#emc-devel | Logs for 2006-10-11

[00:09:05] <A-L-P-H-A> apparently the G203 (new stepper driver) is coming out soon... ETA... from Mariss 'when it's done... should be in a month'. :)
[06:35:49] <alex_joni> jepler: I hope to be available during the afternoon today (around your morning coffee time)
[12:20:54] <alex_joni> morning all
[12:21:03] <A-L-P-H-A> good morning Alexandru.
[12:21:10] <alex_joni> * alex_joni is finally not that busy
[12:21:45] <anonimasu> morning alex
[12:23:53] <alex_joni> quite some crazy days lately :/
[12:28:46] <jepler> hi alex
[12:29:31] <jepler> did you take a look at improved-dpkg branch?
[12:31:11] <alex_joni> jepler: depends on your definition of look
[12:31:15] <jepler> my concerns are: did I get the right files in -common? what will happen to people upgrading, now that there's no package named 'emc2' anymore? what will happen to people who don't upgrade?
[12:31:23] <jepler> er, who don't want to upgrade
[12:31:26] <alex_joni> I actually read the commits, and looked at the files in cvs using the webinterface
[12:31:32] <alex_joni> but I never checked it out to try them
[12:31:34] <alex_joni> :-/
[12:31:54] <jepler> are you comfortable with the approach -- to generate the debian/ files with a shell script that is separate from src/configure?
[12:32:03] <alex_joni> somehow
[12:32:31] <alex_joni> I mean we have 2 systems which we are supporting currently (2.6.12-magma and 2.6.15-magma)
[12:32:46] <jepler> for 2.1 we'll add -sim and maybe whatever is in edgy
[12:32:57] <alex_joni> some special tweaks will be necessary for both of them
[12:33:14] <alex_joni> special tweaks not existing in CVS; but beeing distributed with the source packages
[12:33:37] <jepler> I have to admit I don't understand why it's being done that way
[12:33:40] <alex_joni> so there is still some work to build the packages
[12:33:56] <jepler> is that this 'extra' directory?
[12:33:57] <alex_joni> well, the main reason is to have the default debian/ build debian packages
[12:34:07] <alex_joni> jepler: from what I remember the extra
[12:34:26] <alex_joni> it's actually a decision which hasn't been really tackled
[12:34:46] <alex_joni> option 1. have the debian/ stuff relate _only_ to our ubuntu settings (breezy & dapper so far)
[12:35:13] <alex_joni> option 2. have the debian/ stuff be generic as much as possible, and have a bit manual trouble to build the ubuntu packages
[12:35:31] <jepler> what are these extra files? Do they harm a debian system?
[12:35:42] <alex_joni> jepler: the extra files probably not
[12:35:51] <alex_joni> some desktop shortcuts & images & the like
[12:36:12] <alex_joni> I would have to look inside a sourcepackage to remember what it was
[12:37:04] <alex_joni> also the udev & parport fixes
[12:37:12] <jepler> emc2.desktop emc2-doc.desktop emc2icon.png cnc.directory udev/.../emc2.rules modprobe.d/emc2 cnc.menu
[12:38:20] <jepler> I would suggest this: arrange for configure to be told the distribution type, or get it from /etc/lsb-release if unspecified
[12:38:51] <jepler> then it selects the right extras/ directory, such as extras-Ubuntu or extras-Ubuntu-6.06 and substitutes it into debian/rules
[12:38:57] <jepler> now we can put all that stuff in CVS where it belongs
[12:39:56] <jepler> and append those filenames to xxx.files when generating them
[12:41:23] <jepler> but it's more complicated than that, because we don't want udev/rules.d/emc2.rules installed for a -sim configuration
[12:41:50] <alex_joni> right
[12:43:46] <jepler> find_extras () {
[12:43:46] <jepler> for candidate in extras-$KERNEL_VERSION-$DISTRIB_ID-$DISTRIB_RELEASE \
[12:43:46] <jepler> extras-$DISTRIB_ID-$DISTRIB_RELEASE extras-$DISTRIB_ID; do
[12:43:46] <jepler> if [ -d $candidate ]; then echo $candidate; return 0; fi
[12:43:46] <jepler> done
[12:43:48] <jepler> return 1
[12:44:16] <jepler> so if we have a "special case" for sim-Ubuntu-6.06 we can accomodate that
[12:44:21] <alex_joni> ahaa
[12:44:23] <alex_joni> cool
[12:44:36] <jepler> we'd just add extras directories when we discovered we needed them
[12:44:39] <jepler> bbl, it is coffee time as you predicted
[12:44:47] <alex_joni> ;-)
[13:02:22] <alex_joni> jepler: say when java time is over :)
[13:11:33] <jepler> hmm -- 'find' fails if you add directories while it is running
[13:11:58] <alex_joni> jepler: can you look at something?
