| 00:00.23 | mikelp | scroll up a bit, there should be a load of lines starting Configuring: |
| 00:00.44 | mikelp | just before that it'll say if it couldn't find package: blah |
| 00:00.49 | JediMaster | nope, not a single line with it in |
| 00:00.54 | JediMaster | it's just a big bash script |
| 00:00.58 | JediMaster | sure it's the right log file? |
| 00:01.02 | mikelp | :q |
| 00:01.03 | mikelp | gar |
| 00:01.16 | mikelp | in tmp/work/opie-image-1.0-r0/temp |
| 00:01.31 | mikelp | It starts off looking like a shell script |
| 00:01.38 | JediMaster | log.do_rootfs.3275 |
| 00:01.48 | JediMaster | it starts and ends looking like one lol |
| 00:03.24 | mikelp | Oh well, should be at least a hint in there to what went wrong |
| 00:03.42 | JediMaster | nothing but shell script |
| 00:04.38 | mikelp | first line is: + rm -rf /home/phillim/zaurus/20040819/tmp/rootfs |
| 00:04.38 | mikelp | ? |
| 00:04.49 | mikelp | well, similar anyways |
| 00:04.54 | JediMaster | is there a "make clean" type command? |
| 00:05.19 | JediMaster | not in mine |
| 00:05.20 | JediMaster | #!/bin/sh -e |
| 00:05.20 | JediMaster | export CPPFLAGS="-I/root/oe/build/tmp/staging/arm-linux/include" |
| 00:05.22 | JediMaster | is mine |
| 00:05.50 | mikelp | Hmmm, that looks like a run.do_blah, not log |
| 00:06.20 | JediMaster | oops, my bad lol |
| 00:06.30 | mikelp | ~lart JediMaster |
| 00:09.12 | JediMaster | well... |
| 00:10.23 | JediMaster | it starts failing after the huge long ipkg-cl line |
| 00:10.46 | mikelp | I just started a oemake opie-image, lets see if it fails |
| 00:16.56 | JediMaster | mikelp: this is openzaurus/collie btw |
| 00:17.25 | mikelp | mine's husky, close enough |
| 00:17.45 | *** join/#oe qip|Virus|da (~Thomas@p5091B404.dip.t-dialin.net) |
| 00:18.33 | mikelp | grrrm cvs update of opie-*, its rebuilding them all |
| 00:18.48 | JediMaster | lol |
| 00:18.57 | mikelp | won't take long, just irritating |
| 00:19.05 | treke|home | the checkout is the bad part |
| 00:19.11 | treke|home | the building is pretty quick |
| 00:19.21 | JediMaster | I presume it's not possible to take advantage of a multi processor system to build it? |
| 00:19.35 | mikelp | JediMaster: certain packages hate -jN |
| 00:19.45 | mikelp | the install part of glibc is the biggest culprit |
| 00:19.48 | JediMaster | I'd heard |
| 00:19.50 | mikelp | that's on the todo list |
| 00:19.52 | treke|home | opie doesnt have that big of a problem with it |
| 00:20.11 | treke|home | and if you dont mind babysitting, you can build multiple opie packages at a time |
| 00:21.20 | mikelp | yeah, farming out packages to different processors would be the biggest gain |
| 00:21.29 | mikelp | there's not a lot that takes *that* long to compile |
| 00:21.54 | JediMaster | yeah agreed, really sucks to see 20% cpu usage when building on a dual xeon |
| 00:22.24 | treke|home | what I've always found to be a big gain is running a second oemake process to download all the source in advance |
| 00:22.39 | mikelp | python should be taking more that that |
| 00:22.40 | treke|home | so you dont spend time waiting between builds |
| 00:24.49 | mikelp | my opie-image just finished fine |
| 00:25.03 | JediMaster | mikelp: python takes about 95% cpu time on 1/4 cpus |
| 00:25.13 | JediMaster | works out at about 20% overall |
| 00:25.16 | mikelp | holy crap, one of the folding@home's has take 2130 minutes of cpu time so far |
| 00:25.26 | JediMaster | technically it's only 2 real cpus with hyperthreading |
| 00:26.01 | mikelp | nod, mine's a dual opteron setup |
| 00:26.56 | *** join/#oe mithro (~tim@dsl1-83.gw1.adl1.airnet.com.au) |
| 00:33.05 | zecke | good nite guys |
| 00:56.07 | cterry | hmm |
| 01:16.01 | JediMaster | how do you build the kernel for the zaurus with 64mb ram and 0 for ramdisk? |
| 01:16.18 | JediMaster | rather than the default 32-32 |
| 01:31.37 | *** join/#oe qfh (~qfh@ppp-82-135-0-192.mnet-online.de) |
| 01:42.30 | *** join/#oe ChanServ (ChanServ@services.) |
| 01:42.30 | *** mode/#oe [+o ChanServ] by irc.freenode.net |