jimfleig: not in the more recent machines, yours probably requies SDRAM memory which most likely does not need to be added in pairs
you need to make sure it's the same type and speed of memory
is this printed on the memory?
I recently added memory to some similar machines (P3, 750MHz or something like that) -- I think they took "PC133 SDRAM" in single pieces.
the memory I added had stickers
jimfleigh: yes it is printed on, mhz and type
thanks, gotta go
Less-wires ROCKS at home...
how's it going?
Good. I was thinking y'all might be able to help me with something: I want to change the file system type on my root partition. Do you know a way to do that without messing up everything?
from what to what?
short of doing a backup and restore of the partition, I don't know of a way. especially if the filesystems are different (e.g., ntfs to ext3).
From ext2 to something that journals... It only takes a few kernel panics before you appreciate journalling
ext2 can easily be converted to ext3, I think
maybe these instructions will help? http://www.troubleshooters.com/linux/ext2toext3.htm
ooohhh. let me check...
along the lines of tune2fs -j
just found it in the man page
not sure if there are gotchas or not
Oh, that rocks -- thanks much!
one gotcha is to make sure support is in the kernel
often the ext3 module is in the initrd
I'm pretty sure Paul put it in, but I'll check
a few years ago, at least, an ext3 filesystem with a clean journal can be mounted by a kernel with only ext2 support
do you have backups or another way to mount the disk if it goes wrong?
Not really, but everything can be regenerated.
It's just a PITA
yep always is
wonder if that means he got it to work
Well, that was easy. Thanks, guys
it's amazing how tasty Entenman's chocolate covered donuts are
sorry - that was all I could type before I licked all the chocolate off my fingers :)
typing with your elbows?
the other hand
the milk glass hand
we went out to dinner saturday night, got "triple chocolate cake" for dessert... just finished it off a little while ago
hm, that's no good. in this vmware machine, every keystroke is being echo'd to the console as "^@"
even pressing "capslock" or "shift" echos ^W
* jepler restarts
what guest OS>
wonder what lloyd wilson is working on
he doesn't give us much to go on to help him...
I was kind of wondering that myself
on the users list
oh, the "adding code" guy
is there a wiki page that tells the files you have to touch to get a new module to compile?
no, I don't think there is
(within the make - not how to get it to work, of course ;) )
you're guessing a hal module?
I think so
yeah, can't tell whether he wants to add another kernel module, another user space prog, or just another source file linked to existing code
could be userspace program as well though
I think it's just the makefile
"just the makefile" is still non-trivial
the top-level makefile is where all the kernel hal modules are listed
yes, in a few places though, with kbuild and non-kbuild
obj-$(CONFIG_MODMATH) += modmath.o
modmath-objs := hal/components/modmath.o $(MATHSTUB)
* mtimmerm is AWAY at 22:09:10 : I need a reason?
mtimmerm is now known as mtimmerm_AWAY
../rtlib/modmath.o: $(addprefix objects/rt,$(modmath-objs))
I don't even begin ton understand the makefieles anymore, I just examine how some other module is done and copy it
the first declares the module for kbuild; the second gives the object files in it; the third declares it for non-kbuild
that's an excellent strategy
now go forth and write the wiki page on it
for userspace stuff, see Submakefile.skel
if you'll hang around for 20 mins to review it, I'll do just that
if not tonight then tomorrow
I always glance at the wiki changes between reading webcomics
hmm, linuxcnc.org is taking a long time to come up
indeed - hmmm
skunkworks sent an email saying it's down
but I didn't check anything yet
pinging it also comes up empty
but you want the wiki anyway, which is on sf
dinero is down
I know I want the wiki, I was just gonna use SF to get there, I'm not sure I have the wiki bookmarked
[02:03:37] <SWPadnos> http://wiki.linuxcnc.org
heh, I guess that doesn't need a bookmark
just one or two working brain cells
I got used to the long url when it was on robin's server
as of about 15 minutes ago, they were trying to fix some "after-effects" of a problem caused by a switch went south on them (among other things)
SWPadnos: until I figure out you meant a network switch, it was quite funny that a problem with a single (pushbutton? toggle?) switch could lead to whole websites being down.
