| --- Log opened Wed Aug 17 00:00:18 2011 |
| --- Day changed Wed Aug 17 2011 |
| 00:00 | <@ToxicFrog> | Struct crushing? |
| 00:01 | <@Tamber> | What happens if you pull the lever linked to the supporting int. |
| 00:02 | <@Namegduf> | Or, alternatively, if you release the lever. |
| 00:02 | <@Namegduf> | Choose wisely. |
| 00:04 | | celticminstrel [celticminst@Nightstar-5d22ab1d.cable.rogers.com] has joined #code |
| 00:05 | <@ToxicFrog> | Hmm. |
| 00:05 | <@ToxicFrog> | I wonder if variable references would be a useful feature in vstruct. |
| 00:05 | <@ToxicFrog> | Eg, instead of: |
| 00:06 | | cpux_ [chatzilla@Nightstar-c5874a39.dyn.optonline.net] has joined #code |
| 00:06 | <@ToxicFrog> | local toc = struct.unpack(fd, "n:u4"); struct.unpack(fd, toc.n .. "* c2", toc) |
| 00:06 | <@ToxicFrog> | You could instead do: |
| 00:06 | | Netsplit *.net <-> *.split quits: @cpux, @simon_, @Attilla, @McMartin, @kw-sleep-n, @PinkFreud |
| 00:06 | <@ToxicFrog> | local toc = struct.unpack(fd, "n:u4 $n * c2") |
| 00:06 | | cpux_ is now known as cpux |
| 00:07 | <@TheWatcher> | I doubt they'll be too useful if they involve netsplits~ |
| 00:07 | <@Tamber> | Oh, sheesh. You said that about the atom-splits, too! |
| 00:07 | | Attilla [Some.Dude@Nightstar-f29f718d.cable.virginmedia.com] has joined #code |
| 00:08 | | simon_ [simon@Nightstar-883ecc1d.brahmaserver.dk] has joined #code |
| 00:08 | <@ToxicFrog> | Add temporaries and you can unpack an entire TXTDAT file in one call, using something like this: |
| 00:11 | <@ToxicFrog> | n:u2 magic:u2 x4 $1:$@ $magic * (+$n) $2:$@ $n * (@$1 $3:u4 $1:$@ @$2 +$3 z) |
| 00:12 | <@ToxicFrog> | However, at that point, I'm basically reinventing Perl. |
| 00:12 | <@ToxicFrog> | So a better approach may be to make the struct components first-class. |
| 00:13 | <@Tamber> | hehe |
| 00:16 | | kw-sleep-n [kwsn@Nightstar-635d16fc.org] has joined #code |
| 00:17 | | You're now known as TheWatcher[T-2] |
| 00:23 | | You're now known as TheWatcher[zZzZ] |
| 00:26 | | McMartin [mcmartin@Nightstar-656a00f7.pltn13.sbcglobal.net] has joined #code |
| 00:29 | | AnnoDomini [annodomini@Nightstar-6667002f.adsl.tpnet.pl] has quit [[NS] Quit: leaving] |
| 00:37 | | mode/#code [-o Syloqs_AFH] by ChanServ |
| 00:38 | | PinkFreud [WhyNot@NetworkAdministrator.Nightstar.Net] has joined #code |
| 00:57 | | kw-sleep-n is now known as kwsn |
| 01:06 | | Attilla [Some.Dude@Nightstar-f29f718d.cable.virginmedia.com] has quit [Ping timeout: 121 seconds] |
| 01:17 | | Omega [omegaboot@Nightstar-10752b3e.il.comcast.net] has joined #code |
| 01:18 | | Alek [omegaboot@Nightstar-10752b3e.il.comcast.net] has quit [NickServ (GHOST command used by Omega)] |
| 01:18 | | Omega is now known as Alek |
| 01:30 | | Vash [Vash@Nightstar-f03c5637.sd.cox.net] has joined #code |
| 01:30 | | Vash is now known as Vornicus |
| 01:37 | | Derakon[AFK] [Derakon@Nightstar-c826c088.ca.comcast.net] has quit [Ping timeout: 121 seconds] |
| 02:20 | | SmithKurosaki [smith@Nightstar-87f33f96.cpe.mountaincable.net] has quit [Operation timed out] |
| 02:21 | | gnolam [lenin@Nightstar-202a5047.priv.bahnhof.se] has quit [[NS] Quit: Z?] |
| 02:23 | | Thaqui [Thaqui@27B34E.D54D49.F53FA1.6A113C] has joined #code |
| 02:50 | | Kindamoody[zZz] is now known as Kindamoody |
| 03:04 | | Phox [Phox@Nightstar-3e5deec3.gv.shawcable.net] has quit [NickServ (GHOST command used by Neophox)] |
| 03:04 | | Phox [Phox@Nightstar-3e5deec3.gv.shawcable.net] has joined #code |
| 03:11 | <@ToxicFrog> | Euuuuuuuurgh |
| 03:12 | <@ToxicFrog> | The fundamental problem with this crash I'm seeing is that vstruct handles it very poorly when you have both a readahead and a buffer underrun. |
| 03:12 | <@ToxicFrog> | To fix it, I need to rewrite vstruct's error handling code so that it can distinguish between errors it raises itself (which should be passed through to the user unaltered) and errors raised by the runtime (which should be reported as a bug in the library) |
