SWP_Away is now known as SWPadnos
Thanks again for that line. It was/is well beyond my ability.
I believe that it runs properly now.
I noticed jepler's comment about comments
I also see that sf changed the permissions to executable for it.
do you want me to change it to replace the whole string after the '='?
oh - I hadn't noticed that
No. I think we need to talk about input and output scale to see if we keep or toss that second var.
It seems to work fine as is.
it might be nice to be able to edit the "Machine Name" at some point
but I'll leave it for now
Yes there should be a whole tillie front end to many of those things.
My notion is to move configs to netlisting
and radically simplify the ini.
something like that could work, but I think there need to be "bundles" (grouped signals) first
I find the netlist just about as easy to man read as the several .hal files
add an axis with a single command.
I'd also like to have some standard naming, like using P I and D on the PID block, instead of PGain IGain and DGain
Did you get a chance to look at jepler's ini generator.
that would match the ini settings to the params much easier (automatically)
That naming set me back a ways.
But you've got to remember that jmk was working alone and in almost constant battle with paul.
I've had in mind a halcmd-like program that can read an ini section, and apply all the settings to params
something like pid.0.* = [AXIS_0]*
look at all the param names, and try to find identically named variables in the ini. if found, set the param
but there would need to be some more flags for the params themselves
The whole of configuration is negotiable.
If you look at emccalib, much of that ability is there.
But klunky as hell.
It's almost like we have to try or start a configurator in order to see the needs.
itteration or as someone said, eric raymond I think, evolution.
halcmd was a great idea.
I'm torn between the desire to have very good config manipulation tools, and a not-necessarily-human-readable confg file
is a great idea.
and having a human-readable file, which allows us to mostly ignore config tools
You know I lean heavily toward a totally invisible variable/value storage.
that's almost a requirement for dedicated machines
XML would be great in many ways
there's also a more readable / compact version of it
Then make several editors, plain text, tillie, ...
called PXSL or something
kate has an XML validation plug-in
[14:00:20] <SWPadnos> http://community.moertel.com/pxsl/
Sure any system that is able to display and save values.
validation is a little beyond that - you can specify what kinds of values are allowed
I'm not sure exactly how far it goes
If we did the system right it could even be displayed in a web browser.
but I think you can do things like String, number, list-of-options, etc
Hell we could build a webmin interface.
it could actually be edited with a browser
That would shoot my work with emccalib.
and halcmd ;)
and all the .inis and .hals
But such a thing would be a giant leap forward because it generalizes the whole config process.
Not just a little part of it.
at this point, the only thing loaded from the runscript is io, and even that should be loadable from halcmd loadusr
other than some measure of backwards compatibility, there's almost no need for the ini file at all (since all the functionality is really in the .hal files)
Right. That would get all of config back into a single place.
And it would allow users to build a system from scratch and save as they work.
I think it's possible to ask tillie to edit a file, but telling her that she needs to edit 6 of them, and tell one file when to load the others, is bad
And not be able to tell her which of the 6 contains the problem she's fighting.
"tillie - just use grep to find where the P for the rotary table is being set" ;)
I sure do miss the file content search stuff available in konqueror.
grep is another thing I do not do well, nor read the result well.
it should still be there in nautilus
isn't nautilus the gnome filesystem browser?
I think so but there isn't a find in the stock version from ubuntu.
hmm - strange
I looked for plug in's and found some. But when I tried to apply them, it was a bust.
well, at least it should be there in the 2.14 release ;)
I saw quite a few improvements to gnome. I hope that ubuntu doesn't rip em out.
That was my big issue with fedora when RH started saying we have a better way.
ubuntu bases their releases on the gnome release
they both do 6-month release cycles, I think
Yes. It's just that ubuntu edits the list of gnome packages included.
Most are available in the universe thing but you are at the mercy of users.
Universe repository is user generated.
at least it uses a preference order for the repositories
Like I could get a LyX from there for ubuntu but it isn't current with the main ubuntu libraries so it doesn't load.
Yes. And the core stuff seems to be rock solid.
Very few crashes.
Even with the real-time.
I've liked it so far
unfortunately, RT and power management don't seem to get along, so extra care needs to be taken on laptops ;)
Yep. All the folk I've sent that way seem to be happy with it.
Yes. I had to turn off all battery monitoring on the old gateway to make it live.