I also had a grammatical mistake, which may not have helped
I forgot how bad interlaced screens are
cradek: why are you using interlaced?
I'm working on the 7? 9?" monitor that I might use on the lathe
no, a tube
jmkasunich, no, PCP
I have a stand that will hold the monitor and a small keyboard up off the desk
* jmkasunich has sworn off CRTs...
* jepler summarizes recent improvements to AXIS: http://axis.unpy.net/index.cgi/01153186355
jmkasunich, oh come on, you know you like a good back ache every now and then.
ops, and I typo'd it pretty badly
CRTs are fine if you don't have to carry them
and if you don't mind sacrificing desk space to them
Heh, by 32" tv is like 120 # and no handles on it
jepler: support for ... is
"you might be a programmer geek if your computer monitor is bigger than your biggest TV screen"
who uses TVs anymore?
jepler: (last line)
these are really excellent changes
jepler: I'm curious.. what happesn with the "manual toolchange" component if you have Tnn many lines before the M6?
cradek: found it
it actually uses the hal pins doesn't it? it'll work right
I don;'t think tool prep time will matter much with manual tool changes
jmkasunich: I don't know. hal_manualtoolchange with the example hal file just displays whatever number is put on the pin iocontrol.0.tool-prep-number
I never tested with them on different lines
the tool prep request HAL pin goes active (and the new tool number comes out on a pin) as soon as the Tnn is seen, the tool change pin goes active when the M6 comes along
I guess as long as you trigger on the tool change pin and ignore the tool prep pin it will be fine
linkpp hal_manualtoolchange.change iocontrol.0.tool-change
linkpp hal_manualtoolchange.changed iocontrol.0.tool-changed
linkpp hal_manualtoolchange.number iocontrol.0.tool-prep-number
jmkasunich, what if my computer monior doubles as a TV ?
jmkasunich: what I wondered was, how do I know when to stop asserting "changed"?
the dialog pops up when "tool-change" goes true, right?
as soon as "tool-change" goes away, you can release "tool-changed"
classic two wire handshake
want to read the last few lines of the python and tell me if it looks right?
[02:22:05] <jepler> http://unpy.net/cgi-bin/viewcvs.cgi/axis/scripts/hal_manualtoolchange.py?rev=1.1
it would help if I could actually read python ;-/
if 'change' has become asserted but we weren't asserting h.changed, go through the change logic which sets .changed.
otherwise, if 'change' is not asserted, deassert 'changed'
I need to use a 1/8" bit, but need to go into a 1/4" slot that's about 3/4" deep. Any suggestions?
wow... california's hot... 41C... where it's like 23C here.
if change and not changed, do your thing and set changed
if changed and not change, clear changed
Jymmm: use my slot subroutine, it's exactly what you need for that
Jymmm: it's in useful-subroutines.ngc in cvs
or... 30C right now
only 28C here
at 10:30 PM
cradek, Sorry, I mean in respect to having the clearnace and depth for the bit (it's a 1/4" shank)
you need a bit with at least a 3/4" cutting length
jmkasunich, no other way (not being sarcastic)
I think we have no idea what you're talking about
I think I do (but it makes no sense)
ok jmk is one ahead of us
are you trying to cut a 1/8" wide groove in the bottom of an existing 1/4" wide x 3/4" deep groove?
using a tool with a 1/4" shank?
I have an upper slot thats .220 wide, then a lower slot that's .195 wide
I think I want a home weather station.
just for kicks.
ok, so you have to cut both slots with something smaller than 1/4"
A-L-P-H-4: I've thought that too
like 3/16 (0.188) or 1/8
jmkasunich, Yes. it's hard to find long reach 1/8" bits
the total depth of both slots is 3/4"
you need a bit with a cutting edge at least 0.7 long then (or at least, with a 0.7" long section that is less than 0.195 diameter
The best I can find (w/o having one made) is .5"
what are you cutting? wood?
