| 00:01.47 | brlcad | we need to get back on schedule to monthly postings regardless of any specific item |
| 00:13.06 | brlcad | I think we've missed two releases now |
| 00:13.22 | brlcad | we "should" be on 7.18.4 :) |
| 00:16.20 | *** join/#brlcad R0b0t1 (~Enigma@64-136-219-55.dyn.everestkc.net) |
| 00:16.20 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) |
| 01:42.38 | CIA-55 | BRL-CAD: 03brlcad * r41428 10/brlcad/trunk/sh/conversion.sh: |
| 01:42.38 | CIA-55 | BRL-CAD: add summary count statistics on what percentage and how many objects |
| 01:42.38 | CIA-55 | BRL-CAD: successfully converted. this required a reworking of how objects are iterated |
| 01:42.38 | CIA-55 | BRL-CAD: one line at a time (so objects with spaces in their name are handled correctly) |
| 01:42.38 | CIA-55 | BRL-CAD: using a here document. also add a GED parameter so you can specify which mged |
| 01:42.38 | CIA-55 | BRL-CAD: you want to use instead of whatever is in your path. |
| 01:51.25 | CIA-55 | BRL-CAD: 03brlcad * r41429 10/brlcad/trunk/sh/conversion.sh: ah, right. support VARIABLE=value arguments. also give more informative failure if we can't find MGED. |
| 02:35.56 | CIA-55 | BRL-CAD: 03brlcad * r41430 10/brlcad/trunk/sh/conversion.sh: |
| 02:35.56 | CIA-55 | BRL-CAD: boo yah. add elapsed conversion times. this does even a better job than the |
| 02:35.56 | CIA-55 | BRL-CAD: benchmark since it counts seconds from the beginning of the century and should |
| 02:35.56 | CIA-55 | BRL-CAD: track time across days/weeks. it reports both cumulative time elapsed, average |
| 02:35.56 | CIA-55 | BRL-CAD: per object (including process startup overhead), and real time per nmg/bot |
| 02:35.56 | CIA-55 | BRL-CAD: conversion. |
| 02:50.26 | *** join/#brlcad stevegt_1 (~stevegt@c-69-181-134-76.hsd1.ca.comcast.net) |
| 04:05.22 | brlcad | hells yeah |
| 04:05.49 | CIA-55 | BRL-CAD: 03brlcad * r41431 10/brlcad/trunk/sh/conversion.sh: (log message trimmed) |
| 04:05.49 | CIA-55 | BRL-CAD: damn I'm good. implement some scary mad shell scripting here in order to kill |
| 04:05.49 | CIA-55 | BRL-CAD: long-running conversions while still timing them and capturing their output. to |
| 04:05.49 | CIA-55 | BRL-CAD: make this happen, we can no longer just use a here document fed to the 'while |
| 04:05.49 | CIA-55 | BRL-CAD: read' loop because the kill signals sent to the children processes jack it up |
| 04:05.49 | CIA-55 | BRL-CAD: good. instead, temp override stdin with our object list so the loop continues |
| 04:05.50 | CIA-55 | BRL-CAD: unabated. add in some timer cleanup for the instances when we finish the |
| 05:31.04 | *** join/#brlcad cosurgi (~cosurgi@atak.bl.pg.gda.pl) |
| 06:23.10 | CIA-55 | BRL-CAD: 03brlcad * r41432 10/brlcad/trunk/sh/conversion.sh: put 'instructions' before 'help' so we can get to it without a .g specified. |
| 07:40.59 | *** join/#brlcad d_rossberg (~rossberg@BZ.BZFLAG.BZ) |
| 07:41.50 | d_rossberg | brlcad: thanks! |
| 08:04.49 | *** join/#brlcad merzo (~merzo@193.254.217.44) |
| 08:58.35 | *** join/#brlcad mafm_ (~mafm@36.Red-79-159-0.staticIP.rima-tde.net) |
| 11:46.08 | *** join/#brlcad ibot (~ibot@rikers.org) |
