Back
[01:32:19] <SkullWorks_PGAB> I am having a Linux Noob flashback moment.
[01:47:46] <owhite> jepler: could I ask you an axis question?
[01:49:50] <jmkasunich> just ask - if he's here he might answer the axis question
[01:50:08] <jmkasunich> he's not likely to answer the "may I ask a question" question
[01:50:48] <owhite> jepler: if I ask you a question about axis will you answer it? :-)
[01:51:19] <jmkasunich> if he's not here he won't answer anything - but if you ask now, and he sees it an hour from now, he'll probably answer
[01:51:44] <owhite> jepler: what I was wondering if you could point me to the area in the code that's responsible for the menu -- I couldnt dig that out of it.
[01:51:55] <jmkasunich> if he sees you asking to ask a question an hour from now, he'll probably say "well ask it already" (at least to himself)
[01:52:32] <jmkasunich> I can answer that question
[01:52:36] <jmkasunich> the answer is "yes"
[01:53:09] <owhite> hm. I searched for the occurance of the string "yes" in the code, and couldnt find that either.
[01:53:17] <SkullWorks_PGAB> ha ha
[01:53:33] <jmkasunich> your question was "could jepler point to the area in the code"
[01:53:38] <jmkasunich> the answer is "yes, he can"
[01:53:44] <owhite> oh great.
[01:54:07] <owhite> well I'll just sit here and hang by my thumbs.
[01:54:09] <jmkasunich> sorry, I'm being a bit of a smartass, and I have no good reason for it
[01:54:35] <SWPadnos> too much eggnog
[01:55:00] <SkullWorks_PGAB> who needs a reason - and you where correct JMK
[01:55:22] <jmkasunich> correct, but not helpful
[01:55:41] <jepler> owhite: in the source tree, share/axis/tcl/axis.tcl defines the axis menu bar.
[01:56:06] <owhite> jepler: thanks. just found that.
[01:56:34] <owhite> would you mind giving me general guidance on how to run another command in axis? I'd like to load it into the menu.
[01:56:59] <jepler> .menu.machine add command \
[01:56:59] <jepler> -command {exec halscope -- -ini $emcini &}
[01:56:59] <jepler> setup_menu_accel .menu.machine end [_ "Ha_l Scope"]
[01:57:34] <jepler> here's the code that puts the "Hal Scope" option in the machine menu; in tcl, "exec" starts any commandline program, and & puts it "in the background" (so that the AXIS GUI is still responsive while it's running)
[01:58:10] <owhite> thanks much. and what happens if the code that is called prints out text to stdout?
[01:58:28] <jepler> it will go to the same stdout as if axis printed it
[01:58:42] <JymmmEMC> jepler: you guys using CVS or subversion?
[01:58:50] <jepler> JymmmEMC: CVS
[01:59:12] <JymmmEMC> jepler: If you add a web (RO) iface to it, google can index it easily enough.
[01:59:25] <jepler> JymmmEMC: yes indeed.
[01:59:30] <SWPadnos> do you mean like
http://cvs.linuxcnc.org ?
[01:59:36] <jepler> in this case, I suspect owhite didn't know what to search for
[02:00:05] <JymmmEMC> jepler: Nah, google didn't return results on site:linuxcnc.org
[02:00:35] <SWPadnos> indexing a CVS repo can be dangerous - things are supposed to change from time to time :)
[02:00:42] <jepler> I don't follow you. Google says 'Results 1 - 10 of about 7,850 from linuxcnc.org. (0.13 seconds)'
[02:00:49] <jepler> result 4 is the web interface to CVS, even
[02:01:24] <jepler> (but SWPadnos is right; in fact, robots.txt partially disables indexing of cvs.linuxcnc.org because they were overloading the server it's on)
[02:01:25] <JymmmEMC> jepler:
http://www.google.com/search?num=100&hl=en&safe=off&client=firefox&rls=org.mozilla%3Aen-US%3Aunofficial&hs=Sgw&q=axis+menu+site%3Acvs.linuxcnc.org&btnG=Search
[02:02:20] <JymmmEMC> it's text, what's to overload?
