#emc-devel | Logs for 2007-01-08

[02:32:55] <jepler> huh -- how long has "halcmd show | less" been broken?
[02:33:10] <jepler> apparently it now writes to stderr, but didn't before?
[02:33:15] <jepler> I must have busted it
[02:33:34] <jmkasunich> I have no idea
[02:33:48] <jepler> oh yeah -- I did
[02:33:48] <jmkasunich> it certainly didn't go to stderr orginally
[02:33:54] <jepler> rtapi_print went to stdout and rtapi_print_msg went to stderr
[02:34:01] <jepler> but I broke it so that rtapi_print went to stderr too
[02:34:04] <jepler> for sim only, I think
[02:34:55] <jepler> yep that's it
[05:58:26] <cradek> ok someone check something in!
[05:59:30] <cradek> SWPadnos: you know, now I think Reply-To should be -devel
[05:59:44] <SWPadnos> that would make some sense
[06:00:00] <SWPadnos> I liked the fact that I could see who made the commit in my folder list though
[06:00:21] <SWPadnos> I could easily filter out only commits by alex, on the interp ...
[06:00:30] <cradek> well I'm not going to be able to set From so forget that (although I liked it too)
[06:00:46] <SWPadnos> heh
[06:01:17] <cradek> because I think that will make sf not like the messages
[06:01:24] <cradek> I think
[06:01:39] <cradek> I'm fuzzy about this now, it's been too long since we figured out the hoops
[06:02:40] <SWPadnos> did you have to register the email the CVS server uses with SF?
[06:02:47] <cradek> yes
[06:02:57] <cradek> it's subscribed to the commit list with no-email turned on
[06:03:14] <SWPadnos> right. you need an SF account to send messages to any of the lists
[06:03:38] <cradek> no you don't
[06:03:43] <SWPadnos> really?
[06:03:48] <cradek> you just need to be subscribed
[06:03:58] <cradek> 90% of our subscribers do not have sf accounts
[06:04:07] <SWPadnos> ah - will your ISP block a spoofed FROM address?
[06:04:38] <SWPadnos> you also need the same thing for the unswitched line
[06:04:39] <cradek> I think sf would (rightfully) block it
[06:04:41] <SWPadnos> oops
[06:05:08] <SWPadnos> how would SF know - they don't necessarily get the mail from the developer's actual ISP
[06:05:09] <cradek> because it would have to say @users.sf.net
[06:05:19] <SWPadnos> hmm
[06:05:50] <cradek> but maybe they do allow that - some people might use it that way
[06:05:59] <SWPadnos> heh - be sneaky. reply-to = Chris Radek <emc-developers@lists.sourceforge.net>
[06:06:15] <SWPadnos> just change the name that's sent, not the address
[06:06:33] <cradek> From: even
[06:06:37] <SWPadnos> sure
[06:06:50] <SWPadnos> From: Developer Name <cvs-adm@cvs.linuxcnc.org>
[06:07:03] <SWPadnos> and add in the reply-to field if you can
[06:07:06] <cradek> I wonder if I get $USER in my shell
[06:07:13] <cradek> I already put in the Reply-To
[06:07:16] <SWPadnos> ok
[06:07:26] <cradek> (doing anything inside the chroot is hard)
[06:07:42] <cradek> do you have a checkout of documents?
[06:07:51] <SWPadnos> I don't see any reply-to
[06:07:58] <SWPadnos> or did you just do that?
[06:08:02] <cradek> I just did it
[06:08:06] <SWPadnos> ok
[06:11:22] <SWPadnos> damn. I need to figure out how to get cvs/SSH to work on this stupid windows machine
[06:11:26] <cradek> darn
[06:11:34] <cradek> no $USER in the environment
[06:11:39] <SWPadnos> bummer
[06:11:49] <cradek> it just went from easy to hard
[06:12:44] <SWPadnos> well, the reply-to worked at least
[06:12:48] <cradek> yeah
[06:12:55] <cradek> that's easy because it's a fixed string
[06:12:58] <SWPadnos> yep
[06:13:16] <cradek> see author:cradek? that may be the only way to know the user
[06:13:26] <SWPadnos> right
[06:13:39] <cradek> and that's a nightmare because the headers were already generated long ago when that diff is made
[06:15:29] <SWPadnos> and it's not a "real" login
[06:15:35] <SWPadnos> hence no user name
[06:15:39] <SWPadnos> $USER name
[06:16:29] <cradek> no, it is a real login, but I bet the env is (rightfully) cleaned
[06:16:46] <cradek> maybe I can put $USER in loginfo...
[06:17:21] <cradek> I don't know where to find the spec for loginfo though
[06:17:39] <SWPadnos> http://www.cvsnt.org/manual/html/loginfo.html
[06:17:41] <SWPadnos> ?
