! ... I had to take a moment to remember what I was so ready to rant about |
| 09:27 | <@Vornicus> | *snrk* |
| 09:27 | | Vornicus is now known as Vornicus-Latens |
| 09:27 | <@Vornicus-Latens> | I sleep! You rant! Later, I read! |
| 09:28 | < Janus> | Why is coding things that actually do things, so hard? Like making physics stuff and parsers and managers and all sorts of little things. But when it's time to actually use those things, I always find an excuse to rewrite something. And then the alligator sa-- SLEEP |
| 09:28 | < Janus> | *physics and parsers and managers being among the things that don't do anything in this case |
| 09:29 | <@Vornicus-Latens> | No plan survives contact with the enemy. |
| 09:29 | <@Vornicus-Latens> | And the real world, let me tell you, is the most manxome foe of all. |
| 09:31 | < Janus> | I hate going against it, it always plays dirty |
| 09:35 | | * Janus is rewriting the silly physics thing a 6th time. And it worked perfectly fine beforehand (sort of). |
| 09:35 | < Janus> | Set out to make it easy for myself to use this time. Which I am quickly thwarting as I go along. |
| 09:39 | < Janus> | Starting to wonder if I should just use someone else's library, so I can just be done with it, aha |
| 09:41 | < Janus> | Don't get me wrong, I actually enjoy it. Well, at least the non-math-y bits anyway. Sadly anything at polynomials and beyond, and you lose me. |
| 09:41 | < Janus> | *writing the physics stuff that is |
| 09:42 | < Janus> | I wonder how important math is to a programmer anyway. It uses very little of it, I've noticed. Even here, the most complex thing I had to work with was a cross/dot product |
| 09:45 | < Janus> | ... okay getting back from that tangent, need to keep this monologue on track. I wonder if I am over-complicating this thing. There are indie games with entirely acceptable, bug free platformer physics, they make it look effortless enough. Like it's a really simple thing |
| 09:50 | < froztbyte> | how much complexity are you adding? |
| 09:52 | < Janus> | I'd think I've actually stripped it away, I think. There are static bodies which can move, but don't react to collisions, and dynamic bodies that do. It's only 2D rectangles. ... m-mostly 2D rectangles. The top part can be sloped. ... and any of the sides may be passable. And bodies can have callback functions assigned to them. |
| 09:57 | < Janus> | ... maybe that's the problem. A lot of the trouble has been making the fake shape of a sloped rectangle work |
| 10:15 | < Janus> | ... aha, suddenly changing it so that it's triangles and rectangles, fixed a lot of the falling through and sudden stop bugs. I will need to investizzz |
| 10:15 | | Janus [NSjavaIRC@Nightstar-283cd59d.clmboh.sbcglobal.net] has quit [[NS] Quit: Nightstar's Java Chat http://www.nightstar.net] |
| 10:22 | | Kindamoody[zZz] is now known as Kindamoody |
| 10:43 | | Kindamoody is now known as Kindamoody|out |
| 10:50 | | Ior [NSwebIRC@Nightstar-4764665d.tbcn.telia.com] has quit [Ping timeout: 121 seconds] |
| 11:37 | | Serah [Stalker@26ECB6.A4B64C.298B52.D80DA0] has joined #code |
| 12:14 | | celticminstrel [celticminst@Nightstar-f8b608eb.cable.rogers.com] has joined #code |
| 12:17 | | celticminstrel [celticminst@Nightstar-f8b608eb.cable.rogers.com] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 12:25 | | Thaqui [Thaqui@27B34E.D54D49.F53FA1.6A113C] has quit [Connection reset by peer] |
| 12:31 | | AnnoDomini [annodomini@Nightstar-91f3d793.adsl.tpnet.pl] has joined #code |
| 12:31 | | mode/#code [+o AnnoDomini] by Reiver |
| 12:33 | | gnolam [lenin@Nightstar-38637aa0.priv.bahnhof.se] has joined #code |
| 13:45 | | Kindamoody|out [Kindamoody@Nightstar-4764665d.tbcn.telia.com] has quit [[NS] Quit: This computer has gone to sleep] |