[13:12:03] <jepler> alex_joni: sure
[13:12:31] <alex_joni> www.robcon.ro/emc/debian.tgz
[13:13:02] <jepler> what am I looking at specifically?
[13:13:13] <alex_joni> basicly having control and control.modules
[13:13:22] <alex_joni> the later is handled by modass
[13:14:14] <jepler> you mean "module_assistant"? There seem to be two problems with that approach: first, the lack of good documentation; second, the fact that we need to build not only modules that are kernel-dependent, but also some files (module-helper, for instance)
[13:14:20] <jepler> and third, we need to build the -sim version
[13:14:27] <jepler> three problems.
[13:15:40] <jepler> plus it's not a solution for the extras/ problem
[13:15:42] <jepler> four problems.
[13:16:01] <cradek> and a fanatical devotion to the pope
[13:17:49] <alex_joni> pope?
[13:18:03] <anonimasu> cradek: that's a major problem.
[13:18:16] <cradek> it's a joke
[13:18:19] <cradek> I didn't mean to derail
[13:18:24] <alex_joni> jepler: module-helper can't be built by module-assistant?
[13:18:28] <anonimasu> I'm just kidding ^_^
[13:18:39] <cradek> jepler: I'm glad you're reexamining this packaging
[13:18:55] <jepler> alex_joni: I have no clue -- refer to the first of the two^H^H^Hfour problems again, please
[13:19:06] <alex_joni> problem #3 is not related to the module-assistant
[13:19:08] <jepler> meanwhile, to explain the joke: http://www.jumpstation.ca/recroom/comedy/python/spanish.html
[13:19:21] <alex_joni> it can be in the regular user-space package
[13:21:29] <alex_joni> jepler: but it was only a thing I stumbled upon recently..
[13:21:45] <alex_joni> I kinda agree module-assistant has no documentation which is useable
[13:21:49] <jepler> sorry to run, but work is calling in an important way
[13:22:38] <jepler> if you want to work on module_assistant please do, but make sure that the result builds a -sim package and that any user can rebuild identical packages to the ones we would provide, with well-documented and simple steps
[13:22:53] <jepler> it's hard to beat 'apt-get -b source', and you won't get that with module_assistant
[13:23:00] <alex_joni> jepler: right
[13:23:07] <alex_joni> I'll investigate further :)
[13:24:02] <alex_joni> although I think 'apt-get -b source' & 'apt-get -b source.modules' would be the needed thing with ma
[13:24:26] <jepler> I don't think you're right.
[13:24:48] <cradek> can someone briefly tell me what problems we're trying to solve?
[13:25:33] <cradek> is it that there are separate emc2 packages for the different kernel/rtai versions?
[13:25:57] <jepler> OK, work problem seems to be solved
[13:26:59] <jepler> cradek: right now the debian/ files in HEAD only work for one combiation: 5.10, 2.6.12-magma. and even then, some files that are not in CVS are used to create the official package
[13:27:03] <alex_joni> heh, that was quick
[13:27:23] <cradek> ok
[13:27:25] <jepler> for 2.1, I envison that we might build as many as 6 ubuntu packages, but probably fewer: 5.10, 6.06, and edgy; realtime and sim
[13:27:47] <cradek> also we have to have separate repositories for the different ubuntu releases, would it fix that?
[13:27:51] <jepler> more realistically, we'd build 5.10 RT, 6.06 RT and sim, and maybe ignore edgy, but that's still more than now
[13:28:20] <cradek> currently I boot 5.10 to build the 5.10 package etc. would that still be necessary?
[13:29:40] <jepler> I think the 6.06 and 5.10 versions would still be separate builds, because the versions of libraries in /usr/lib might be different
[13:30:20] <cradek> as for the missing files, we could always just check them in...? a directory per version?