[06:18:06] <SWPadnos> hmmm - maybe the apple version would be closer
[06:18:55] <cradek> hmm, that has $USER in the example but not in the documentation
[06:20:09] <SWPadnos> http://www.merten-home.de/FreeSoftware/lcvs/manual.html#loginfo_options
[06:20:15] <SWPadnos> not many there though
[06:22:08] <SWPadnos> heh - does it use the new or the old version for the commit message? :)
[06:22:36] <cradek> I think it screws up
[06:22:40] <cradek> I try not to think about it
[06:23:12] <cradek> well hey look at that, it works
[06:23:34] <SWPadnos> it does?
[06:23:53] <cradek> yeah look at X-User:
[06:24:06] <cradek> echo "From: $USER <cvs-adm@cvs.linuxcnc.org>"
[06:24:10] <SWPadnos> ah yes - there it is
[06:24:11] <cradek> is this what we want?
[06:24:42] <SWPadnos> that's preferable to everything being from the EMC CVS server
[06:24:59] <SWPadnos> if you can use the "real names" that would be better, but that's likely a real pain
[06:27:01] <cradek> hmm, even id doesn't give that
[06:27:34] <SWPadnos> hmmm - I guess I should go break down all those boxes before going to bed :(
[06:27:39] <SWPadnos> cleaning the office is so much fun
[06:37:11] <cradek> while read i; do case $i in $USER:*) NAME=${i#*:*:*:*:}; NAME=${NAME%%:*}; echo $NAME;; esac; done </etc/passwd
[06:37:17] <cradek> who needs awk
[06:38:56] <cradek> for some reason IFS=: set -- $i doesn't work. I wish I understood how to use IFS
[06:40:39] <cradek> kick ass
[06:40:44] <cradek> bedtime!
[06:40:48] <SWPadnos> what the hell did that code do?
[06:40:55] <cradek> ?
[06:41:08] <cradek> oh
[06:41:12] <cradek> username => "User Name"
[06:41:14] <SWPadnos> while read i; do ${lots of scary shit}!
[06:41:34] <cradek> without calling any subprograms (nothing is available in the chroot unless I build it static)
[06:41:39] <SWPadnos> cool! that's perfect for me. thanks :)
[06:41:41] <SWPadnos> ah
[06:42:12] <cradek> alex wanted that since day 1 and I always said "maybe someday"
[06:42:16] <SWPadnos> heh :)
[06:42:19] <cradek> ok goodnight
[06:42:33] <SWPadnos> well, more box-breaking for me. good night to you
[15:48:29] <cradek> I made two changes to the commit emails: they now show the committer's name in From:, and have Reply-To: set to emc-devel.
[16:37:27] <maddash> are there any precompiled rtai-enabled kernel imgs for debi etch?
[16:38:19] <cradek> you could try the ubuntu or bdi kernels, either might work, but I doubt either is known for sure to work
[16:38:46] <cradek> I think there are instructions for building it yourself on the linuxcnc wiki
[16:39:07] <jepler> yes, there is: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Debian_Etch_Compile_RTAI
[16:39:22] <jepler> I can't vouch for their accuracy, but acemi (I think) has done a lot of work and posted his instructions
[16:41:48] <alex_joni> cradek: thanks for the commit script changes :)
[16:42:10] <cradek> alex_joni: welcome. I know you wanted that since the beginning.
[16:42:25] <alex_joni> yeah, it makes for easier reviewing commit messages
[16:42:30] <cradek> I don't know why I decided to do it at 1am last night though...
[16:42:38] <alex_joni> heh
[16:58:09] <maddash> cradek, I know about the compilation instructions, but I was hoping for precompiled images so that I wouldn't have to twiddle my thumbs for ~6 hrs waiting for make-kpkg to finish up.
[16:58:58] <cradek> maddash: like I said, there are some you could try. that's all the help I have for you, sorry
[16:59:25] <cradek> I don't like building kernels either - I avoid it whenever possible!
[16:59:28] <maddash> aren't bdi kernels emc1 only?
[16:59:52] <alex_joni> maddash: the kernel is not emc dependent
[16:59:54] <cradek> I don't know anything about bdi except some people use it to run emc2
[17:00:17] <maddash> alex_joni, so I would be abble to use the bdi w/ emc2. great.
[17:00:22] <jepler> bdi 4.50 is one of the systems we constantly test in our "compile farm". So it at least compiles emc2. http://linuxcnc.org/compile_farm/
[17:01:01] <jepler> please don't ask the bdi maintainer any emc2-specific questions, though.
[17:58:11] <skunkworks> I ran emc2 on bdi for a bit. Only issues where my lack of understanding of linux.
