Back
[04:30:13] <SWPadnos> SWPadnos is now known as SWP_Away
[12:07:28] <lilo> [Global Notice] Hi all. We'll be restarting services in just a moment. There will be a brief delay, and we hope to have upgraded services online pretty quickly. Thanks.
[14:23:24] <steve_stallings> steve_stallings is now known as steves_logging
[15:26:35] <cradek_> cradek_ is now known as cradek
[15:40:39] <SWP_Away> SWP_Away is now known as SWPadnos
[16:08:34] <rayh> morning tomp
[16:10:07] <tomp> hello ray
[16:11:21] <rayh> Roland said that he moved that Japax into our area, just in case we wanted it.
[16:18:49] <tomp> ok Ray, I'll try to visit early. I was working on the sink idea.
[16:19:34] <tomp> I'm working on 8255 io cards, is all emc i/o realtime? for non-crictical & slow i/o like pumps, pumpbtns... does this change driver code?
[16:19:55] <jepler> it can be either
[16:20:03] <tomp> how to decide?
[16:20:22] <jepler> I don't see any reason *not* to make it realtime
[16:20:32] <tomp> ok
[16:21:18] <jepler> in theory hal allows the same code to produce a realtime driver and a non-realtime driver -- hal_parport is an example of this, but since the Makefiles changed the non-realtime version isn't built anymore
[16:21:46] <tomp> ok, i wont try the impossible :)
[16:23:04] <tomp> if of interest, decisioncomputing.com has a 192 bit 16 timer/counter for 108$ US (88 euroids), 8255s and 8254s, pci
[16:25:34] <tomp> rayh: any word from 'mac' who specialized in JAPAX? I havent heard from for many weeks.
[16:28:20] <rayh> Don't think he was able to come to fest.
[16:29:40] <rayh> Wow. that card sounds like a lot for a little.
[16:30:26] <SWPadnos> I can't find their site - do you have a URL?
[16:31:09] <rayh> I've got several older ISA cards we could use if we had a mobo.
[16:31:56] <rayh> If new and PCI, I favor a m5i20 although we'd have to write a pure IO driver for it.
[16:32:05] <tomp> lookin... was in germany but card is thai
[16:32:24] <SWPadnos> I actually got through the build process for the m5i20 FPGA a couple of days ago
[16:32:41] <SWPadnos> up to the point of the xilinx tools complaining that my trial periodis over at least
[16:32:54] <SWPadnos> tomp, I think it's decision.com.tw
[16:33:06] <SWPadnos> or pci8255.net
[16:33:11] <tomp> http://www.decision.com.tw/our%20products-a3.htm
[16:33:22] <tomp> but the price is at... lookin
[16:34:42] <tomp> http://www.pci8255.net/ has the linux driver, still lookin for the de site
[16:35:04] <SWPadnos> I see a UK site at £118
[16:35:35] <tomp> http://www.decision-computer.de/dio/dio8255pmv.htm
[16:36:18] <tomp> yeh, less than a buck a point with the hi speed timers for the edm ...
[16:36:50] <tomp> uh, didnt er in the vat and mailing...
[16:37:01] <tomp> but cheap
[16:38:00] <tomp> SWPadnos, is there an 8255 module for the m5i20 'do anything' card?
[16:38:19] <SWPadnos> no, but there's 8255-like I/O anyway
[16:38:30] <tomp> and 8254 like?
[16:38:50] <SWPadnos> it would be trivial (once I can do anything with it) to make all 72 I/Os input or output ports
[16:39:12] <SWPadnos> and faster, since it can operate in 16 or 32-bit mode, rather than 8 bits at a time
[16:39:17] <SWPadnos> timers are also very easy
[16:39:23] <tomp> cost per point?
[16:39:24] <rayh> Don't much like the db connectors. Do they have wires and breakout cards?
[16:39:40] <SWPadnos> 72 I/O is $199 in single quantity
[16:39:54] <SWPadnos> of course, that's with the FPGA as well, so you can do a lot on the card
[16:40:31] <tomp> SWPadnos: thanks, nice stuff i didnt grok the xilinx tools yet.
[16:40:58] <SWPadnos> me either ;)
[16:41:09] <jepler> If you just want the dumb IOs you can get a better price than $199. The cheapest I know of is
http://www.futurlec.com/PCI8255.shtml and I think there's good documentation on the PCI chip it uses, the TIGER 320.
[16:41:13] <SWPadnos> you can actually put PIDs in the card
[16:41:15] <jepler> disclaimer: I don't own one
[16:42:21] <rayh> Linux and DOS driver are included!
[16:42:37] <SWPadnos> that's still an 8255 card, so all I/O is done in 8-bit chunks
[16:43:31] <tomp> rayh: I use the furturlec, but got direct from mfctr. it uses ide type cnx and i use breakouts from winford w custom cables
[16:43:55] <tomp> and linux drivers dont work the 'hal' way
[16:43:58] <rayh> That works.
[16:44:07] <SWPadnos> unless there's source ;)
[16:44:10] <rayh> No but the company is not opposed to linux
[16:44:18] <tomp> right arm
[16:48:29] <rayh> http://www.decision-computer.de/zubehoer/term.htm
[16:50:04] <tomp> http://ett.co.th/product/1002.html has the 3x8255 card, outdated pix tho, and in thai
[16:50:17] <SWPadnos> I like thai (food)
[16:50:29] <tomp> ray, whats the little black thingy in lower left?
[16:50:56] <jepler> tomp: I think a lot of futurlec's products may come from ett.co.th -- their "arm stamp" board does, for instance.