| 03:18 | | cpux is now known as shade_of_cpux |
| 03:18 | < Vornicus> | What's readahead in this case? |
| 03:19 | <@ToxicFrog> | vstruct will attempt to coalesce adjacent read operations where possible (ie, where it can determine ahead of time the size of the read) |
| 03:20 | <@ToxicFrog> | So, something like [[1024 * u4]] will result in one 4092-byte read and then 1024 substring operations, rather than calling fread() 1024 times. |
| 03:23 | <@ToxicFrog> | The problem is that the generated code for this looks similar to: |
| 03:23 | <@ToxicFrog> | buffer = fread(size) |
| 03:23 | <@ToxicFrog> | io(operation, buffer, ...) |
| 03:23 | <@ToxicFrog> | And if we're at EOF, fread(size) returns nil rather than "" |
| 03:25 | < Vornicus> | BUt if you're not at EOF but you just plain don't have data (reading off a slow pipe) you get a different problem. |
| 03:25 | < Vornicus> | Or in that case does it just block until you've got 4kB |
| 03:25 | < Vornicus> | ? |
| 03:26 | <@ToxicFrog> | It'll do what fread does normally, which is block. |
| 03:26 | < Vornicus> | Okay. |
| 03:26 | < Vornicus> | (I didn't know what fread does in that situation, I have never written any sort of robust network or piping code.) |
| 03:29 | < Vornicus> | Okay, so you've got two possible problems: 1. the buffer might be short; 2. the buffer might be nil. |
| 03:30 | < Vornicus> | So you're just going to have to check that. --wait, do your exceptions, when vstruct makes them, do you get to add in a string like you do in Python, "raise ValueError, 'divide by zero'"? |
| 03:32 | < Vornicus> | If so, you can add strings to yours that go "vstruct" in there. Idunno, I'm kind of meandering now. |
| 03:33 | <@ToxicFrog> | Different error semantics. There's no Error class, but you can throw anything (including a string or an object) |
| 03:34 | <@ToxicFrog> | So yeah, it's totally doable |
| 03:34 | <@ToxicFrog> | It's just effort |
| 03:37 | <@ToxicFrog> | (and I actually have way more possible problems than that) |
| 03:37 | < Vornicus> | Ah, so |
| 03:37 | <@ToxicFrog> | The underlying thing that needs fixing is that at present, all errors are treated as internal errors |
| 03:37 | <@ToxicFrog> | So assert(buffer, "buffer underrun in vstruct.unpack") will result in a giant === INTERNAL ERROR IN VSTRUCT LIBRARY, REPORT THIS AS A BUG === banner. |
| 03:37 | < Vornicus> | So you're assuming everything is a bu--- yes. How fun. |
| 03:46 | <@ToxicFrog> | to be fair, it is |
| 03:47 | <@ToxicFrog> | the bug is "error handlers are not present for generated code"~ |
| 03:47 | < Vornicus> | Heh |
| 04:07 | | Stalker [Z@Nightstar-3602cf5a.cust.comxnet.dk] has quit [Ping timeout: 121 seconds] |
| 04:23 | | Rhamphoryncus [rhamph@Nightstar-14eb6405.abhsia.telus.net] has joined #code |
| 04:59 | | Kindamoody [Kindamoody@Nightstar-5507a6b5.tbcn.telia.com] has quit [Ping timeout: 121 seconds] |
| 05:16 | | Kindamoody|out [Kindamoody@Nightstar-5507a6b5.tbcn.telia.com] has joined #code |
| 05:17 | | Kindamoody|out is now known as Kindamoody |
| 05:30 | | Vornicus [Vash@Nightstar-f03c5637.sd.cox.net] has quit [Connection closed] |
| 05:43 | | kwsn is now known as kw-sleep-n |
| 07:13 | | You're now known as TheWatcher |
| 07:19 | | celticminstrel [celticminst@Nightstar-5d22ab1d.cable.rogers.com] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 08:11 | | You're now known as TheWatcher[afk] |
| 08:11 | | Kindamoody is now known as Kindamoody|out |
| 08:30 | | ToxicFrog [ToxicFrog@ServerAdministrator.Nightstar.Net] has quit [Operation timed out] |
| 08:31 | | ToxicFrog [ToxicFrog@ServerAdministrator.Nightstar.Net] has joined #code |
| 09:27 | | AnnoDomini [annodomini@Nightstar-3edcbb87.adsl.tpnet.pl] has joined #code |