* cradek giggles at the letter-writing campaign quip
Carbide Router Bit for Wood and Laminates Up-Cut, 3/16" Cut Dia, 3/4" Cut Length, 2-1/2" L O'all
In stock at $17.95 Each
pn or url ?
thats the pn
mtimmerm_AWAY is now known as mtimmerm
if you can use 3/8 or 3/16" shanks, then you have more choices... there are eight 3/16 dia endmills with 3/4" long cutting edges
jmkasunich, ok cool. But if I had needed .120" slot what options are there? I can't be the only one ever to need a deep/narrow slot
or .130 as example
they have 3/8 shank endmills with 3/4" long cutting edges in 1/8, 5/32, 11/64, and 3/16 dia
oops, thats 3/16" shank
break often I'd suspect.
in general tho, smart designers avoid narrow and deep slots
I need to make a video of this lathe - it is extremely cool
if you _must_ have a narrow and deep slot, try to make it straight, and have non-critical ends... then you can use a slitting saw
with a slitting saw you can make slots 0.030 or less wide and an inch deep (if you are carefull)
milling any slot that's deeper than about 2-3x the dia of the tool is always rough going
my biggest issue is clearnace from the shank itself, not so much the cutting edge itself.
you can get 3/16" shank end mills with 3/16" diameter cutting edges less than 3/4" long
wow, you can get a slitting saw 1" diameter, 3/8" hole, and 0.006" thick
sounds like a circle of tinfoil
circle of HSS foil
heh, I forgot I had these... 25mm dia x 0.2mm thick (0.008")
and 50mm dia x 0.3mm
I don't think I've ever used those.. but I have used the 3" x 1/32"
anybody remember the wiki syntax for preformatted text? (ie. use monospace font, don't screw with whitespace and linebreaks)
I think you can just use <pre> like in html
but I always dig around for another page and look at its source
yeah, thats what I'm doing
ah, you indent it with one space
stupid thing thinks "src/Makefile" is a link to another wiki page, puts a question mark after it
I think you can undo that ... somehow
I hate wikis even though they're ok for some stuff
my thread gauge has 11.5 TPI, wtf
that's worse than Jumm's nut
03jepler 07HEAD * 10axis/scripts/axis.py: don't throw a fit if the splash/demo file doesn't exist
[03:03:29] <jmkasunich> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?MakefileDeMystified
the title is a bit ambitious...
oops, found a mistake already
ok, now you can look
it looks great
what's the difference between a component and a user_comp?
user_comp is a user space program
not a kernel module
I didn't make that connection for some reason
so user means userspace
I thought it meant the user (person)
naming leaves a bit to be desired
as in that directory being the "human interface" stuff
(and CVS makes fixing it not easy)
no problem though, I get it now
yeah, but if it confused you, it will confuse others
assuming I'm average or smarter, yes
I forget what jepler was adding, but we couldn't figure out which was the right directory
although it's obvious now that you say how it works
* cradek shrugs
are you going to reply to lloyd now? the wikipage looks great to me
(assuming that's what he's asking)
I'm going to bed a half hour ago - goodnight
I just realized one additional step
gotta add it to the module helper whitelist
not sure that's still true
it whitelists a directory?
yes I think so
better check the source
or the module is in the directory /home/chris/emc2.head/rtlib/
if he wants to add to 2.0.1 it's going to be different
if so he can choose to do that the cvs way or the debian way
I'm assuming he's working with a cvs checkout
I should add that assumption to the wiki page
I can sure see someone wanting to do it the debian way too (make new packages with a customization or added module)
you get to write that part
only if someone asks for it
actually there are probably lots of resources for that already
"the debian way" involves doing it the CVS way first, right? then make the packages?
our packages are built the 100% standard and open way
apt-get source emc2
oh, I see
(make new changelog entry)
and pow, you have new packages
there is no ordinary configure/build step?