[16:51:00] <rayh> don't know.
[16:52:18] <tomp> any suggestions on the 8255 driver? i see no need for the union used in par port, and think I'll set max 8225's to 16 for now.
[16:54:21] <rayh> That ought to be enough.
[16:54:47] <tomp> ok, bye for now, thanks all.
[16:54:52] <rayh> How will we address a card with more than one of them?
[16:55:03] <rayh> fingers to slow.
[16:55:34] <SWPadnos> the 8255 driver could take multiple addresses, like parport
[16:55:39] <SWPadnos> (not sure if it does)
[17:01:02] <rayh> jepler, Do you know how tk's -underline 0 works with msgcat?
[17:05:21] <jepler> rayh: it doesn't.
[17:06:07] <rayh> Okay. I though not.
[17:06:13] <rayh> Thanks
[23:08:25] <jmkasunich> evening folks
[23:11:49] <cradek> hi john
[23:12:25] <cradek> not only does my "new" laptop agree with rtai, I also now have working accelerated opengl
[23:12:40] <jmkasunich> cool
[23:12:46] <cradek> not bad for a throwaway!
[23:12:56] <jmkasunich> and all for the stunning price of FREE!
[23:13:08] <jmkasunich> cd ..
[23:13:08] <cradek> free plus a new keyboard and new battery
[23:13:16] <cradek> well I paid some for it actually
[23:13:20] <jmkasunich> oh, forgot about that
[23:13:27] <jmkasunich> still, under $100?
[23:13:30] <cradek> I like to pay at least a token amount when I get something from work
[23:13:46] <cradek> battery is $125, waiting to make sure the keyboard works before I pay it
[23:14:04] <cradek> even with that, the whole thing will be <$200
[23:14:20] <jmkasunich> nice
[23:14:26] <jmkasunich> ok, I gotta cvs question
[23:14:40] <cradek> uh-oh
[23:14:45] <jmkasunich> I checked out using -r TESTING
[23:14:55] <jmkasunich> you committed a change to the 2.0 branch
[23:15:03] <jmkasunich> I didn't get it when I updated
[23:15:09] <cradek> that's right
[23:15:22] <jmkasunich> so what is the incantation to check out the branch?
[23:15:27] <cradek> if you want a branch checkout, you have to ask for it specifically: -r2_0_branch
[23:15:37] <cradek> if you want a branch checkout, you have to ask for it specifically: -rv2_0_branch
[23:15:40] <cradek> whatever we called it
[23:15:51] <jmkasunich> duh, because you didn't actually move testing yet
[23:16:00] <jmkasunich> eventually testing will be the head of the branch again
[23:16:10] <cradek> TESTING is not a branch, it's a fixed set of revisions (that happen to all be on a branch)
[23:16:11] <jmkasunich> and after release testing will be a spot on the trunk again
[23:16:31] <cradek> yes I think so
[23:16:42] <jmkasunich> ok, gotta change the farm checkouts
[23:17:00] <cradek> there will be a RELEASE_2_0 tag when we release
[23:17:11] <cradek> or whatever we decide to call it
[23:17:34] <cradek> there are several (4-5) fixes in the branch since TESTING
[23:17:35] <jmkasunich> I think we want the farm building the head of the branch
[23:17:47] <jmkasunich> (and the head of the trunk of course)
[23:17:49] <cradek> I agree
[23:18:22] <jmkasunich> ok, to change the testing checkout to a branch checkout, I think I can just do
[23:18:44] <jmkasunich> cvs up -r v2_0_branch
[23:19:40] <jmkasunich> then cd .. ; mv emc2testing emcv2branch
[23:19:58] <jmkasunich> change the list of trees to build (trivial) and change index.shtml
[23:20:12] <jmkasunich> zat sound right to you?
[23:20:25] <cradek> yep
[23:21:00] <jmkasunich> I have 5 ssh terminals on my screen, one monitoring each farm slot
[23:21:12] <jmkasunich> beats the keck out of standing in front of the rack
[23:21:30] <jmkasunich> I can't believe it took me so long to catch on to ssh
[23:22:05] <cradek> * cradek tries to not make a wisecrack about "old dogs"
[23:22:36] <jmkasunich> go right ahead, I know darn well what I am
[23:22:38] <jmkasunich> ;-)
[23:23:17] <cradek> nah, it's not nice
[23:23:55] <cradek> fred sent that driver box overnight tuesday, but it still hasn't come... I hope it wasn't lost
[23:24:13] <jmkasunich> he didn't send a tracking number?
[23:24:18] <cradek> nope
[23:25:40] <cradek> * cradek 's eyes bleed on keystick.cc
[23:25:55] <jmkasunich> the code or the on-screen appearance?
[23:26:00] <cradek> the code
[23:26:03] <jmkasunich> (never looked at either one)
[23:26:08] <cradek> it looks great running
[23:39:30] <jmkasunich> did you get an invoice from the CNC workshop?
[23:39:54] <cradek> yes I did yesterday I think it was
[23:39:59] <jmkasunich> ok
[23:40:13] <jmkasunich> I called Roland yesterday morning, he said he was working on it
[23:40:17] <jmkasunich> got one later
[23:40:30] <jmkasunich> I wonder if he really was working on it, or if he needed me to remind him
[23:41:06] <jmkasunich> oops, time for dinner
[23:41:15] <cradek> good question
[23:41:27] <cradek> I still haven't paid though - maybe I"ll go dot hat
[23:46:40] <jepler> I should do that too
[23:49:30] <cradek> it's official, I'm going