Back
[11:14:05] <KGB-linuxcnc> 03Jeff Epler 05jepler/anonymous-net c20c94c 06linuxcnc 10src/hal/hal_lib.c 10src/hal/hal_priv.h hal: new private APIs to 'newsig', 'linkps' with lock held * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=c20c94c
[11:14:05] <KGB-linuxcnc> 03Jeff Epler 05jepler/anonymous-net 0ec9720 06linuxcnc 10src/hal/utils/halcmd_commands.c halcmd: remove race conditions in 'net' * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0ec9720
[11:14:05] <KGB-linuxcnc> 03Jeff Epler 05jepler/anonymous-net 51476c5 06linuxcnc 10src/hal/utils/halcmd.c 10src/hal/utils/halcmd_commands.c 10src/hal/utils/halcmd_commands.h 10src/hal/utils/halcmd_completion.c halcmd: allow net command to create anonymous nets * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=51476c5
[11:14:08] <KGB-linuxcnc> 03Jeff Epler 05jepler/anonymous-net 72530ad 06linuxcnc 03tests/net-anon/README 03tests/net-anon/expected 03tests/net-anon/test.hal tests: test of anonymous net command * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=72530ad
[12:41:28] <jepler> alternately, maybe it should be a new command, not 'net'...
[12:53:45] <KGB-linuxcnc> 03Jeff Epler 05jepler/2.6/proposed bd2577c 06linuxcnc 10tests/halui-jogging/.gitignore 04tests/halui-jogging/sim.var tests: no need to track this var file * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=bd2577c
[13:34:23] <linuxcnc-build> build #100 of 4017.deb-wheezy-rtai-i386 is complete: Failure [4failed shell_1] Build details are at
http://buildbot.linuxcnc.org/buildbot/builders/4017.deb-wheezy-rtai-i386/builds/100 blamelist: dummy, Sebastian Kuzminsky <seb@highlab.com>, Jeff Epler <jepler@unpythonic.net>
[13:42:01] <jepler> huh
[13:42:03] <jepler> Failed to fetch
http://buildbot.linuxcnc.org/dists/wheezy/scratch-rt/source/linuxcnc_2.6.2~jepler.2.6.proposed~bd2577c.dsc 403 Forbidden
[14:16:21] <cradek> jepler: thanks for responding to EBo assuming good faith
[14:47:27] <jepler> cradek: luckily, Ingrid pulled me away just as I was about to send a first draft
[14:47:32] <jepler> there wasn't much resemblance to the final e-mail
[15:11:04] <jepler> pcw_home: next you'll need to modify the Arduino IDE so that it can target the d8 cpu and you can sell a billion units. blue soldermask optional.
[15:19:20] <memleak> sweet, kernel 3.8.13 with 64-bit RTAI is working again :)
[16:08:00] <cradek> jepler: he retracted his offer to help in 5-6 months.
[16:11:59] <jepler> hmm wikipedia describes gentoo's portage system as being inspired by freebsd ports
[16:12:03] <jepler> cradek: maybe you'd like it
[16:14:11] <cradek> there's no way is freebsd ports is better than apt
[16:14:32] <cradek> the only thing that makes freebsd good is that it's not linux - it's not that it's not debian
[16:18:47] <jepler> cradek: in its concrete form (jepler/anonymous-net) what do you think?
[17:28:05] <cradek> jepler: I'm not sure I understand the problem it's solving. If it's something like "I don't care about this net's name and I want to explicitly make it unnamed" I'm sympathetic to John P's idea that the command should be more explicit. He suggests "anet" but I'd rather paint my bikeshed "net anonymous pin1 pin2". On the other hand I think the problem that might have led to this suggestion is "I'm sad I don't understand the net ...
[17:28:11] <cradek> ... command, and I didn't read the docs, and I'm sad that what I typed (pin names only) didn't work". If so, I really don't think that's one that needs fixing, at the cost of new implicit/magic behavior.
[17:29:41] <cradek> but frankly I think both are nonproblems, so I might not be the right one to ask.
