huh someone on the mailing list is complaining about the unreliability of sf (mailing list archives in this case) :-P
[00:34:15] <jmkasunich> http://jmkasunich.dyndns.org/pics/sourceforge_sucks.png
should call that "sourceforge_spams"
I submitted a gripe about that issue on 5/26 - two months later, there is nothing on that tracker except their initial canned response
they didn't comment on it, they didn't dismiss and close it, nothing
at least the Dice ads are working
yeah - the adserver is the most reliable part of the site
it's probably offsite ;)
nah, its just a higher priority
I can only blame them a little bit
the "provide free services and hope people donate to support the costs" business model didn't work
now they're doing whatever they can to milk some cash out of their traffic
man. nothing makes a hot day seem hotter than (a) eating raw garlic (marinated in a hot pepper oil) and (b) washing dishes
I find it interseting that spicy food tends to come from places with hot climates
"damn, this Mexican desert sun is hot" "Here, have a nice Habenero and cool off"
I think it has to do with keeping little beasties away from the food
ISRT reading something a while back about that
they tested several spices for preservative qualities
did it work?
don't recall the details
some did, some didn't
I'm pretty sure that's how spices were used in the middle ages and thereabouts - mostly to either help preserve the food, or to cover up the "aged" taste :)
cradek: "I don't fault anyone for not knowing about PKC." either, but I do fault someone who makes the claim multiple times since it advances some sort of agenda he has.
now I wonder what you could be talking about ... :)
looks like we ended up with a good set of board candidates
yep. short list though
just make the board 7 members, and we don't need no steenkin' election :)
jepler: me too, and I think in the next sentence I covered that sentiment, more or less
well, I feel bad for jmk, but that's pretty funny.
heh - indeed
huh -- "python-modbus" gets only 3 hits on google, and the CVS server named in one of them doesn't work, at least at the moment
lots more if you remove the hyphen
[13:52:44] <cradek> http://lintouch.org/static/download.html
is is true that there are daily tape backups of the CVS server, that jmk still has "cvs2" waiting in the wings, and there's a vmware CVS server image available somewhere?
I make tape backups (but not daily) AND there is a vmware image available to all of us
ok, periodic then :)
when you weren't looking, the original hardware died and jepler and I used the vmware image
that must have been the 5-minute outage a few months ago :)
yeah it went pretty smoothly.
did you and jmk ever get his cvs2 VM up and running?
he had a machine, I don't recall if it was vm, but I think he doesn't right now
but, he has access to the vmware image anytime
(so do you and alex)
once I get my server configured, I can set up another backup here
cradek: how often is ~emcboard/var.cvs.backup updated?
I don't think we need anything else, but feel free if you like
is all the user login information in the VM or in the data backup?
or somewhere else
ok, so you (/the cvs admin) need to update the VM image whenever a user is added
or not, since the data restore will overwrite the appropriate files
I think they happen to match today, but the VM will get out of date eventually
right, things like the developer pubkeys are in the nightly
the VM is from 11/26/06, so Chris Morely should have been added since then, I think
err - where is that?
jepler did that upload - I'm moving it to ~emcboard now
rm: cannot remove `/home/jepler/cvs.linuxcnc.org-20070706.tar.gz': Permission denied
cradek: you're ready for me to remove that file now
it would be very nice if the unix permission structure supported "sub-root" access
specifically in the case of shared hosting :)
lately I ran into an NIS limitation where users can only be in a small number of groups (8?)
that really makes it worse
maybe it was more than 8, but still caused me trouble
permissions are one thing that Novell really got rigjht
but it was a bit complex
yeah weren't there about 14 bits?
it was more the inheritance that seemed to work well
novell also got some obvious stuff right, like how a copy didn't have to go into the workstation and back out
I think you could add attributes as well
complex like WNT permissions, or complex yet useful?
complex yet useful
(oops, is my bias showing?)
I didn't notice any
the unix model is surprisingly useful/flexible for its apparent simplicity
so much that ACLs never caught on
yes, but its lack of hierarchy is starting to make it less "competitive"
I liked that Novell had separate rights for create, delete, read, write, modify, and I think append
in a sense, modify without delete is silly, since you coulr modify the file to contain 0 bytes (or one NULL byte)
and that rights to change rights were separate - so someone could have the right to modify a file, but not to change permissions
(or vice versa)
I was never a netware admin, but I suffered through some abends and 6-hour-long bullet-sweating filesystem rebuilds with someone who was
well, those were the bad parts
but, there was the cool snake
that made up for it
I'm kind of bummed. my old company used to use Novell + groupwise, but groupwise backups took way too long. so they switched over to Win2k3 for everything
groupwise sucked back when I was involved with it
they would use Linux + open source stuff, but they can't find people to support it in this area, and they're too interoperability-impaired to look very hard
I think they never quite got smtp
we had to hide it behind a sendmail which carefully spoon-fed it
actually, the features and functionality of GW were great (recently), it was just the 8-hour long backups that sucked
also, it would do the abominable thing where it would get an email to an invalid account, accept it, and then generate a bounce (spam) to whoever was in "From:"
at least with Symantec backup, it seemed that it did a database query for each message instead of backing up the data
(surely they've fixed it by now)
which could be the right thing to do, since the mail system remains up while the backup takes place
I liked how it would refcount attachments instead of copying them around
(if only we could have that with imap today)
yes, and the shared mailbox thing was nice. also "unsend" :)
and "did he read it yet?"
not that I care about either of those, since I never make mistakes
of course not
lintouch looks quite nice
too bad it's been a year since the last update
I still didn't find their python modbus interface ..