[13:30:39] <cradek> (I don't remember if they're different across versions)
[13:31:35] <jepler> just guessing based on the extras/ directory from the breezy package, it looks like they'll be the same for different versions of ubuntu
[13:31:51] <jepler> but for instance there's no use including modprobe.d and udev/rules.d in the sim package, and it might do harm
[13:32:02] <jepler> ("my printer doesn't work after I installed the simulator version of emc2")
[13:32:06] <cradek> very true
[13:32:14] <alex_joni> right
[13:32:22] <alex_joni> sorry.. had a workproblem to fix too :)
[13:32:29] <alex_joni> but I agree so far
[13:32:30] <cradek> the sim will be short a lot of files
[13:32:48] <alex_joni> jepler: you might be right about m-a
[13:33:04] <cradek> I don't really like that the user will have to upgrade two packages
[13:33:17] <cradek> (and could get bogus things by upgrading only one of them)
[13:33:20] <jepler> emc2-sim and emc2-common?
[13:33:29] <cradek> emc2 and emc2-kernelversion-modules
[13:33:33] <jepler> there's some way to tell the packaging system "the version has to match exactly"
[13:33:43] <cradek> ok
[13:33:53] <jepler> I think my improved-dpkg doesn't do that yet, though
[13:34:23] <jepler> I mean, I know it doesn't
[13:35:16] <jepler> when does ubuntu breezy support end?
[13:36:15] <jepler> interesting -- "
[13:36:16] <jepler> In contrast to previous releases, Canonical has announced that it will not provide free CDs of Edgy Eft through its popular ShipIt programme which will continue to ship Dapper Drake."
[13:37:15] <alex_joni> jepler: breezy will continue about 12-18 months I think
[13:38:32] <cradek> I haven't found it yet
[13:44:22] <cradek> I'm guessing april 2007 (I think the pattern is non-LTS are 18 months)
[13:45:34] <alex_joni> Ubuntu releases are supported for at least 18 months after release. Ubuntu LTS (Long Term Support) releases will be supported for 3 years on the desktop, and 5 years on the server.
[13:46:01] <jepler> what's that mean? They stop supporting gnome but keep supporting apache?
[13:46:56] <cradek> I can't find a clear statement of the support times, but I do find stuff like "do you think a computer CPU is more or less efficient than a heater at generating heat?" on their wiki
[13:47:08] <alex_joni> lol
[13:47:32] <alex_joni> https://wiki.ubuntu.com/Releases <- that's what I found
[13:47:55] <alex_joni> cradek: according to that EOL for Breezy is april 2007
[13:48:42] <cradek> yeah I don't know what "on the server" means either
[13:48:48] <alex_joni> jepler: for our kind of users I think edgy is stay-away
[13:49:11] <alex_joni> cradek: the ubuntu-server install
[13:49:24] <alex_joni> it uses slightly different packages I think
[13:50:03] <alex_joni> err.. no
[13:53:33] <jepler> so for emc2.1 we want to support: breezy/rt, dapper/sim, dapper/rt? skip edgy and breezy/sim?
[13:54:46] <alex_joni> I'd skip edgy for now
[13:54:53] <alex_joni> not sure about breezy sim
[13:55:11] <alex_joni> because breezy is somehow more forgiving on older HW
[13:57:38] <jepler> but the use I see for sim is newish desktop machines, not the old machine out in the garage with the mill
[13:58:22] <cradek> you think sim will run on most/all debian installs?
[13:58:42] <jepler> I run it on an fc4 machine
[13:58:51] <jepler> and redhat9
[13:58:59] <cradek> I mean the package you're talking about building
[13:59:04] <jepler> oh -- nfc
[13:59:11] <A-L-P-H-A> anyone besides me think datsumer (sp) is odd?
[13:59:54] <cradek> A-L-P-H-A: it might just be a language obstacle, not sure
[14:00:44] <A-L-P-H-A> he's using adelphi
[14:00:48] <A-L-P-H-A> where's adelphi?
[14:00:55] <A-L-P-H-A> I thought that was a state provider
[14:01:00] <cradek> oh
[14:01:01] <cradek> no idea
[14:01:20] <alex_joni> A-L-P-H-A: traceroute if you really care
[14:01:34] <A-L-P-H-A> alex_joni: hmmm..