I'm thinking of the other problem - can't do much development on a plane ;)
Yep. I have not found aircraft or airports to be very code friendly places.
not with RT kernels
mu giant laptop lasts 3-4 hours when doing development work otherwise
it's old and clunky - PIII 1GHz, 512M
15" 1400x1050 screen
You mean only four batteries and we could code all the way to Bejing
nah - for beijing I'd go first class, and use my power adapter ;)
around-the-world coding, baby!
hmmm - maybe codefest should be on a cruise ship next year ;)
Not a bad plan. I've got a lake and a 10' cruiser. I'll do the rowing.
maybe we need two fests then
NorthernFest and SouthernFest ;)
One for the poor and ...
we needa wealthy benefactor
Yes we do.
actually any benefactor would do for a start ;)
I like what I see of PXSL so far.
it looked pretty good to me, but I'm not seeing any DTD equivalent
No and even with the xml converter, it would have trouble in the read/edit/save environment with anything but their own editors.
I thought they had a library for program use
and it's almost editable as text, for the power user
I can just see tillie getting out bluefish and writing an ini.
how is it?
Good and getting better. How's your cold?
a bit better, was at work today
good to work but you can still feel pretty used up when the day is over.
* chinamill is away: eat
SWPadnos is now known as SWP_Away
wonder where cradek disappeared
he said the heck with you guys - making him do all the hard stuff. That is what he told be anyways :)
rayh: seems we both thought about MDI :-)
Yes it seems like it would handle the task if it is a simple one.
* chinamill is back
rayh: you around?
Yes I am.
What can I do for you, alex_joni?
nothing ;) just wanted to let you know I added a search box for the site, and wanted to get your oppinion about it :)
actually I'm bored, and not feeling like doing anything :(
Ah I'll look.
Wow. That is great.
I think we reached about 15% of what joomla can do :))
I've had some folk say that I should have included a google search but I didn't want to.
This is very nice. Results laid out consistent with the look.
I'm sure we will discover a lot of stuff.
I'm glad we did this
I'm really glad that you took the time to find joomla and do the work.
And thank your friends for me as well.
the site looks very good, quite professional!!!!!
Roguish: well, thank you
cradek: good morning ;)
The search is really nice. We can add a lot of stuff to documentation and it should make browsing painless.
rayh: I added a category a while ago, for Technical Articles
maybe some days we'll start adding those ;)
describing HAL & CL & such
Oh. Good plan. Would they be html, pdf or what formats are available.
doesn't matter, you can paste into that editor you liked
I tried various programs, and the paste worked very good each time
If they were html they would have the same look as joomla or would that start a browser.
you only need to put the text in there, the rest is taken care of by the CSS (to feel consistent)
I was thinking about things like Hal_Introduction
for big stuff like that, pdf is best
most of them would be for offline reference
at least for most users
Then the user would click the link and it would raise their own viewer?
yeah, something like that
but for HAL, I would like a short description, and the HAL_Introduction.pdf linked
Okay. How do I find this technical articles.
rayh: right now you don't ;)) only from the backend (admin interface), but I can make it look somehow like the Events interface (http://www.linuxcnc.org/index.php?option=com_content&task=blogcategory&id=15&Itemid=12)
That would look really good. And the link to it would be on the documentation page.
give me a couple of minutes
Nice job Alex.
now we only need to add technical articles ;)
btw, there is a statistic module int he backend, for people who wonder about visitors & such
How about getting the man pages in there.
heh, I actually found it amusing
45% doze users vs. 10% linux users
probably I was most of the 45%
surfing while at work = windoze user
what are the other 55%?
(well not for me, but most I bet)
23% unknown ;)
I'd think that is a pretty heavy use of linux. Compared to the world average installs.
jepler: your math is a bit off ;)
1.10% Mac Os X
0.31% Win 2k3
0.16% Free BSD
0.16% Win 95
oops; what are the other 45%?
Debian is different than Linux?
seems like it ;)
I suppose it depends upon how the linux user has his browser configured to respond.
rayh: how is emccalib?
It's running and saving stuff.
About as far as I want to go with it right now.
There is the problem with changing input scale at an unknown location.
I'm working on halconfig right now. I've got tabs going so it looks like the other but a ways to go yet.
ok, I think we might be ready for a branch
what do you think?
I'm worried that we keep pushing the emc2 release
Right. I don't really have a good idea when halconfig will be ready. It is okay now for viewing hal just not for saving.
yeah, but I don't think that's crucial for the release
As far as I'm concerned we can branch.
it's not like we won't have any other versions ;)
In fact I really like the branch idea. We should label that release xx
cradek: any other concearns?