[02:02:28] <SWPadnos> DSL
[02:02:31] <SWPadnos> uploads
[02:02:49] <JymmmEMC> come on isn't text... cna be done on a 14.4 modem
[02:02:55] <JymmmEMC> s/isnt/it's/
[02:03:14] <jmkasunich> there are several megs of "text" in the EMC source code tree
[02:03:20] <SWPadnos> it's aeveral megs of text, which google would want to index every day or two
[02:03:24] <jmkasunich> and many trees in CVS - all version of every file
[02:03:25] <SWPadnos> yeah, several
[02:04:00] <JymmmEMC> ok, so how big is the whole cvs tree? 10Megs?
[02:04:06] <JymmmEMC> 100Megs?
[02:04:20] <SWPadnos> DSL generally has upload rates of 256-512kbits/sc
[02:04:25] <jmkasunich> JymmmEMC: its not just the size of the tree
[02:04:45] <SWPadnos> which is 32k-64k/second, let alone the CPU load since that's actually a set of dynamically generated pages
[02:04:46] <jmkasunich> the webcvs interface has links that let you click through to EVERY version of EVERY source file in EVERY branch
[02:04:57] <jmkasunich> and a robot would blindly follow them all
[02:05:22] <jmkasunich> so two versions of a 20K file that diffet by one line, would still be uploaded twice
[02:05:27] <SWPadnos> however, note that some parts of the tree are in google's index, so it's all good
[02:05:30] <jmkasunich> some of our files have 100 or more versions
[02:05:33] <JymmmEMC> Sure, there will be an initial surge, but after that?
[02:05:49] <jepler> IIRC the problem was also CPU load. cvsweb and lxr are pretty simplistic cgi scripts, and at one time the server simply ran out of RAM when a crawler hit it with a dozen requests at once.
[02:05:50] <JymmmEMC> ok, what about limiting to the last 5 revs?
[02:05:57] <SWPadnos> note that jepler said "they were overloading ..."
[02:05:59] <JymmmEMC> jepler: ah
[02:06:10] <SWPadnos> not that they could overload
[02:06:20] <jmkasunich> JymmmEMC: a human may have a very good reason to request an ancient version
[02:06:21] <JymmmEMC> * JymmmEMC hands jepler a 486 to double cpu capacity =)
[02:06:25] <SWPadnos> heh
[02:06:36] <SWPadnos> you need a - err - 972 to do that ;)
[02:06:39] <JymmmEMC> jmkasunich: for google indexing purpose only the last 5 revs
[02:06:47] <jmkasunich> but a human isn't so stupid/rude as to go down a page with 100 links clicking on every one just to see what happens, and then doing the same on the next page
[02:06:53] <SWPadnos> I'm not sure you can limit it that way
[02:07:03] <jmkasunich> so you're suggesting we deliver a different version of the pages to google?
[02:07:23] <JymmmEMC> jmkasunich: Sure, why not robots.txt can do that for you easily enough.
[02:07:31] <jepler> robots.txt is too blunt an instrument to do that.
