#emc-devel | Logs for 2010-10-16

[00:18:52] <NTU> Hey! I cannot find the upstream (non-ubuntu specific) source for EMC 2.4.5 in a tarball and was wondering if somebody could push 2.4.5 to sf.
[00:19:46] <cradek> NTU: seems it is not on sf yet. you can get it from our git: git.linuxcnc.org
[00:20:13] <cradek> by the tag v2.4.5, click commit to see it, click snapshot to get a .tgz
[00:20:54] <NTU> seems as if v2.4.5 isn't merged in master. only in v2.4_branch
[00:21:42] <cradek> yes v2.4_branch is periodically merged into master
[00:22:17] <cradek> v2.4.5 is a particular commit on v2.4_branch
[00:22:29] <cradek> you can get a tgz of any of these from our git server
[00:23:32] <NTU> seems to take about 3 minutes to ask me if i want to download it
[00:23:54] <cradek> yes - the server generates it for you - it is a fairly intensive task
[00:25:59] <jepler> afaik the only difference between our debian source .tgz and the one that goes on sourceforge is a single file in the debian/ directory that's been customized for an ubuntu version and sim vs rtai
[00:26:06] <jepler> it should be entirely suitable for building your own
[00:27:20] <cradek> bbl
[00:28:35] <NTU> oh ok, thanks for the info jepler. :)
[01:09:11] <NTU> NTU is now known as Godmode
[02:11:26] <qq-> jepler, found a typo in 2.4.5 debian/configure line 128: KERNEL_HEADERS="$KERNEL_HEADERS" < extra "" ?
[02:13:16] <jepler> I agree it's inconsistent with the other lines near it, but I don't think it causes any problem
[02:14:40] <qq-> ok , i suggest to add in : 'case $DISTRIB_NAME in' "Debian-6.0*)
[02:14:40] <qq-> EXTRA_BUILD=lyx
[02:14:40] <qq-> TCLTK_VERSION=8.4"
[02:15:32] <qq-> well even in squeeze TCLTK may be 8.5 too
[02:16:51] <jepler> most debian/ubuntu systems have more than one tcltk available. The one you need is the one that /usr/bin/python's Tkinter module uses.
[02:19:16] <jepler> according to http://packages.debian.org/testing/python/python-tk it'll be 8.5 that you want to specify.
[02:20:52] <qq-> jepler, my squeeze system use 8.4, but that is maybe because i upgraded from lenny
[02:23:22] <qq-> other suggestion is to add " 2.6.32-*-rtai)" \ KERNEL_HEADERS=$KERNEL_HEADERS,gcc-4.4 , for kernels from series .32 build with -rtai
[02:25:06] <jepler> ok
[02:25:30] <jepler> once you're satisfied with the squeeze debian/configure changes, can you send it to me as a patch?
[02:25:35] <jepler> or better yet, send it on the mailing list
[02:25:54] <jepler> * jepler is about to call it a night
[02:26:09] <qq-> as used in ubuntu-cnc 10.04 , or debian squeeze home build as mine "-2.6.32-5-686-squeeze-rtai"
[02:27:01] <qq-> jepler, sure , now just to a study case for debian squeeze
[02:27:14] <qq-> , now just do *
[02:31:08] <qq-> * qq- here too is deep 'night'
[13:22:07] <cradek> jepler: Packages* in http://www.linuxcnc.org/emc2/dists/hardy/emc2.4/binary-i386/ were not regenerated
[13:22:40] <cradek> symptom is that hardys are not getting 2.4.5
[13:27:06] <jepler> cradek: hm, let me double check..
[13:27:07] <jepler> should be quick to fix
[13:27:30] <jepler> fixed.
[13:29:09] <cradek> thanks
[15:34:36] <andypugh> My code needs to write to a register on the Mesa card, then wait for that register to be cleared. This might not always happen. What is the neatest way to code a timeout in kernel code? Normally I would not the time that the function was entered, then keep checking, but that needs time.h which is not in the includes, as far as I can see.
[15:35:19] <andypugh> is there something in the rtapi libraries?
[15:37:18] <andypugh> As I was typing that, I thought "why don't I just read the rtapi header files and look" and as if by magic, there it is.
[15:37:24] <andypugh> Sorry for wasting your time.
[15:43:05] <qq-> * qq- if you really sorry , don't ask
[15:44:05] <andypugh> I becale sorry after realising how easy it was to answer my own question.
[18:36:04] <SteveStallings> SteveStallings is now known as steves_logging
[20:05:29] <mhaberler_> mhaberler_ is now known as mhaberler