is this the right way to wire multiple switches? http://emergent.unpy.net/files/sandbox/switch-nc-series.png http://emergent.unpy.net/files/sandbox/switch-no-parallel.png
or "a" right way, at least
looks good to me
it may be good to have a series resistor to ground for the parallel case, so you're never connecting directly to a supply rail
same with NC probably
though adding those would require you to put in some R values, since you want a 10:1 or higher ratio between the pull-up and the series resistor
I'll just call them "simplified" diagrams in the documentation where I'm incorporating those images
since you might also want to add a C for debounce, or schmitt triggers, or god knows what..
sure, plus diodes and the like, and other stuff that may depend on the supply voltage
I think matt S had this
yeah, that's the one that pXXl says is no good for unspecified reasons
well, it's french
I don't remember if matt had it working or not
this is a bigger list: http://www.modbus.org/tech.php
the modicon kernel driver may be a good place to start
isn't there also a modbus driver for RTAI/RTAI-LAB
hmmm. maybe not
SWPadnos: p**l just said if we? want help with firmware stuff we should just ask. Isn't that what you recently did, and it only set off one of his tirades?
actually, I asked the question like 6 months ago
and I think I got a response something like "anyone with a clue can do it, bonehead"
oh - maybe it was only 3 months ago
ha, if you look at "more details" it shows that p**l often checks in 100,000 line changes all at once. That makes him a big contributor I guess (or more likely, it means he screwed up a lot of formatting using indent, or added a bunch of /*~!@$\\ another #if 0 */ -type turdlets)
for stepconf, is it most natural to enter step waveform timings in microseconds or in nanoseconds? Different driver datasheets seem to give different values.
I'm leaning towards ns, because the user will also have to enter the latency number they got from running the latency test, and that's in ns
I bet really fast ones are nanoseconds, and the rest are microseconds
having 4 "us" and one "ns" entry field is worse than anything else
cradek: "more details" where?
how hard is it to have a drop-down for the units?
or a radio button pair
(from his non-seq response to Mark)
I like that CVS tree view
yuck: "StatCVS is written in java"
too bad I'd never get it to work on my system
but it has prettly little 'x'es on deleted folders
new prerelease of stepconf is mostly changed documentation, some of it documenting features I haven't added yet. http://axis.unpy.net/01185733075
jepler: got a small request for you if you can spare 5 minutes
can you update the axis webpage with the spanish translation?
jepler: can be some other day too
I'll mail you some screenshots to remind you of this in a couple of days.. is that ok?
thanks a bunch
alex_joni: if that translation is in the branch as well, it may have the same encoding problem (specified iso-8859-15 but was actually utf-8) ..
no, it's only in TRUNK
I thought it's not a bugfix ;)
it's too bad one has to modify files in debian/ and so on to add a new translation
ideally it would not be at all invasive and we could do it on stable branch releases
otoh, we don't add translations daily
no, not often
hmm.. I had no idea I wrote so many lines of code :P
heh.. paul's majour work is documentation?
according to those stats
[20:33:12] <alex_joni> http://eneas.juve.ro/~juve/stats/
those stats look a little outdated :)
SWPadnos: what do you mean?
oh - maybe not. duh
they are about 14 minutes old
I was looking at all Paul's work, which ended in 2005 or so
it's fun to see what you've been working on
alex_joni: nice. it pulls all that from the cvs server?
awallin: yeah, from cvs log
SWPadnos: it's quite accurate too.. the commit hours for me are dead on
[20:41:02] <alex_joni> http://eneas.juve.ro/~juve/stats/activity_time_alex_joni.png
alex_joni: you clearly have it on european time. mine looks similar http://eneas.juve.ro/~juve/stats/activity_time_awallin.png
but ofcourse much less densely sampled than yours...
haha.. jeff's top work was on classicladder
well that's useful to know
at least I'll know who to assign CL bugs to
Dallur is pretty precise about when he does commits. Saturday at 14: http://eneas.juve.ro/~juve/stats/user_jarl.stefansson.html
same day though
yeah, just realized...
jmk has a pretty long history if you look at the dot-cloud graph thingy
third from bottom on this page http://eneas.juve.ro/~juve/stats/developers.html
what happened in Nov-2005 ? activity is suddelny boosted
I think we got voted back then
27 nov 2005
after that we got serious about releasing emc2
hm, that is also where paul_c's commits end
nothing too revealing there (?)
nope, but you can select 2005-12, and so on
it was in the middle of getting emc2 ready for 2.0.0
which was done in march 2006
there must have been emc1 before Oct-2003 ?
oh sure, and even after
this is just a stat for emc2 only
ok, oct-2003 was when emc2 was moved to sourceforge?
when emc2 was added
emc was the one moved to sourceforge
in 2000 or so
alex_joni: thanks for the chat! your stats may be interesting for others too. need to sleep now.
I wrote a hal driver for our fpga PCI card. I would like to make it part of EMC under GPL licence.
Who should I contact to submit the files or get cvs access
this is for the ARCNC100 CNC appliance
ok - I've heard of it but not seen one
not seen in person :)
I have been discussing this on the EMC channels and folks there suggested I submit the hal driver for inclusion into EMC2
yep - good idea. how big is the code?
as well as configs files for our hardware stepgen
just 3 files in the hal/drivers plus some config files in a arcnc directory under configs
well, I'd offer to take a look at it, but I'm probably not the best person to do it at the moment
maybe you should send a message to the emc-users list announcing that you have the driver written, and ask there for one of the developers to take a look at it for inclusion
will do that
sorry I can't help more - I'm kind of swamped at the moment
its ok, just wanted to know how to proceed and you answered that