buildpackage does it for you
and you can't test by run-in-place, you have to install the new package to test?
doesn't seem like the path I'd encourage
well if your goal is to make a package, that's what you do
that's not usually what you do for the devel work itself
I'd prefer people to submit their changes to us (or get developer status and commit them)
like I said, I'll write that if someone asks for it (or, find a nice link that already explains it)
if they want to use it only for themselves, I don't care what they do, but if they're gonna distribute, oddball packages floating around are just asking for trouble
you may not like it, but it's important to me that it works like other debian packages
one of the things people must be able to do is change it and repackage
not arguing about that at all
ok I understand, not trying to pick a fight
I'm sensitive about that since not everyone seems to understand it's a gpl requirement :-/
ok bedtime for me, goodnight all
jmkasunich: thanks for finding the QCTP for me, I like it
(videos coming soon!)
can't wait for the vid
bedtime for me too...
This nickname is owned by someone else
If this is your nickname, type /msg NickServ IDENTIFY <password>
A-L-P-H-4 is now known as A-L-P-H-A
so i guess using Ubuntu is a must to use emc huh
no, but it's the easiest sol'n... as it's packaged that way... not dapper drake yet htough.. just breezy
im just having a hard time getting the ubuntu install to just work
shouldn't... works on standard hardware pretty easily.
I've only had video card issues.
ive tried it in 4 diffrent boxes and the same thine happens, so i reburned the disc and same thing happened
so i asked a friend about it and he said he had the same problem
wow, what an astounding lack of understanding from joshuakerns76 on the users list
I was just about to - erm - suggest that he's incorrect
and point him at some light reading
a great idea
I got sidetracked by a bunch of forum posts I have open in Mozilla :)
it's probably better to correct the discussion sooner than later
cradek: thanks for the info on linuxcnc.org
sure, looks like it's fixed today
Hi Chris. I was off reading log.
this morning when I was supposed to be getting out of bed I figured out how to do DTG, it's not as hard as I thought
Great. I'm amazed at how fast things are developing these days.
the way I have in mind, it will always be in "external" (ini file) units though even if the program is in the other units
so I'm not too happy with the idea, but it'd be about 80% right
(I seem to be way behind in acronym complieance these days)
"distance to go" (how far until the end of this move)
ray says people will want that on lathes
(but I'll do it for mills too)
reagrding the josh email - one thing to note is that the source code is available, it's just not in a highly publicized location
or is it missing stuff?
(I haven't checked)
whare can I read this post?
it's missing stuff
on the emc users list
specifically the scripts controlling compile/packaging
ok, and you don't get those with apt-get source emc2?
those are also missing in bdi4emc packages
(or whatever the correct apt command would be)
no, they aren't in the files you get
maybe you should get the exact details from jepler, he studied this more
I think the "source" he provides for emc2 is exactly our tar file, with none of his changes
right. I'm not sure aI want to go that far into the details in my email - just a not-quite-polite reminder to actually read the GPL before posting such tripe
yeah a pointer to gpl2 and "find clue here:" is maybe all that's required
"look for clue starting here?"
we shouldn't guarantee that he will "find" it.
something like that
rayh: you're assuming it will be a multi-step process?
I/m assuming that this is a correct link: http://www.gnu.org/licenses/gpl.html
I hate the delay that the source forge online mail viewer has.
hate is a strong word
Mindset often requires several steps to fix.
step 1) steal underpants..
step 2) ask folk to secure nomex suits.
step 3) Profit!
damn - maybe that's been my problem. I've been using the wrong version of profit all these years
Humm. Might be a common problem.