[14:01:37] <alex_joni> jepler: it's quite possible the deb runs on any deb
[14:02:11] <alex_joni> any debian system even
[14:04:30] <alex_joni> A-L-P-H-A: looks related to the lackawanna county to me
[14:04:43] <alex_joni> NE Pennsylvenia
[14:05:32] <A-L-P-H-A> I don't have traceroute installed...
[14:05:37] <A-L-P-H-A> nor do I want more stuff installed.
[14:05:42] <A-L-P-H-A> damn... I'm starting to crash now.
[14:05:57] <A-L-P-H-A> as in my sleep is catching up with me.
[14:06:20] <jepler> cradek: what should the driver for the digital I/O pins of a serial port be called?
[14:06:48] <jepler> right now I called it hal_8250 and the pins have names like 8250.0.cts-in but I am not happy with that
[14:07:14] <cradek> it's done??
[14:07:50] <jepler> well, it's not *tested*
[14:08:14] <cradek> I wonder if the pin numbers should be pin-03-in like the parport (assuming 9 pin serial I guess)?
[14:08:31] <SWPadnos> how about hal_serport? (like hal+parport, but not hal_serial since it doesn't use serial data)
[14:08:36] <cradek> does anyone have db25 serial ports anymore?
[14:08:48] <jepler> I dunno, but the concern about db25 vs db9 was what made me use names, not numbers
[14:08:49] <SWPadnos> I do, but I'm not sure I can make that computer work any more
[14:09:00] <jepler> SWPadnos: that's not a bad name
[14:09:07] <cradek> well that was kind of rhetorical, I know they're out there
[14:09:11] <skunkworks> hmm - the issue is if someone uses and 9 to 25 adaptor
[14:09:14] <skunkworks> alos
[14:09:15] <skunkworks> also
[14:09:39] <skunkworks> could you designate what plug you have when calling it :)
[14:09:51] <SWPadnos> I'd use the 9-pin numbers
[14:09:53] <cradek> ideally, as few people as possible should have to look up a pinout
[14:09:55] <skunkworks> or you name the actualy pins
[14:10:04] <skunkworks> request to send and such
[14:10:07] <skunkworks> rtd
[14:10:10] <skunkworks> rts
[14:10:13] <jepler> skunkworks: that's what I did: cts-in dsr-out etc
[14:10:17] <skunkworks> nice
[14:10:27] <cradek> I think I like hal_serport
[14:10:27] <alex_joni> harder for the user to find them
[14:10:45] <alex_joni> cradek: how about hal_serio ?
[14:10:47] <cradek> and I'm in the 9-pin numbers camp
[14:10:55] <alex_joni> same here.. 9-pin :)
[14:11:10] <skunkworks> all in favor :)
[14:11:18] <cradek> I like how serport is "parallel to" parport
[14:11:33] <cradek> to me, serio means the IO is serial
[14:11:37] <alex_joni> yeah, but it's not treated as a port
[14:11:44] <jepler> pin-9 or pin-09? They only go to 9
[14:11:47] <alex_joni> hal_serport_io ?
[14:11:58] <alex_joni> jepler: I'd stick with 09 for consistency with other drivers
[14:12:11] <cradek> I have no opinion on 9 vs 09
[14:12:44] <A-L-P-H-A> no one uses db25 serials anymore... as it's not common anymore. hasn't been since P2.
[14:12:48] <A-L-P-H-A> Pentium 2s.
[14:12:56] <alex_joni> A-L-P-H-A: care for some cables?
[14:13:00] <alex_joni> * alex_joni has plenty
[14:13:10] <A-L-P-H-A> alex_joni: serial cables? Nah...
[14:13:17] <alex_joni> A-L-P-H-A: the ones you used to plug into the motherboard
[14:13:34] <alex_joni> with a DB25 connector screwed to a metal plate
[14:13:42] <A-L-P-H-A> alex_joni: only serial device was/is an interface to my satellite reciever to get free satellite.
[14:13:46] <skunkworks> alex_joniL I have a box of those.
[14:13:59] <SWPadnos> hey - I still have a machine that has an AT keyboard connector, and a serial mouse connected via an adapter to the 25-pin serial port ;)
[14:14:06] <skunkworks> mostly from pcchips mother boards
[14:14:26] <cradek> SWPadnos: in everyday use?