[02:07:42] <jmkasunich> remember - all the pages are dynamically generated content
[02:07:52] <jmkasunich> cvsweb would have to be involved, not just robots.txt
[02:08:13] <SWPadnos> heh - and don't forget the pyramid problem with the .tar.gz file available on every page ;)
[02:09:01] <SWPadnos> so each of the few dozen directories has an option to download it and everything below in a tarball :)
[02:09:13] <SWPadnos> for each version visible
[02:11:39] <jepler> even if google had an index of everything that was viewable under
http://cvs.linuxcnc.org/ I still don't think it would have helped owhite figure out which source file he needed to read or modify any better than grep on the copy of the source tree he already had would
[02:11:50] <SWPadnos> details, details :)
[02:11:52] <jmkasunich> agreed
[02:12:28] <jmkasunich> google is designed to index text, not source code - it doesn't know anything about the structure of the code
[02:12:45] <jepler> well neither does grep but it does OK
[02:12:57] <SWPadnos> ctags is good
[02:14:56] <jmkasunich> google is somewhat smarter than grep, about text
[02:15:47] <jepler> on the other hand, I have a fighting chance of writing a grep that will help me find the place that y is subtracted from x (x.*-.*y won't give too many false positives)
[02:39:11] <JymmmEMC> Hey, why not mirros the svc onto alex server, then let google index it and have an offsite backup at the same time
[02:39:32] <SWPadnos> nah
[02:39:32] <fenn> i wonder why they dont just ask for the time last modified, and only download if it's new
[02:39:58] <SWPadnos> they do, but all those pages are dynamically created when they're requested - they're always new
[02:39:59] <fenn> it wouldnt work with autogenerated pages, but at least they wouldnt download .tgz files
[02:40:10] <SWPadnos> the .tgz is also created on demand
[02:40:31] <JymmmEMC> add a simple text captcha fo rthe gz's
[02:40:50] <JymmmEMC> 42 -12 = [ ]
[02:41:17] <SWPadnos> google knows math ;)
[02:41:33] <JymmmEMC> but it's indexer want grab things from forms
[02:41:33] <fenn> google knows all
[02:41:40] <JymmmEMC> but it's indexer wont grab things from forms
[02:41:45] <fenn> have you accepted google into your heart yet?
[02:42:28] <fenn> * fenn squints to spot the one-letter difference
[02:44:07] <JymmmEMC> <form action="
http://cvs.linuxcnc.org/download/somegeneratedtarball.gz">
[02:44:07] <JymmmEMC> <input type="submit" value="Download tarball">
[02:44:07] <JymmmEMC> </form>
[02:44:30] <JymmmEMC> google will NOT follow that as it's a form, not a link.
[02:47:47] <fenn> you could exclude *.tgz in robots.txt
[02:48:15] <JymmmEMC> fenn: I suggested that, but they dismissed my idea.
[02:49:30] <SWPadnos> um - guys, there isn't a problem here. parts of cvs are indexed by google, others aren't
[02:49:39] <fenn> yay
[02:49:55] <SWPadnos> it was changed as a result of excessive robot traffic before, so there was a problem, but there isn't now ...
[02:50:02] <JymmmEMC> and is there an mirror somewhere?
[02:50:20] <SWPadnos> yes, there's a backup on the DH servers
[02:50:31] <JymmmEMC> ah
[02:50:45] <SWPadnos> and there's a VM image of the server, so anyone with vmware player can set up their own emc CVS server
[02:50:59] <SWPadnos> (actually, I think the VM has the CVS data as well)
[02:51:15] <JymmmEMC> how big is the vm image?
[02:51:38] <SWPadnos> we were quite careful to insure that the data would be as safe as at SF, or more so
[02:51:48] <SWPadnos> I'm not sure - I can check if it's important
[02:52:06] <JymmmEMC> SWPadnos: Just an estimate would be cool
[02:52:23] <cradek> 100-200 MB
[02:52:44] <JymmmEMC> cradek: for the vm image ANd cvs date combined?
[02:52:49] <JymmmEMC> data
[02:52:56] <cradek> yeah I think so (compressed)
[02:53:06] <JymmmEMC> ok, cool thanks cradek
[02:53:26] <JymmmEMC> figure around 800MB uncompressed or so
[02:53:48] <cradek> http://pastebin.ca/954783
[02:53:57] <cradek> this is from the running machine fwiw
[02:54:33] <cradek> so about 1G unpacked and running
[02:54:52] <JymmmEMC> yeah. that's not much at all.
[02:55:19] <JymmmEMC> you could toss that on a thumbstick if you ever needed to
[02:55:26] <SWPadnos> that includes the emc, rcslib, and emc2 archives as well
[02:56:59] <JymmmEMC> you gys played with subversion yet?
[02:57:01] <JymmmEMC> guys
[02:57:17] <cradek> nope
[02:57:38] <cradek> I tried doing something with git recently - it was inscrutable
[03:00:04] <SWPadnos> git has some very nice features
[03:00:16] <SWPadnos> the commands are similar to CVS, aren't they?