| 11:46.08 | *** topic/#brlcad is BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || #brlcad logs: http://ibot.rikers.org/%23brlcad/ || BRL-CAD release 7.16.10 is posted! (20100805) |
| 11:49.36 | d-lo | hehehe, looks like BRLCAD is rather proud of the last few commits :) |
| 11:53.15 | *** join/#brlcad juanman (~quassel@unaffiliated/juanman) |
| 12:15.21 | ``Erik | I have a feeling that d-lo is awfully close to doing something that'll confuse him |
| 12:15.53 | d-lo | ?? |
| 12:16.03 | ``Erik | reading backlog and trying to catch up |
| 12:16.23 | ``Erik | um, it's a consumer/producer problem, the kernel will queue up to a certain size |
| 12:16.24 | d-lo | nah, tracking down a stilly bug |
| 12:16.25 | ``Erik | and then stop |
| 12:16.58 | ``Erik | so if you try to blast a monster packet down, it'll "lock up", but it's just waiting for the consumer to evacuate the queue a bit |
| 12:17.16 | ``Erik | keith was just messed up by this, with an inter-tcl mapping |
| 12:18.30 | ``Erik | nuffim pacific, but this is a common category of issue, don't spool yourself up if'n ya see it :) |
| 12:19.17 | d-lo | whatcha talking about? the whole byte arrival assurance thingy brlcad and i were yacking about? |
| 12:19.28 | ``Erik | yeh |
| 12:19.33 | d-lo | kk |
| 12:19.38 | ``Erik | like I said, catching up on backlog |
| 12:19.47 | d-lo | orite |
| 12:19.51 | d-lo | forgot :) |
| 12:20.50 | ``Erik | hopefully, you're not choking the pipe mechanism, but if things suddenly go wonky, stop and step back, aight? |
| 12:21.47 | ``Erik | (and yeh, brlcad seems to have slewn up some hubris with those commits... kinda tempted to figure out what he did wrong, but I doubt I'd find anything) |
| 12:22.11 | d-lo | "slewn up some hubris" .....lol |
| 12:22.32 | d-lo | oh yea. |
| 12:22.39 | d-lo | Astro just olpened a new server, you on it? |
| 12:22.40 | ``Erik | can ya think of a better way to put it? ;) |
| 12:22.44 | ``Erik | no |
| 12:22.58 | ``Erik | I lost both my mobiles this weekend |
| 12:23.12 | d-lo | ack, big ass battles? |
| 12:23.23 | ``Erik | major crash on epsi, we lost |
| 12:23.32 | ``Erik | and on fenix, they'd given a "go hide" order and I got saw |
| 12:24.06 | d-lo | You still with F.A.T.E. ? |
| 12:24.15 | ``Erik | on epsi, yeh |
| 12:24.22 | ``Erik | on fenix, I'm on the opposing side |
| 12:24.27 | d-lo | lol |
| 12:24.28 | d-lo | nice |
| 12:25.16 | ``Erik | I fight for the flag I'm under, ain't gonna be deceitful |
| 12:25.41 | d-lo | so everyone's aware your on different sides on different servers? |
| 12:25.54 | ``Erik | no |
| 12:26.04 | ``Erik | no one's asked and it isn't anyones business |
| 12:26.30 | d-lo | that's awesome :) |
| 12:26.42 | d-lo | get any good intel that way? |
| 12:26.55 | ``Erik | nope, different sets of people |
| 12:27.09 | ``Erik | and even if I did, I wouldn't be using it *shrug* |
| 12:27.44 | d-lo | not accusing you of anything man, I just think its kinda funny :) |
| 12:28.35 | ``Erik | I find it kinda ironic myself, but I don't think there's any real conflict of interest |
| 12:29.08 | ``Erik | ah, there's the recyc truck |
| 12:30.27 | ``Erik | so yeh, 30m fleet this weekend, without getting either pile |
| 12:30.53 | d-lo | ouch |
| 12:31.00 | d-lo | that's gonna cost ya in rebuild time. |
| 12:31.05 | d-lo | still got base defense? |