[14:14:39] <SWPadnos> well, I'm hoping I can fix it ;)
[14:14:50] <SWPadnos> then it will be in everyday use as a file server
[14:14:53] <A-L-P-H-A> I have a box of old computer toys too... like floppy drive cables, 3.5" floppy drive... ummm... ide cables (as opposed to EIDE 66/100 cables)...
[14:15:06] <A-L-P-H-A> parallel port stuff from XT days... and other junk
[14:15:30] <SWPadnos> but I wouldn't run emc on it (though it probably could - it's a dual PPro-overdrive 333MHz)
[14:15:44] <alex_joni> oh.. I just found something looking through cables :D
[14:15:56] <A-L-P-H-A> alex_joni: old porn mags?
[14:16:17] <skunkworks> overdrive proccessors - now that brings back memories.
[14:16:29] <SWPadnos> heh
[14:16:38] <alex_joni> nah.. 2C87
[14:16:39] <SWPadnos> Pentium Pro, too ;)
[14:16:44] <alex_joni> 10 MHz
[14:16:55] <alex_joni> wonder who knows what it is :)
[14:17:03] <A-L-P-H-A> anyone remember to turn turbo on/off you hit alt-ctrl- + and alt-ctrl- - to toggle on/off turbo mode on a CPU.
[14:17:20] <A-L-P-H-A> atmel 90s2313-10mhz! or a 286.
[14:17:28] <alex_joni> 287 actually :)
[14:17:29] <cradek> no, my machines always had a turbo button
[14:17:33] <alex_joni> co-processor
[14:17:35] <A-L-P-H-A> hahahaha
[14:17:43] <alex_joni> was about 1000$ at it's prime-time
[14:17:58] <A-L-P-H-A> cradek: yeah... but if you used the keybaord, you didn't need to reach to hit the button.
[14:18:03] <alex_joni> manufactured by IIT ? wonder if they still exist
[14:18:24] <skunkworks> yah - we had some 386's that had soft turbo buttons. One you could get into the bios at any time doing a 3 key combo
[14:18:26] <A-L-P-H-A> Thompson was another CPU manufacturer... like Intel, AMD, and VIA.
[14:19:08] <jepler> OK -- i'll call it serport, use pin-N notation with 9-pin numbers
[14:19:17] <SWPadnos> heh - good plan :)
[14:19:21] <A-L-P-H-A> oh... they're not called ST, not thomson anymore
[14:19:59] <alex_joni> http://www.cpu-world.com/CPUs/80287/L_IIT-2C87-20.jpg
[14:20:18] <alex_joni> only mine is -10 not -20 :)
[14:20:39] <alex_joni> but still from 1988 :D
[14:21:11] <A-L-P-H-A> you got the ghetto one. :D
[14:21:58] <alex_joni> A-L-P-H-A: this one : http://www.cpu-museum.com/Bilder/IIT-2C87-10_1.jpg
[14:23:09] <alex_joni> http://cpu.in.ua/chips/fpu_80287_iit_2c87-10.jpg
[14:24:21] <alex_joni> running home.. bbl
[14:24:58] <jepler> http://pastebin.ca/197745
[14:26:38] <cradek> jepler: *cool*
[14:27:04] <cradek> does it do any uart detection/verification?
[14:27:07] <jepler> no
[14:27:26] <cradek> how do you specify the port?
[14:27:39] <jepler> loadrt serport io=0x3f8,0x2f8
[14:28:10] <cradek> I wonder what else one has to do to the machine to make it safe to run
[14:28:21] <cradek> obviously turn off gettys etc
[14:28:42] <cradek> does it do the request-region stuff?
[14:28:57] <cradek> might have to disable the serial module
[14:29:12] <jepler> yes yes
[14:29:45] <SWPadnos> interestingly wnough, the serial port can provide nicely debounced inputs. if the changed bit is off, then the input was stable since the last read
[14:29:50] <SWPadnos> s/wnough/enough/
[14:30:26] <jepler> http://pastebin.ca/197752
[14:30:40] <SWPadnos> also, for encoders, you can see if there was noise or a missed count, because if the changed bit is on, but the bit is in the same state, then something weird happened
[14:30:54] <jepler> hm that manpage shouldn't show 'count=N'
[14:31:02] <jepler> no way to fix that in an autogenerated manpage
[14:31:06] <jepler> * jepler sighs
[14:31:37] <jepler> well I'll fix that up sometime
[14:32:14] <cradek> it says all the outs are ins and the ins are outs?