[17:59:30] <alex_joni> you can go straight to hell for that
[18:03:08] <skunkworks> usning bdi or lack of linux knowledge?
[18:03:18] <skunkworks> both?
[18:04:56] <alex_joni> guess you'll never find out now that you've been cured
[19:04:01] <jepler> oh man fuck you too
[19:04:47] <SWPadnos> heh
[19:08:25] <lerneaen_hydra> ey wtf=
[19:08:38] <lerneaen_hydra> another paulc handle?
[19:09:09] <SWPadnos> it's in San Antonio, TX, so probably not
[19:09:23] <lerneaen_hydra> maybe a default exit-message then
[19:10:30] <cradek> remind me not to ever get rutex stuff
[19:10:43] <SWPadnos> heh - ok :)
[19:11:09] <SWPadnos> it's cool that you can tune via software (or pots)
[19:11:20] <lerneaen_hydra> what's messed up with them, oh the person on the mailing list and windows tuning?
[19:11:23] <SWPadnos> it's not cool that you have to use their Windows-only app to do it
[19:11:48] <lerneaen_hydra> step/dir for servo control just feels strange
[19:12:14] <lerneaen_hydra> like compressing an image with a very lossy algorithm and then trying to look at the details
[19:12:15] <cradek> that's one word for it
[19:12:30] <SWPadnos> yeah. I've got geckos, but I'm considering trying a completely analog system instead
[19:25:05] <skunkworks> I agree
[20:00:23] <alex_joni> Im having a little trouble compiling the v2_1_branch, when I run 'sudo make install'
[20:00:27] <alex_joni> I get this error:
[20:00:29] <alex_joni> install: cannot stat `../lib/python/*.so': No such file or directory
[20:00:29] <alex_joni> any ideas around?
[20:02:10] <jepler> looks like it's trying to install the python stuff, but it wasn't built
[20:02:36] <jepler> probably there should be an install-python rule, and if the python stuff is built, install would depend on install-python
[20:05:19] <jepler> something like this, though I can't test it right now
[20:05:20] <jepler> http://pastebin.ca/310472
[20:05:35] <jepler> (oops, tabs got turned to spaces in the pastebin)
[20:07:11] <alex_joni> ahh .. ok
[20:08:47] <alex_joni> thx
[21:41:31] <cradek> hmm, I wanted to point Mike Cinquino to the page in the hal user manual that shows how to load the parport, but it's unclear/inconsistent
[21:42:40] <jepler> then fix it
[21:42:47] <cradek> I don't understand if cfg="" is needed or not
[21:43:02] <cradek> but I'd rather whine than look at the source
[21:43:39] <SWPadnos> cfg= is needed, AFAIK
[21:43:43] <cradek> (we've changed this several times and I don't know what it's supposed to be now)
[21:43:42] <SWPadnos> if you want more than one port
[21:44:15] <SWPadnos> I don't think it's just a number since you need the in/out spec there as well
[21:44:51] <cradek> user: hal_parport 378 in 278
[21:44:51] <cradek> realtime: insmod hal_parport.o cfg="378 in 278"
[21:44:57] <cradek> (from the source comment)
[21:45:06] <cradek> it doesn't even work as user now does it?
[21:45:40] <cradek> maybe it can - there's a main()
[21:45:53] <SWPadnos> it could, depending on root provileges (not that it's a good idea to attach any hardware)
[21:46:14] <SWPadnos> * SWPadnos is also too lazy to look at the source
[21:46:36] <jepler> we don't build the userspace binary anymore
[21:46:38] <cradek> spaces and _ are allowed ... somehow
[21:46:52] <SWPadnos> it's explicit in the source
[21:46:52] <cradek> I wonder which is preferred (and why)
[21:47:17] <SWPadnos> the underscore is allowed because kernel 2.6.mumble terminates string parameters at the first space
[21:47:39] <SWPadnos> so space and underscore are accepted as token separators
[21:48:14] <cradek> I think spaces work though
[21:48:47] <SWPadnos> they do, if the kernel version is appropriate. underscores would also help with shell quoting issues for the userspace version
[21:49:23] <cradek> -bash: lyx: command not found
[21:49:44] <SWPadnos> apt-get install lyx-qt
[21:49:52] <cradek> I'll let someone else do this one. I've answered a lot of emails lately. :-/
[21:49:56] <SWPadnos> heh
[21:50:45] <alex_joni> * alex_joni heads to bed
[21:50:47] <alex_joni> night all
[21:50:52] <cradek> night alex
[21:50:52] <SWPadnos> good night
[22:23:34] <skunkworks> skunkworks is now known as skunkworks_
[22:30:55] <skunkworks_> skunkworks_ is now known as skunkworks