| 14:44 | | Attilla [Some.Dude@Nightstar-c23b61b3.threembb.co.uk] has joined #code |
| 14:45 | | mode/#code [+o Attilla] by Reiver |
| 16:28 | | Rhamphoryncus [rhamph@Nightstar-473f8685.abhsia.telus.net] has joined #code |
| 16:39 | | You're now known as TheWatcher[afk] |
| 17:50 | | Derakon [chriswei@Nightstar-cfae48c3.ca.comcast.net] has joined #code |
| 17:50 | | mode/#code [+o Derakon] by Reiver |
| 17:51 | <@Derakon> | Seems like there must be some better way to do this... |
| 17:52 | <@Derakon> | I have a set of child classes that all implement the same function, which returns a tuple of (program name, argument list). |
| 17:52 | <@Derakon> | The invoker of that function then combines those values with a list of files to iteratively invoke the program on each of them. |
| 17:52 | <@Derakon> | I've discovered now that one of the child classes has to provide an additional argument that is specific to each file (basically an additional output file). |
| 17:53 | <@Derakon> | Which means the tuple has to be extended, which in turn means either hacking the invoker to check the tuple length, or modifying every child class to insert a null value in their tuple. |
| 17:53 | <@Derakon> | Or I suppose I could add an additional function to the base class getFileSpecificArguments and override that for just the one child class... |
| 17:54 | <@Derakon> | Seems kinda a weird approach though. |
| 17:54 | <@Derakon> | Any thoughts? |
| 17:55 | < Rhamphoryncus> | What's the tuple look like? |
| 17:55 | <@Derakon> | (program name as a string, program arguments as a string) |
| 17:55 | < Rhamphoryncus> | Where's the output file specified? |
| 17:55 | <@Derakon> | I'd extend it to be (program name as a string, program arguments as a string, list of file suffix strings) |
| 17:56 | <@Derakon> | Currently (without the extra suffices), the invocation is % programName inputFile inputFile arguments |
| 17:56 | <@Derakon> | That is, the output overwrites the input. |
| 17:56 | <@Derakon> | For this specific case, I need something like %programName inputFile inputFile inputFile-plot arguments |
| 17:56 | < Rhamphoryncus> | hrm |
| 17:57 | < Rhamphoryncus> | a collections.namedtuple might help clean that up |
| 17:57 | < Rhamphoryncus> | Or let that class override the tuple-handling method |
| 17:57 | <@Derakon> | Or return a dict instead, maybe? |
| 17:58 | < Rhamphoryncus> | I'd prefer a namedtuple over a dict |
| 17:58 | < Rhamphoryncus> | Or a full class over a dict, which could have some other helper methods |
| 17:58 | <@Derakon> | Mm, yeah. |
| 17:58 | <@Derakon> | That'd work. |
| 17:59 | <@Derakon> | Of course, it'd involve changing all the other child classes too, but it'd also be nicely future-proofed. |
| 17:59 | <@Derakon> | When in doubt, make your data smarter~ |
| 17:59 | < Rhamphoryncus> | It's also possible to subclass a namedtuple, so you could name it Foo now, then later when you have extra methods you rename it to FooBase and add a new Foo |
| 18:00 | < Rhamphoryncus> | although by that point you might want to drop the tuple-ness |
| 18:00 | <@Derakon> | Yeah. |
| 18:09 | | Tarinaky [Tarinaky@Nightstar-f349ca6d.plus.com] has joined #code |
| 18:52 | | SmithKurosaki [smith@Nightstar-e94c2dff.dsl.teksavvy.com] has quit [Ping timeout: 121 seconds] |
| 19:00 | <@Derakon> | Hrm, one problem with the smart-container approach: I have to put all the details of invoking the program onto the commandline at some point, to pass them to a second Python program that actually puts them on the command line "for real". |
| 19:01 | <@Derakon> | That is, local Python program gets user input => connect to remote server => % ./invokeProcessor.py programName arguments files => invokeProcessor.py uses subprocess to run programName with the specified arguments for each file. |