[14:32:45] <cradek> brb
[14:32:47] <SWPadnos> RI is a HAL output, since the physical pin is an input
[14:33:16] <jepler> cradek: yeah I know -- sucks, doesn't it
[15:05:11] <alex_joni> back
[15:28:43] <jepler> * jepler ponders direct support for hardware drivers with io= in comp
[15:29:11] <jepler> the EXTRA_SETUP and EXTRA_CLEANUP will be nearly identical in each hardware driver, except for the number of bytes requested
[15:29:29] <jepler> I guess that assumes that each one registers a single contiguous range of ports..
[15:29:53] <SWPadnos> unless you allow for multiple ranges (named, perhaps)
[15:31:35] <SWPadnos> hmmm - actually, for hardware you have to allow for multiple ranges, since even the parport/serport driver may want to register multiple pieces of hardware at discontiguous addresses
[15:31:47] <SWPadnos> bbiab
[15:32:27] <jepler> the case I'm worried about is where the EPP parport uses some addresses at base and at base+400
[15:32:52] <jepler> the EXTRA_SETUP and EXTRA_CLEANUP in serport.comp accomodates one range per instance, not two
[15:33:00] <jepler> er, base+0x400
[15:35:34] <jepler> I guess I mean ECP anyway
[15:36:02] <cradek> EASDFP
[15:37:57] <jepler> of course, good support for hardware-drivers would also make automatic support of -invert and -in-not desirable
[15:40:16] <alex_joni> http://pastebin.ca/197838
[15:40:17] <jepler> should setting serport.0.pin-X-out and -invert to FALSE put +V or -V on the pin?
[15:40:21] <alex_joni> minimal ini file
[15:40:38] <cradek> jepler: very good question
[15:40:49] <alex_joni> * alex_joni would expect -V intuitively
[15:40:58] <cradek> yes but the rs232 way is +V
[15:41:07] <alex_joni> I think RS232 defines low as +V
[15:41:10] <alex_joni> right
[15:41:17] <jepler> yes, a level-shifter chip like max232 will output 0V if the input is +V and 5V if the input is -V
[15:41:26] <cradek> arg
[15:41:54] <alex_joni> cradek: seen that pastebin?
[15:42:22] <alex_joni> I think it can be further reduced ..
[15:44:34] <cradek> I love it
[15:44:46] <alex_joni> cradek: really?
[15:45:00] <cradek> is default acceleration used for anything?
[15:45:09] <alex_joni> hmm.. probably not
[15:45:52] <alex_joni> default & max accel from TRAJ can be missing too
[15:52:33] <alex_joni> http://pastebin.ca/197856 <- this works too
[15:52:54] <alex_joni> but it's with tkemc (axis assumes some fields present: like NC folder)
[16:04:53] <SWPadnos> if we get rid of the azxis scale offsets, it can be reduced even further :)
[16:04:58] <SWPadnos> axis
[16:06:44] <alex_joni> cradek: would it make sense to add that to CVS ?
[16:06:54] <alex_joni> sim/minimalistic.ini or the like ?
[16:07:37] <SWPadnos> I'd leave the comments in, I think
[16:07:53] <SWPadnos> but maybe only for the items present, not for all the other stuff that could be in there
[17:11:30] <cradek> Set Break. When set to "1", the transmitter begins to transmit continuous Spacing until this bit is set to "0". This overrides any bits of characters that are being transmitted.
[17:11:37] <cradek> wooo
[17:11:51] <cradek> * cradek just found another output
[17:11:56] <SWPadnos> heh
[17:11:57] <jepler> 6 Send Break -- Sends SPACE until this bit is reset again
[17:12:03] <jepler> this one?
[17:12:10] <cradek> yep
[17:12:11] <SWPadnos> can you set it to MARK as well?
[17:12:17] <cradek> that's the default state
[17:12:35] <SWPadnos> ah, right. unless someone writes to the TX register :)
[17:12:40] <jepler> yes
[17:12:53] <jepler> what pin# is that? 3?