[03:00:44] <cradek> nope
[03:00:51] <SWPadnos> hmmm
[03:01:06] <cradek> I'm sure it's nice once one learns it
[03:01:28] <cradek> (but the web interface seems poor)
[03:01:38] <SWPadnos> the offline stuff is very nice, and the ability to make separate changesets, then commit them as a group (but still with individual comments) is nice
[03:01:51] <SWPadnos> (but most of what I know I learned from jepler)
[03:01:52] <SWPadnos> :)
[03:02:35] <cradek> offline would be terrific. changesets might be good
[03:02:58] <cradek> the whole thing about not having any real master repository, I'm not so sure about
[03:06:12] <fenn> i like how git tracks movement of code from one file to another
[04:03:03] <tomp> SWPadnos: i spent the day getting sshfs to run, all my boxes could talk to one ancient masq server, but not to each other. DOH! server had sshd others didnt!
[04:03:19] <SWPadnos> heh - oops
[04:24:32] <eric_U> integer math, you gotta love it
[04:25:32] <eric_U> thinking to myself, "why is that number always zero?"
[04:50:21] <JymmmEMC> what util can I use to image a 120gb hdd to a 250gb hdd ?
[04:50:44] <SWPadnos> ghost?
[04:50:48] <JymmmEMC> nix
[04:51:03] <eric_U> nix ghost?
[04:51:21] <SWPadnos> unix executable or unix filesystem?
[04:51:28] <SWPadnos> ghost understands ext2 and I think ext3
[04:51:36] <JymmmEMC> nix utility,
[04:51:43] <SWPadnos> http://freshmeat.net/projects/g4l/
[04:51:45] <JymmmEMC> dont have ghost available to me atm
[04:52:03] <SWPadnos> there's another one tha tI don't remember the name of
[04:52:20] <JymmmEMC> parted?
[04:52:28] <SWPadnos> partimage
[04:52:28] <eric_U> hammer
[04:52:36] <SWPadnos> http://www.partimage.org/Main_Page
[04:52:42] <JymmmEMC> partimage only does to file not to another hdd
[04:53:01] <SWPadnos> well, I bet you could use /dev/disk2 as the file ...
[04:53:18] <SWPadnos> I don't know about direct copy software
[04:57:38] <cradek> if you're mirroring a unix system you can use dump and restore
[04:57:43] <cradek> it's just like making and then restoring a backup tape
[04:58:00] <JymmmEMC> cradek: destination is twice as big
[04:58:09] <cradek> you will have to partition and then format each partition on the new disk
[04:58:13] <cradek> that doesn't matter one bit
[04:58:15] <SWPadnos> can you dump < partition1 | restore > partition2 ?
[04:58:21] <cradek> yes
[04:58:26] <SWPadnos> ok, then I'd do that ;)
[04:58:43] <cradek> mount the newdisk
[04:59:01] <cradek> cd /mnt/newdisk; dump 0af - /old/partition | restore -rvf -
[04:59:13] <JymmmEMC> dump < /dev/sdb1 | restore > /dev/sda1
[04:59:41] <cradek> that's the idea, but you have to get your command lines right, check the man pages and my example
[05:00:42] <cradek> you'll also have to install grub
[05:00:57] <cradek> to make the new disk bootable
[05:01:15] <JymmmEMC> I already have the MBR backed up to a file
[05:01:48] <cradek> I doubt it will work since the disk is different
[05:02:10] <cradek> might be easiest to boot from cd after removing the old disk, and install grub (as if something (windows install?) nuked it)
[05:03:31] <cradek> goodnight!
[05:03:36] <cradek> good luck JymmmEMC
[05:03:49] <SWPadnos> night cradek
[06:28:55] <[1]a-l-p-h-a> [1]a-l-p-h-a is now known as a-l-p-h-a
[06:29:02] <lerman_> lerman_ is now known as lerman
[06:29:21] <SkullWorks_PGAB> wow - got it all working... installed "DOSbox" and made 2 folders in my home dir - made 2 copies of Bobcad V12 Gold Edition for Dos one Inch config - the other in metric config. So I have a very suitable small CAD/CAM running on my EMC2 system.