[17:31:47] <jepler> heh
[17:32:43] <cradek> a "net anonymous pin1 pin2" implementation seems like it would be simpler than what you had to do
[17:34:49] <cradek> but I guess if you want to fix the "I'm sad" problem, don't let me stop you, particularly if the work is already done
[18:06:53] <jepler> bah, "uspace" created a lot more portability problems for linuxcnc on debian kfreebsd :-/
[18:07:46] <jepler> the good news is, it's passing a lot of runtests now, after an hour or two of hacking
[18:08:01] <jepler> (based on a tree I made a year or two ago and never pushed upstream..)
[18:28:26] <memleak> if anyone has a chance feel free to try this branch of RTAI out:
https://github.com/NTULINUX/RTAI
[18:29:53] <memleak> its the latest and greatest IPIPE code, 3.10 and 3.14 support will hit it soon, it works here on 64-bit w/ linuxcnc with isolcpus=2,3 set
[18:31:31] <memleak> IPIPE release -4 is what that branch is made for, so 3.4.55-1 will not work.
[18:34:03] <KGB-linuxcnc> 03Jeff Epler 052.6 6f93901 06linuxcnc 10tests/halui-jogging/.gitignore 04tests/halui-jogging/sim.var tests: no need to track this var file * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=6f93901
[18:34:06] <seb_kuzminsky> thanks jeff
[18:35:40] <memleak> ill bump 3.4 support to 3.4.103 and add the ubuntu 3.8 extended release kernel as well
[18:37:05] <memleak> http://kernel.ubuntu.com/git?p=ubuntu/linux.git;a=shortlog;h=refs/heads/linux-3.8.y
[18:37:46] <memleak> once i add 3.10 and 3.14 support i'll need someone to help bump up the rtapi code to the new proc changes.
[18:39:16] <memleak> (that rtai repo doesnt have ANY of the new 3.10+ proc changes, its a starting point for future RTAI development)
[18:53:06] <jepler> Runtest: 142 tests run, 142 successful, 0 failed + 0 expected
[18:57:32] <memleak> seb_kuzminsky, are building RTAI + kernel packages for debian pretty easy?
[18:58:35] <memleak> ive got a few debian linuxcnc systems and if i wanted to try the new IPIPE code on those without having to recompile and every system, is it straight forward to make the debs?
[19:09:59] <jepler> memleak: now that linuxcnc 2.6.x is out as stable, we don't have much latitude to switch kernel versions
[19:10:52] <jepler> we want to preserve the property that a user can go from e.g., linuxcnc 2.6.x+1 back to linuxcnc 2.6.x if they discover a bug
[19:11:07] <jepler> .. at best, that would require mucking about in the bootloader configuration or intervening at grub everytime
[19:11:42] <jepler> but it would ultimately be seb's call whether some details of a newer kernel were compelling enough to force this kind of inconvenient upgrade on the users at large
[19:13:14] <jepler> bbl
[19:14:05] <jepler> otoh it would be dandy to get all the way up to 3.16 by the time debian jessie is shaping up. probably pie-in-the-sky.
[19:14:08] <jepler> https://bits.debian.org/2014/07/kernel-version-for-jessie.html
[19:26:34] <memleak> when 3.10 proc changes are supported by linuxcnc i can get 3.16 working
[20:13:19] <jepler> memleak: might just rip the proc code out
[20:13:31] <jepler> I mean, if worse comes to worse
[21:11:33] <jepler> but I guess I should put that on my stack of tuits, it's not like anybody's actually familiar with our /proc code
[21:14:08] <jepler> http://emergent.unpythonic.net/files/sandbox/odd-latency.png
[21:14:14] <jepler> oddest latency graph I've seen lately
[21:14:44] <jepler> (freebsd kernel, not expected to have proper RT performance)
[21:15:03] <jepler> (but more than that, it seldom gets within 500us of the right time!)