[17:12:59] <cradek> yes
[17:13:05] <SWPadnos> 2 or 3, depending on the connector type
[17:13:07] <cradek> I think it's 3 on DTE
[17:13:14] <SWPadnos> don't remember which is which though
[17:13:22] <jepler> I'm doing everything from the PC's point of view
[17:13:34] <cradek> no I think it's 2
[17:13:36] <SWPadnos> depending on 9 or 25 pin connectors
[17:13:56] <cradek> oh do they swap? ffs
[17:14:00] <SWPadnos> yep
[17:14:08] <cradek> my BB says TD is 2 DCE on DB25
[17:14:15] <SWPadnos> one of the other confusing things about serial :)
[17:14:16] <cradek> so it must be 3
[17:14:39] <SWPadnos> http://www.aggsoft.com/rs232-pinout-cable/pinout-and-signal.htm
[17:15:31] <jepler> 4 Break received BREAK was received
[17:16:01] <jepler> it seems a lot more dicey to try to receive a bit on RX using this
[17:16:07] <cradek> yeah
[17:16:15] <cradek> it may take a long time to realize it's a break
[17:16:19] <cradek> I'm not sure what the algorithm is
[17:16:32] <cradek> it would obviously be at least a character's time
[17:16:36] <SWPadnos> should be one byte + 1 bit, I think (including start, stop, and parity bits)
[17:17:06] <cradek> The Break signal must be of a duration longer than the time it takes to send a complete byte plus Start, Stop and Parity bits. Most UARTs can distinguish between a Framing Error and a Break, but if the UART cannot do this, the Framing Error detection can be used to identify Breaks.
[17:17:08] <SWPadnos> actually, the first stop bit that is 0 should be it
[17:17:54] <SWPadnos> there may also be a delay in the status bits, since the serial chips are usually clocked from some multiple of the serial clock
[17:18:07] <SWPadnos> ie, the bits may be slower if the port is set to 110 baud
[17:18:18] <cradek> it might be ok for a pushbutton, but it's a dodgy use
[17:18:19] <SWPadnos> status bits, that is
[17:18:39] <jepler> you'd set the baud rate as high as possible
[17:19:05] <SWPadnos> right - I'm not sure if the status bits are dependent on the clock, but in case they are, it's best to do that
[17:19:23] <SWPadnos> of course, 115200 baud is faster than the fastest thread we're likely to see out of emc any time soon
[17:19:33] <SWPadnos> (and some chips go up to 921.6k or higher)
[17:20:35] <cradek> do you know offhand the highest programmable rate on the regular 8250?
[17:20:43] <SWPadnos> 115200 baud
[17:21:23] <SWPadnos> 16C450 or 550 chips can usually go to 230.4k, and the higher end chips can go to 460.8k or 921.6k
[17:22:43] <SWPadnos> hmmm - I may be wrong about that. the old 8250 may only go to 57600 or 19200
[17:23:34] <cradek> I'm trying to get google to tell me...
[17:25:42] <skunkworks> what are these serial port things anyways?
[17:26:47] <skunkworks> What else is there for i/o... Cradek: have you eeked out every bit of i/o on your computer? :)
[17:27:54] <cradek> SWPadnos: 0x1c200/divisor (115200)
[17:28:13] <SWPadnos> ok
[17:29:15] <cradek> skunkworks: the three keyboard lights
[17:29:56] <jepler> cradek: ??
[17:30:03] <cradek> and obviously, the speaker
[17:30:11] <cradek> hsync and vsync from the vga
[17:30:26] <cradek> hey three D-A converters built in!
[17:30:58] <jepler> cradek: believe me, I've thought about that
[17:31:15] <skunkworks> :) 3 axis machine
[17:31:54] <cradek> maybe you could latch into external storage the pixel value of the first scanline or something
[17:31:58] <cradek> very funny
[17:32:12] <cradek> is a 75Hz servo update fast enough?
[17:53:18] <jepler> cradek: shall we test the RX pin as digital output or shall I just check it in?
[17:53:57] <cradek> do you mean TX? or input?
[17:54:02] <jepler> TX
[17:54:04] <jepler> I meant TX
[17:54:31] <cradek> sure we can easily try it
[17:55:01] <jepler> using TX also adds about 1uS to the typical serport.0.write.time because it writes another port
[17:55:10] <jepler> er, 1000 cycles
[17:55:29] <jepler> 1000-1200 typical before, 2000 typical after
[17:55:42] <cradek> can you tell whether a (hal) pin is connected?