| 19:01 | <@Derakon> | So basically I have to serialize the smart container in a commandline-friendly fashion. |
| 19:01 | <@Derakon> | Ick. |
| 19:02 | | Syloqs-AFH [Syloq@NetworkAdministrator.Nightstar.Net] has quit [Connection closed] |
| 19:11 | | Syloqs_AFH [Syloq@NetworkAdministrator.Nightstar.Net] has joined #code |
| 19:12 | | Syloqs_AFH is now known as Syloqs-AFH |
| 19:52 | < Rhamphoryncus> | Can you make the container smart enough to produce its end result at that point? |
| 19:53 | <@Derakon> | The problem is that its result has to be generated once for each file, but invokeProcessor.py iterates over each file. |
| 19:53 | <@Derakon> | So I have to basically give invokeProcessor.py an instance of the container, which means serializing it. |
| 21:10 | <@Derakon> | Any suggestions for serializing/deserializing an object without using eval, that fit the serialization onto one line? |
| 21:10 | < Rhamphoryncus> | json? :/ |
| 21:10 | < Rhamphoryncus> | I wouldn't end up doing a one-liner |
| 21:11 | < Rhamphoryncus> | I'd have a helper method |
| 21:11 | < Rhamphoryncus> | (or two. serialize and deserialize) |
| 21:11 | <@Derakon> | No, no. |
| 21:11 | <@Derakon> | I mean, the specification for this object is an argument to a script. |
| 21:11 | <@Derakon> | Unless I want to write the serialization to a file, transfer the file across to the remote server, and then pass the path to the file. |
| 21:12 | < Rhamphoryncus> | ugh no |
| 21:12 | < Rhamphoryncus> | Unfortunately I still can't picture all of what you're doing |
| 21:13 | <@Derakon> | Let me write up a pastebin outlining things... |
| 21:20 | <@Derakon> | http://derakon.dyndns.org/~chriswei/temp2/processor.txt |
| 21:20 | <@Derakon> | Sorry about the lack of newlines. |
| 21:22 | | Kindamoody [Kindamoody@Nightstar-4764665d.tbcn.telia.com] has joined #code |
| 21:23 | < Rhamphoryncus> | huh |
| 21:24 | <@Derakon> | It's the transition from the program that's running on the local machine to the program that's running on the cluster that's giving me trouble here. |
| 21:24 | < Rhamphoryncus> | You actually could upload a file to deserialize.. but I'm not sure I'd bother. Since it's small you could just as well use one blob as a single argument |
| 21:25 | <@Derakon> | So what do I use to deserialize, without relying on eval? Or do I just give up and assume that my users are not hackers? |
| 21:25 | < Rhamphoryncus> | I'd use json |
| 21:25 | <@Derakon> | Argh, being called away. >.< |
| 21:25 | <@Derakon> | This is really inconvenient timing. |
| 21:25 | <@Derakon> | (I think I'm using Python 2.5, BTW, so no json...) |
| 21:26 | < Rhamphoryncus> | I'd become python invokeProcessor.py blob |
| 21:26 | < Rhamphoryncus> | Available as a third-party install |
| 21:27 | < Rhamphoryncus> | I'm not sure it's that hard to do client-side though |
| 21:28 | < Rhamphoryncus> | It looks like invokeProcessor.py is somewhat like xargs. You have a list of files, running the target program for each, and you have a list of arguments for that program |
| 21:30 | < Rhamphoryncus> | xargs lets you use placeholders. You might do something like invokeProcessor.py file1 file2 file3 file4 file5 -- {input} {output1} {output2} arg1 arg2 |
| 21:55 | | Attilla [Some.Dude@Nightstar-c23b61b3.threembb.co.uk] has quit [Ping timeout: 121 seconds] |
| 22:04 | | Vornicus-Latens is now known as Vornicus |
| 22:19 | | Attilla [Some.Dude@Nightstar-516247f5.threembb.co.uk] has joined #code |
| 22:19 | | mode/#code [+o Attilla] by Reiver |
| 22:55 | | RichardBarrell [mycatverbs@Nightstar-689c9c54.cable.virginmedia.com] has joined #code |
| 23:13 | | celticminstrel [celticminst@Nightstar-f8b608eb.cable.rogers.com] has joined #code |
| 23:38 | | shade_of_cpux is now known as cpux |
| --- Log closed Wed Jan 12 00:00:47 2011 |