| 09:34 | | Kindamoody|out is now known as Kindamoody |
| 09:37 | | Attilla [Some.Dude@Nightstar-f29f718d.cable.virginmedia.com] has joined #code |
| 11:16 | | You're now known as TheWatcher |
| 12:54 | | gnolam [lenin@Nightstar-202a5047.priv.bahnhof.se] has joined #code |
| 13:07 | < gnolam> | Ah, readability. |
| 13:07 | < gnolam> | Naturally, "DBCN 17J 2." means "use step-specific straggling logic if available.". |
| 13:08 | <@Tamber> | /Naturally/ |
| 13:10 | <@TheWatcher> | Isn't that obvious? |
| 13:11 | | mode/#code [+ooooooooooooooo Attilla Phox simon_ kw-sleep-n PinkFreud AnnoDomini McMartin ToxicFrog Thaqui gnolam Syloqs_AFH Alek Rhamphoryncus Kindamoody shade_of_cpux] by Tamber |
| 13:16 | | RichardBarrell [mycatverbs@Nightstar-3b2c2db2.bethere.co.uk] has joined #code |
| 13:29 | <@Kindamoody> | Whoops, what did Tamber do?!? |
| 13:30 | <@Tamber> | :) |
| 13:36 | | SmithKurosaki [smith@Nightstar-87f33f96.cpe.mountaincable.net] has joined #code |
| 13:43 | <@TheWatcher> | Argh, godsdamnit, why is this tag filtering code not working properly?! |
| 13:43 | <@froztbyte> | wrong tags? |
| 13:44 | | Stalker [Z@Nightstar-3602cf5a.cust.comxnet.dk] has joined #code |
| 13:55 | < SmithKurosaki> | I love it when things just decide to work for me. Thanks for the help before guys |
| 13:59 | <@TheWatcher> | froztbyte: alas, something far more complicated related to splitting data up into blocks and then filtering it - only I was doing filtering first, and breaking shit |
| 14:04 | <@froztbyte> | TheWatcher: erf |
| 14:04 | <@froztbyte> | TheWatcher: I've been wanting to build a general framework/tool for doing that sort of thing for ages now |
| 14:04 | <@froztbyte> | I tend to often pull off insane hacks with sed, cut, awk, some shell loops and lots of redirection |
| 14:14 | | gnolam [lenin@Nightstar-202a5047.priv.bahnhof.se] has quit [[NS] Quit: Fuck it, I'm going back to bed.] |
| 14:19 | | Thaqui [Thaqui@27B34E.D54D49.F53FA1.6A113C] has quit [Connection closed] |
| 14:40 | | You're now known as TheWatcher[afk] |
| 15:04 | | Kindamoody is now known as Kindamoody|out |
| 15:18 | | SmithKurosaki [smith@Nightstar-87f33f96.cpe.mountaincable.net] has quit [Ping timeout: 121 seconds] |
| 15:37 | | celticminstrel [celticminstre@Nightstar-5d22ab1d.cable.rogers.com] has joined #code |
| 16:38 | | Reiver [orthianz@9C034E.E649EA.3194C7.8381A3] has quit [Ping timeout: 121 seconds] |
| 17:24 | | SmithKurosaki [smith@Nightstar-87f33f96.cpe.mountaincable.net] has joined #code |
| 17:42 | | Stalker [Z@Nightstar-3602cf5a.cust.comxnet.dk] has quit [[NS] Quit: If the world didn't suck, we'd all fall off.] |
| 18:00 | <@ToxicFrog> | <3 git stash |
| 18:07 | | Irssi: #code: Total of 23 nicks [20 ops, 0 halfops, 0 voices, 3 normal] |
| 18:28 | | Vash [Vash@Nightstar-f03c5637.sd.cox.net] has joined #code |
| 18:28 | | Vash is now known as Vornicus |
| 19:27 | | Kindamoody|out is now known as Kindamoody |
| 20:38 | | Kindamoody is now known as Kindamoody[zZz] |
| 20:56 | | thomedy [thomedy@Nightstar-7f7d888c.ca.comcast.net] has joined #code |
| 21:26 | | Stalker [Z@Nightstar-3602cf5a.cust.comxnet.dk] has joined #code |
| 21:42 | <@froztbyte> | I have heard a few people liking it quite a bit |
| 21:42 | <@froztbyte> | but, git |
| 22:12 | | celticminstrel [celticminstre@Nightstar-5d22ab1d.cable.rogers.com] has quit [[NS] Quit: KABOOM! It seems that I have exploded. Please wait while I reinstall the universe.] |
| 22:56 | | thomedy [thomedy@Nightstar-7f7d888c.ca.comcast.net] has quit [Ping timeout: 121 seconds] |
| 23:04 | | * Tamber hmms. |
| 23:27 | | Vornicus [Vash@Nightstar-f03c5637.sd.cox.net] has quit [Client closed the connection] |
| 23:45 | <@ToxicFrog> | froztbyte: ? |
| 23:59 | | shade_of_cpux is now known as cpux |
| --- Log closed Thu Aug 18 00:00:07 2011 |