[17:55:51] <jepler> not really
[17:56:03] <cradek> I guess it would break halcmd
[17:56:08] <jepler> I won't worry about it
[17:56:14] <cradek> yeah
[17:58:03] <Lerneaen_Hydra> serial port interface? cool
[17:58:22] <jepler> Lerneaen_Hydra: this is to use several of the pins as digital I/O
[17:58:40] <Lerneaen_Hydra> so it's being used paralelly?
[17:58:54] <Lerneaen_Hydra> ie. no serial -> paralell circuit needed?
[17:59:21] <jepler> Lerneaen_Hydra: if I understand what you mean, yes.
[17:59:45] <jepler> Lerneaen_Hydra: for instance, the value on the "DTR" pin can be directly set by software, and tha value on the "RI" pin can be directly read by software
[17:59:57] <jepler> that's what this new "serport" component does
[18:00:04] <Lerneaen_Hydra> nice
[18:00:06] <jepler> it will have 4 inputs and 3 outputs
[18:00:11] <Lerneaen_Hydra> hey not bad
[18:00:36] <Lerneaen_Hydra> enough for a jogwheel and spindle on/off
[18:01:42] <jepler> cradek: I think this is the rotary switch I have: http://www.mouser.com/search/productdetail.aspx?R=318-ENC111-20PSvirtualkey14860000virtualkey318-ENC111-20PS
[18:14:46] <jepler> looks like there are a variety of LEDs designed for 2mA or 10mA current
[18:15:03] <jepler> I wonder what current you can get out of a typical serial port
[18:15:58] <cradek> it would not surprise me if the original PC async cards (which could also be configured for 20ma current loop) could give 20mA
[18:16:23] <cradek> the whole point of rs232 is to drive a long noisy communications line
[18:17:00] <cradek> the DS1488 line driver gives current limited +- 10mA
[18:18:05] <cradek> no idea if anything, especially laptops, are that strong today
[18:18:05] <jepler> so a 10mA LED might be an appropriate choice
[18:18:18] <jepler> oh, DS1488 is old?
[18:18:36] <cradek> yeah it's an external driver, I bet most chipsets have them built into the uart now
[18:18:52] <cradek> maybe check the max232?
[18:19:06] <cradek> if the 1488 and the max232 can both do 10mA I'd say you can nearly count on it
[18:19:43] <jepler> max232 output short-circuit current min +-7 typ +-22
[18:20:15] <cradek> nice
[18:20:21] <cradek> so I bet you can (nearly) count on 10
[18:23:13] <Lerneaen_Hydra> so an optocoupler without any problems then
[19:04:44] <jepler> too bad that encoder with button sticks up
[19:05:35] <cradek> ?
[19:56:18] <jepler> forget it
[21:12:34] <alex_joni> hi
[21:26:06] <alex_joni> hmm.. guess no-one else around :)
[21:26:14] <alex_joni> * alex_joni will be leaving tomorrow
[21:26:29] <alex_joni> vienna for the weekend, catch you guys back on sunday I presume
[21:27:25] <skunkworks> have fun
[21:27:28] <A-L-P-H-A> alex_joni: ... hotties?
[21:27:37] <A-L-P-H-A> some hot italian girl I hope...
[21:27:46] <A-L-P-H-A> wait
[21:27:50] <A-L-P-H-A> that's austria
[21:27:55] <A-L-P-H-A> Vienna Austria?
[21:28:05] <A-L-P-H-A> I'm Euro stupid.
[21:28:09] <SWPadnos> Vienna Sausages
[21:28:16] <SWPadnos> Vienna Boys Chior
[21:28:18] <SWPadnos> Choir
[21:28:19] <A-L-P-H-A> Hamburg Sausages.
[21:28:28] <SWPadnos> no, that's Hamburgers
[21:28:29] <A-L-P-H-A> bbiab... steaks.
[21:28:32] <SWPadnos> seeya
[21:28:36] <SWPadnos> have fun Alex
[21:35:09] <alex_joni> thx, I probably will :)
[21:36:04] <jepler> see you alex_joni