I don't know, I haven't tested in a while
We will do bug fixes on the branch so if something isn't up to snuff we can fix. Just no new features there.
Will the branch become the next testing?
can we set up a timeschedule?
Or how do we handle.
rayh: that would be best
can we call it TESTING?
How about now for the branch.
rayh: let me run it first
And sunday for the release.
I've seen some commits from flo-h
and he had some breaks before?
Right. flo broke a few tickle things.
I put the back.
yup, I know
maybe we need to call up our translators after the branch
and then we'll try to not change many strings after that
cradek: how does the timeschedule sound?
branch & testing now
if all ok, release (rc1) on sunday?
that's a fine plan but I think any schedule is nearly meaningless
yeah, but it's still nice to have
ok, I'll test for a while, then do the branch.. everyone ok with this?
fine with me
what are you going to call it?
* alex_joni is thinking
it's not something I'm used to :D
In axis, I call the branches maintX_Y when the releases from it will be X.Yrc1, X.Y, X.Y.1 etc
you can't have .
well, _ then
short for "maintainance"
some call it "release"
or maintenance even
you don't want to give it the same name you'll want to eventually give a tag
cradek: yeah that
if you're going to release 2.0 and 2.0.1, but not 2.1, from that branch, then you should call it 2_0 and not 2_0_0
I like the sound of that.
what will the actual release tag be then?
I assume that once the branch is made
jepler: yeah that
* cradek no longer cares
then head goes on toward 2.1
cradek: you lose interest pretty fast lately ;)
so how about v2_0 ?
that's short enough :)
how about v2_0_branch
I can live with that
we don't want people to mistake it for a release
But it will be the code used in the release, won't it?
yeah, for all the 2_0 releases
rayh: we branch to v2_0_branch
which will continue in parallel with bugfixes
and when we decide (on sunday) we release rc1
then we'll have some more bugfixes and then rc2, and so on
till we say it's ok, and we'll have emc-2.0
then a new bug gets found, more bugfixes on the branch and a emc-2.0.1 release
and so on
That sounds like the way that I think about it.
so we all agree the TESTING tag will be on this branch for a while, right?
rayh: btw, SF staff finally made that emccalib.tcl executable
I see that.
it only took them a week or so :)
Hey I got a email saying paul had posted to the list again but I don't see anything in there.
what does again mean?
he sent an email earlier to both lists, about new bdi4emc packages
This was after the first post and jepler and my replies.
oh, don't think I've seen that
maybe my imagination.
where did you get email saying that?
I got it from sf "nonmember post to a member only list"
* alex_joni looks
Emc-developers post from firstname.lastname@example.org requires approval
Date: Wed, 29 Mar 2006 09:57:02 -0800 (11:57 CST)
Subject: Re: [Emc-users] New BDI-4 packages.
* alex_joni looks if the post needs approval
no pending tasks
something's gone belly up with tkemc. emc errors cause tickle failure.
can you tell me how to replicate?
Try a jog in auto mode.
need to build first :)
I just finished building it. I had not seen that behavior before this build.
I think it's msgcat related
doesn't flo-h test his changes at all?
not sure :)
Worst case when changing input scale is following error on the axis changed.
I wonder what it would do on a real machine.
I don't think you can change the scale without restarting emc
you can in HAL
It does here. Or seems to.
oh does emc not know about scale at all anymore?
no it doesn't
it knows about accel/velocity and you can't change those
it only knows position & feedback
forget I said anything
cradek: I won't :P
I see that changing scale seems to change the returned value okay.
just following error message.
I suppose I should trap changes to input scale and test for 0 axis position.
Having a home offset active does affect the input scale change also.
But I don't see anything that would make a machine run away.
rayh: maybe aunt tillie might succeed, if she's ugly enough
She is ugly. Could easily be the bearded lady in the circus.
well, ugly users can cause machine run aways
but that's not our fault, imho :)
if the machine is smart, that might be our fault :)
we had a solar ecplise today
I hear :)
Oh right. Did you watch.
no, was too lazy to get up from my desk
I know that feeling.
only 65% or so coverage
last time we had a full eclipse, that was nice
I built a projector for it as a kid so I could watch the progress on a paper.
cradek: still there?
tkemc works here now.
* alex_joni attempts branching
hope I don't brake anything :)
ok, done branching & moving TESTING
I'm off to bed.. g'night all
I need help with a system that is refusing to install the apt-get build-dep emc2