| 12:31.21 | ``Erik | yeh, but I've thinned those down to a single dn |
| 12:31.25 | *** join/#brlcad juanman (~quassel@unaffiliated/juanman) |
| 12:31.28 | ``Erik | and one base is occ'd |
| 12:32.06 | d-lo | Hrm, not all THAT bad i suppose. |
| 12:32.11 | d-lo | whats the occ force look like? |
| 12:32.17 | ``Erik | 2m |
| 12:32.44 | ``Erik | woops, he saw my inc, he's dropped to a single fighter |
| 12:32.48 | ``Erik | musta scooped it alll |
| 12:32.51 | d-lo | :) |
| 12:32.56 | d-lo | oh, :( |
| 12:42.50 | ``Erik | http://mindfunction.com/~erik/m82.png |
| 12:45.31 | d-lo | that your listhp project? |
| 12:45.52 | ``Erik | ayup |
| 12:46.38 | d-lo | so.... whats mindfunction.com again? A pardner in crime for the listhp project? |
| 12:46.46 | ``Erik | pretty much |
| 14:34.01 | d-lo | brlcad: you around or on the road for turkey day? |
| 14:36.11 | d-lo | brlcad: found an interesting tidbit that's causing a bit of grief |
| 14:37.22 | d-lo | on a 'read' action, the caller is expected to process,suckin,process and that works just fine. |
| 14:37.49 | d-lo | but on a write action, the 'suckin' function is called, but no process. |
| 14:38.23 | d-lo | and in the _pkg_checkin() function, there's a select call that moves data off the socket into the temp buffer. |
| 14:39.06 | d-lo | and I think that is causing my higher level select() call to occasionally miss something. |
| 14:39.32 | d-lo | it reports 0 since the data has already been moved off the socket's buffer. |
| 14:41.31 | d-lo | simple fix was for me to 'short circuit' the selector loop and make it attempt to read/process each FD each loop pass. |
| 14:44.21 | d-lo | ``Erik: that you going thump thump thump? |
| 14:51.10 | CIA-55 | BRL-CAD: 03X Tin Basher 07http://brlcad.org * r2348 10/wiki/EBM: /* Making an image with GIMP */ |
| 14:51.16 | brlcad | d_rossberg: you're welcome.. what for? :) |
| 14:52.33 | starseeker | wow - koffice is splitting into two groups |
| 14:52.53 | brlcad | ``Erik: that wasn't hubris, it works! |
| 14:52.58 | brlcad | it took several hours to figure out how to do what it's doing without touching disk, so yeah, I'm happy :) |
| 14:54.43 | brlcad | d-lo: I'll be on the road later today, but here online for a bit |
| 15:04.43 | brlcad | NMG conversion: 97.0% (8964 of 9244 objects) |
| 15:04.44 | brlcad | BoT conversion: 96.8% (8948 of 9244 objects) Success rate: 96.9% |
| 15:04.51 | brlcad | Success rate: 96.9% |
| 15:05.07 | brlcad | Elapsed: 1702 seconds |
| 15:05.14 | starseeker | huh - any pattern to the failures? |
| 15:05.29 | brlcad | and that was with a conversion limit of 5 seconds |
| 15:05.35 | starseeker | sweet |
| 15:06.59 | brlcad | that is pre-changes, so next up is to check the latest |
| 15:07.35 | starseeker | ah, cool |
| 15:08.40 | brlcad | that's pretty freaking cool that have it auto-log and summarize stats now |
| 15:09.02 | d_rossberg | brlcad: your fast answer yesterday ;) |
| 15:09.59 | brlcad | oh! heh, forgot about that :) |
| 15:10.43 | brlcad | d_rossberg: and I confirmed, he is still the person to contact |
| 15:12.32 | brlcad | http://brlcad.org/tmp/conversion-11663-run.log |
| 15:13.30 | brlcad | so next, to get that rate up to 100%... |