[21:56:19] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd a7dec4a 06linuxcnc 10(6 files in 4 dirs) Don't use <linux/*.h> blindly * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a7dec4a
[21:56:20] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd 2346ac5 06linuxcnc 10src/emc/motion/dbuf.c 10src/hal/drivers/mesa-hostmot2/bitfile.c Don't use errno constants that aren't portable * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=2346ac5
[21:56:20] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd b7711ba 06linuxcnc 10(9 files in 5 dirs) Drop support for kernels below 2.6.14 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b7711ba
[21:56:22] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd c2726b9 06linuxcnc 10src/libnml/os_intf/_sem.c Avoid printf error due to size of semaphore key type * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=c2726b9
[21:56:26] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd 9200c3b 06linuxcnc 10src/rtapi/uspace_rtapi_app.cc Don't use nonportable linux socket semantics * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=9200c3b
[21:56:30] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd 55a4530 06linuxcnc 10src/configure.in 10src/rtapi/rtapi_pci.cc uspace: make libudev optional * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=55a4530
[21:56:33] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd 580c0ad 06linuxcnc 10src/rtapi/rtapi_uspace.hh uspace: ensure declaration of pthread_t is available * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=580c0ad
[21:56:38] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd 6405746 06linuxcnc 10src/rtapi/rtapi_uspace.hh 10src/rtapi/uspace_rtapi_app.cc uspace: make non-portable setfsuid conditional * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=6405746
[21:56:42] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd 975cff0 06linuxcnc 10src/rtapi/uspace_common.h uspace: Handle undefined SHM_LOCK * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=975cff0
[21:56:45] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd b888f8c 06linuxcnc 10src/hal/drivers/mesa-hostmot2/hm2_eth.c hm2_eth: make linux-specific code conditional * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b888f8c
[21:56:49] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd 02cd19b 06linuxcnc 10(6 files in 3 dirs) rtapi_parport: provide parport mode defines on !linux * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=02cd19b
[21:56:53] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd a2e3718 06linuxcnc 10src/rtapi/uspace_rtapi_app.cc uspace: make non-portable harden_rt conditional * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a2e3718
[21:56:57] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd 555f2d1 06linuxcnc 10src/rtapi/uspace_rtapi_app.cc uspace: make non-portable setaffinity conditional * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=555f2d1
[21:57:01] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd 96d6a29 06linuxcnc 10src/rtapi/uspace_rtapi_parport.cc parport: not portable * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=96d6a29
[21:57:05] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd aef3757 06linuxcnc 10(17 files in 8 dirs) Don't use non-portable integer types like __u32 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=aef3757
[22:04:42] <jepler> linuxcnc-build: I'll check back with you in the morning and see how that went
[22:08:35] <cradek> wow
[22:11:14] <jepler> 42 files changed, 263 insertions(+), 188 deletions(-)
[22:11:22] <jepler> it's not really even that much
[22:11:59] <jepler> some of it should even be in master branch, kfreebsd port aside
[22:12:11] <jepler> I wonder how close it is to building on "real" freebsd
[22:12:49] <skunkworks> then we could have zfs? :)
[22:13:59] <jepler> skunkworks: sure thing (that's why I run debian kfreebsd)
[22:14:35] <jepler> oops, some erroneous __LINUX__ in those patches
[22:18:24] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd 435c066 06linuxcnc 10src/rtapi/rtapi_uspace.hh 10src/rtapi/uspace_rtapi_app.cc uspace: make non-portable setfsuid conditional * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=435c066
[22:18:25] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd 51c27e6 06linuxcnc 10src/rtapi/uspace_common.h uspace: Handle undefined SHM_LOCK * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=51c27e6
[22:18:25] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd ad5f4ff 06linuxcnc 10src/hal/drivers/mesa-hostmot2/hm2_eth.c hm2_eth: make linux-specific code conditional * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ad5f4ff
[22:18:26] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd 8053b4a 06linuxcnc 10(6 files in 3 dirs) rtapi_parport: provide parport mode defines on !linux * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=8053b4a
[22:18:31] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd dc0b626 06linuxcnc 10src/rtapi/uspace_rtapi_app.cc uspace: make non-portable harden_rt conditional * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=dc0b626
[22:18:35] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd 8dceaab 06linuxcnc 10src/rtapi/uspace_rtapi_app.cc uspace: make non-portable setaffinity conditional * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=8dceaab
[22:18:39] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd ff522f2 06linuxcnc 10src/rtapi/uspace_rtapi_parport.cc parport: not portable * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ff522f2
[22:18:43] <KGB-linuxcnc> 03Jeff Epler 05jepler/kfreebsd 6ecc763 06linuxcnc 10(17 files in 8 dirs) Don't use non-portable integer types like __u32 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=6ecc763