SWPadnos: paul's broken source upload includes a "orig.tar.gz" but omits the diff.gz and the .dsc, so he's withheld his changes and also made it play badly with "apt-get source"
hopefully I haven't said anything incorrect in my reply
hm, is paul being the problem again?
still, I think (not a new problem, merely a continuation of an old one)
03cradek 07HEAD * 10emc2/src/rtapi/rtapi_math.h: pow prototype
I know this is off topic, sorry.
rayh: is DTG typically used while going very slowly or in feedhold?
rayh: I ask because I'm not sure what it should display during a blend
Ah. DTG is usually very slow and used during a cut but can be used throughout an approach to a cut.
ok, when running at speed, you often don't get right up to corners, so it won't go to zero
but it'll be changing so fast it won't matter
when going slow, you do get right up to the corners
Sure. My guess is that with a lathe turn the operator would use exact anyway.
03cradek 07HEAD * 10emc2/src/emc/kinematics/tp.c: distance-to-go display
03cradek 07HEAD * 10emc2/src/emc/motion/motion.h: distance-to-go display
03cradek 07HEAD * 10emc2/src/emc/nml_intf/emc.hh: distance-to-go display
03cradek 07HEAD * 10emc2/src/emc/task/taskintf.cc: distance-to-go display
03cradek 07HEAD * 10axis/extensions/emcmodule.cc: distance-to-go display
03cradek 07HEAD * 10axis/scripts/axis.py: distance-to-go display
03cradek 07HEAD * 10axis/tcl/axis.tcl: distance-to-go display
jepler: can you help with the presentation
I don't see why it doesn't line up with the other things
also, the rectangle under the axis readouts is not showing up anymore
I'll look at both
can you figure out if the rectangle went away when I made the cone transparent?
it must have
I can't have borken that
that's what I think too
[15:38:20] <cradek> http://timeguy.com/cradek-files/emc/dtg.png
you probably missed the hack to make 'Vel' line up
yeah I didn't notice that
"install2" is quite a bit slower than my laptop
I guess I'm competing with you running emc too
hey, I even got arcs right
wonders never cease
you're supposed to say "I meant to do that"
nobody would believe me
oh yeah - forgot about that
03jepler 07HEAD * 10axis/scripts/axis.py: fix DTG display, rectangle under readout
glVertex3f(pixel_width+24, ypos - 8 - coordinate_linespace*len(positions), 1)
wow, this code sucks doesn't it
a lot of the code in axis sucks
I'm sorry you feel that way - I think only that one line sucks
do you want me to find some others?
SWPadnos: haha your clue pointer did NOT work
I don't know if you followed it from the diff, but the rectangle disappeared because I turned on GL_CULL_FACE and I guess the quad was drawn backwards before now
oh, no I had no idea
it's natural but wrong to draw stuff clockwise
SWPadnos: and David doesn't understand we're talking about an emc2 release, not a bdi release
I'm so going to keep out of this one.
* cradek plugs his ears and says LALALALALALA
Mornin Gents, gonna be a hot one today!
did you see that we already implemented all the things you asked for yesterday?
can it be true?
where would I see that?
DTG and expressions in "touch off"
oh, and 1% increments for FO
wasn't DTG a PITA to code?
and expressions in touch off too?
that's what we said..
but both of them turned out to be doable
ok, sounds nice
was there large modification or was it an obscure smart change?
DTG wasn't too hard once I thought about it the right way
I don't know about DTG, but supporting expressions amounted to ~50 lines of code
jepler: 50 lines?
jepler: that much?
cradek: oh, ok, so no major changes then?
I'll withhold comment until you try it and tell me if you find ways it doesn't work right
oh, ok. good idea
I won't be able to test it for a while though
it'll still be there when you are ready
I'm going to go on a shorter vacation for a week or so
that sounds nice
no access to net or lathe
goodnight, I'm away for the evening
see you later
a week or so
just so you know paul is in the high sierra being a mountain man or something.....he vanishes into the mountains without computers or anything for a month or two....
lol @ les but but but deliverance, 3 miles from you, mountains, oh nm!
* robin_sz wonders why he just did something crazy
hey les_w ,
did you buy that plane I sent you the advert for?
I'll be back in a week or so
the secret to flying is throwing your self to the ground - and missing.
skunkworks: the best way is if someone distracts you while you are falling
boy it has been a long time since I read the books