[07:32:34] <fenn_> fenn_ is now known as fenn
[10:01:30] <steve_stallings> steve_stallings is now known as steves_logging
[13:39:21] <cradek_> cradek_ is now known as cradek
[13:42:12] <alex_joni> hmm.. I wonder what would be better.. a Q6600 or a Q9450
[15:56:41] <BigJohnT> so the Q9450 will be out in a month, what kind of motherboard would you put it on?
[15:57:10] <SWPadnos> one with the LGA775 socket
[15:57:57] <BigJohnT> I used to know all that stuff back when the 286 came out LOL
[15:58:01] <SWPadnos> heh
[15:58:13] <SWPadnos> I happen to be looking at the Intel product page for it, so it came easy ;)
[15:58:18] <SWPadnos> http://processorfinder.intel.com/details.aspx?sSpec=SLAWR
[15:58:20] <BigJohnT> LOL
[15:58:37] <BigJohnT> I'm looking at an asus board at the moment
[15:59:05] <SWPadnos> I haven't looked at motherboards in about a year, so I think I'm 100% out of date by now
[16:00:49] <BigJohnT> do they still make motherboards with parallel ports?
[16:01:27] <SWPadnos> sure. you can even get ISA slots if you're willing to pay enough
[16:05:48] <alex_joni> I was looking at this one:
http://www.intel.com/products/motherboard/DG33TL/index.htm
[16:12:36] <BigJohnT> they should play nice together if their both intel?
[16:13:15] <BigJohnT> even supports 2 parallel ports
[17:25:10] <tomp2> is Noel saying he has independant Y axis-es on the bridge, and that he lets one float while homing? ( regarding todays mail list and his entries at
http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?FiveAxisGantry )
[17:27:26] <skunkworks_> I think he is using a modified gantry kins.
[17:43:46] <jymm> ISA?! I haven't seen an ISA slot on a P4 mobo yet.
[17:44:24] <skunkworks_> Heh.
[17:48:41] <SWPadnos> http://www.ibase.com.tw/mb800.htm
[17:49:32] <tomp2> if you use sshfs and subsequently use 'fusermount -u blah', then the umount is done, but sneaky gnome(?) has left a carcass icon on your desktop ( i had loads this morning ! )
[17:49:51] <SWPadnos> http://www.l-trondirect.com/aimb-742e2-00a1.html
[17:50:35] <SWPadnos> http://news.thomasnet.com/fullstory/487906
[17:54:46] <tomp2> advantech, good for industrial stuff. CyberResearch... dunno, prices seem high, but got some hard(er) to find stuff
[18:03:50] <SWPadnos> tomp2, I've noticed the "remote directory" droppings on the desktop too
[18:04:13] <alex_joni> shitfs?
[18:04:34] <SWPadnos> interestingly, you can mount a remote file system pretty easily by using an IP address in the <mumble> dialog
[18:05:16] <SWPadnos> ah right - places -> connect to server dialog. select ssh as the service type
[18:06:04] <SWPadnos> then you get things like drag+drop for moving files around between computers
[21:08:11] <tomp2> SWPadnos: it seems the 'correct' way to disconnect is to right click the icon and choose 'disconnect', and to hell with the cmd line users
[21:11:12] <tomp2> caveat usurpo? user beware
[21:13:40] <tomp2> caveat moderator ;)
[22:20:36] <toastydeath> did anyone go to drexel university
[22:29:46] <alex_joni> heh
http://www.shamusyoung.com/twentysidedtale/?p=612
[23:19:03] <tomp2> http://en.wikipedia.org/wiki/Bored_of_the_Rings (harvard lampoon)
[23:32:53] <tomp2> dumbledoor : gandalf : gesundheit thats pretty funny stuff alex :)
[23:41:15] <tomp2> too much to read, bbl