| 15:20.46 | CIA-55 | BRL-CAD: 03davidloman * r41433 10/rt^3/trunk/src/libNet/Portal.cxx: Cleaned up some logger calls. |
| 15:25.33 | CIA-55 | BRL-CAD: 03starseeker * r41434 10/brlcad/branches/cmake/src/fb/CMakeLists.txt: fbthreadtest needs X11/X11.h |
| 15:34.53 | starseeker | huh, kinda neat: http://www.gnu.org/software/libmicrohttpd/ |
| 15:37.20 | CIA-55 | BRL-CAD: 03brlcad * r41435 10/brlcad/trunk/sh/conversion.sh: |
| 15:37.20 | CIA-55 | BRL-CAD: add summary of file and object counts along with the failure counts so we don't |
| 15:37.20 | CIA-55 | BRL-CAD: have to subtract. also quiet the killing of the timer because of the race |
| 15:37.20 | CIA-55 | BRL-CAD: condition where it finishes after we get the pid but before the kill. lastly, |
| 15:37.20 | CIA-55 | BRL-CAD: specifying ksh was just for testing, not required. set sh instead, but go ahead |
| 15:37.21 | CIA-55 | BRL-CAD: and set posix mode too (just because we can). |
| 15:41.33 | CIA-55 | BRL-CAD: 03X Tin Basher 07http://brlcad.org * r2349 10/wiki/Talk:Main_Page: |
| 15:41.46 | CIA-55 | BRL-CAD: 03davidloman * r41436 10/rt^3/trunk/src/libNet/PortalManager.cxx: (log message trimmed) |
| 15:41.46 | CIA-55 | BRL-CAD: Fix a bug that had to deal with the selector loop in PortalManager occasionally |
| 15:41.46 | CIA-55 | BRL-CAD: missing a read. Turns out there is an underlying select() call deep in libPkg |
| 15:41.46 | CIA-55 | BRL-CAD: that reads data from a socket and buffers it internal to libPkg. That select() |
| 15:41.47 | CIA-55 | BRL-CAD: call is called on both high level read and write operations. However, on the |
| 15:41.47 | CIA-55 | BRL-CAD: write op, the data that is read from the socket and NOT 'dispatched', thus the |
| 15:41.48 | CIA-55 | BRL-CAD: callback never gets called. The quick fix for this is to make the |
| 15:42.37 | CIA-55 | BRL-CAD: 03X Tin Basher 07http://brlcad.org * r2350 10/wiki/Talk:Main_Page: |
| 15:44.54 | CIA-55 | BRL-CAD: 03davidloman * r41437 10/rt^3/trunk/src/libNet/PortalManager.cxx: That's INFO not an ERROR! |
| 15:44.54 | ``Erik | neat |
| 15:46.28 | *** join/#brlcad mafm (~mafm@36.Red-79-159-0.staticIP.rima-tde.net) |
| 15:48.14 | CIA-55 | BRL-CAD: 03davidloman * r41438 10/rt^3/trunk/src/libNet/Portal.cxx: Clean up some debug printing calls. Changed the bu_bomb() call in Portal::callbackSpringboard() to a ERROR log call. We don't want to take the whole app down if one buffer reference comes thru as null. |
| 15:48.53 | CIA-55 | BRL-CAD: 03Sean 07http://brlcad.org * r2351 10/wiki/Talk:Main_Page: hi x tin basher |
| 15:57.46 | CIA-55 | BRL-CAD: 03brlcad * r41439 10/brlcad/trunk/sh/conversion.sh: er, need elp before computing avg so reorder. sprinkle a few comments too. |
| 16:01.12 | CIA-55 | BRL-CAD: 03davidloman * r41440 10/rt^3/trunk/src/libNet/ (5 files): Clean up comment types. C++ -> C style comments. |
| 16:01.32 | CIA-55 | BRL-CAD: 03brlcad * r41441 10/brlcad/trunk/sh/conversion.sh: jack up the default MAXTIME to 5 minutes. if an object takes longer than that, it'll probably take a LOT longer. |
| 16:08.42 | *** join/#brlcad merzo (~merzo@193.254.217.44) |
| 16:10.40 | CIA-55 | BRL-CAD: 03davidloman * r41442 10/rt^3/trunk/src/libNet/netMsg/ (16 files): More c++ -> c style comments conversion. |
| 16:40.27 | CIA-55 | BRL-CAD: 03starseeker * r41443 10/brlcad/branches/cmake/src/other/CMakeLists.txt: Completely untested, but start adding logic to get auto_path from system tcl for brlcad_config.h |
| 16:47.49 | CIA-55 | BRL-CAD: 03davidloman * r41444 10/rt^3/trunk/src/libNet/netMsg/NetMsg.cxx: Clay: WS, Formatting. |
| 17:48.07 | *** join/#brlcad Yoshi47 (~jan@64.235.102.210) |
| 18:11.58 | CIA-55 | BRL-CAD: 03brlcad * r41445 10/brlcad/trunk/sh/conversion.sh: back down to 30 until I figure out a better way to kill all of the lingering sleep processes that build up |
| 18:32.15 | *** join/#brlcad merzo (~merzo@50-2-94-178.pool.ukrtel.net) |
| 18:59.04 | CIA-55 | BRL-CAD: 03starseeker * r41446 10/brlcad/branches/cmake/src/bwish/main.c: Equally untested, but try to add the paths from the system tcl to the auto_path for btclsh/bwish |
| 18:59.27 | CIA-55 | BRL-CAD: 03X Tin Basher 07http://brlcad.org * r2352 10/wiki/Talk:Main_Page: |
| 19:08.27 | CIA-55 | BRL-CAD: 03starseeker * r41447 10/brlcad/branches/cmake/src/bwish/main.c: Tweaks to bwish main.c code. |
| 19:17.04 | CIA-55 | BRL-CAD: 03starseeker * r41448 10/brlcad/branches/cmake/src/other/CMakeLists.txt: Whoops, copy/paste typo. |
| 19:23.03 | CIA-55 | BRL-CAD: 03starseeker * r41449 10/brlcad/branches/cmake/src/other/CMakeLists.txt: Need to strip the newline off for the define |
| 19:47.50 | CIA-55 | BRL-CAD: 03starseeker * r41450 10/brlcad/branches/cmake/src/bwish/main.c: |
| 19:47.50 | CIA-55 | BRL-CAD: OK, the real issue here appears to actually be tclcadAutoPath setting |
| 19:47.50 | CIA-55 | BRL-CAD: tcl_library to something Not Helpful - don't need this, although probably still |
| 19:47.50 | CIA-55 | BRL-CAD: want the logic to probe the system auto_path - just need to find the actual dir |
| 19:47.50 | CIA-55 | BRL-CAD: with init.tcl and have tclcadAutoPath set that straight up. |
| 19:56.48 | CIA-55 | BRL-CAD: 03starseeker * r41451 10/brlcad/branches/cmake/src/other/CMakeLists.txt: Switch to actually hunting down the init.tcl file based on the system tcl's autopath list - not tested. |
| 20:00.33 | CIA-55 | BRL-CAD: 03starseeker * r41452 10/brlcad/branches/cmake/src/other/CMakeLists.txt: Oh yeah, might help to tell it what to look for |
| 20:03.03 | CIA-55 | BRL-CAD: 03starseeker * r41453 10/brlcad/branches/cmake/src/other/CMakeLists.txt: Actually we want the path the dir containing the file, not the file itself. |
| 20:18.30 | CIA-55 | BRL-CAD: 03starseeker * r41454 10/brlcad/branches/cmake/src/other/CMakeLists.txt: Fix naming of INITTCL variable. |
| 20:21.50 | CIA-55 | BRL-CAD: 03starseeker * r41455 10/brlcad/branches/cmake/src/libtclcad/tclcadAutoPath.c: Make use of the TCL_SYSTEM_INITTCL_PATH variable and spell out the init.tcl path for system tcl for tclcadAutoPath |
| 20:54.30 | *** join/#brlcad R0b0t1 (~Enigma@64-136-219-55.dyn.everestkc.net) |
| 20:54.30 | *** join/#brlcad R0b0t1 (~Enigma@unaffiliated/r0b0t1) |
| 22:59.47 | *** join/#brlcad Ralith (~ralith@d142-058-094-158.wireless.sfu.ca) |
| 23:00.20 | *** join/#brlcad australian_male_ (~PrAyInG^E@186.212.226.26) |
| 23:00.21 | *** part/#brlcad australian_male_ (~PrAyInG^E@186.212.226.26) |
| 23:13.06 | *** join/#brlcad IriX64 (~kvirc@bas2-sudbury98-1096600637.dsl.bell.ca) |
