dang... SF is estimating the CVS outage to last 48 hours _minimum_
a lot of content from the former linuxcnc website is now missing!
jmkasunich: I have a copy of the links page, since that was the only thing that I ever edited on linuxcnc in the past
gotta go. lightning in the area
jmkasunich: what's missing?
on further investigation, I'm not sure what is missing and what is just massively re-arranged
the bdi page is missing
(I was looking for the list of mirrors so I could get bdi isos for the new farm)
cradek: you still around?
did you get networking to work with qemu?
the docs are _not_ clear
if you only need outgoing connections, just run it with -user-net
it's kind of like nat
I got the impression that -user-net meant no actual internet access, just faked
no, you'll have outgoing
even dhcp worked when I installed ubuntu
I started the ubuntu install without specifying -user-net
I wonder if I should start over?
it won't find networking
how far into it are you?
not too far
probably should start over
probably easiest to start over, otherwise you can set up networking later
it just won't find a "network card"
how long did the ubuntu install take for you?
not sure, I went to bed
it probably takes a couple times as long as a normal install
did you get the "QEMU accelerator" or are you using straight QEMU?
I didn't mess with the accelerator
I tried it on redhat9, and didn't see much difference
the webpage implies perhaps 2:1 or better....
I was using autocad in dos which might not be the kind of thing it helps a lot with
well, I'm just using vanilla qemu right now... but the whole purpose of this excercise is to get faster compile farm results......
I love sf's site status page that says they might try to fix cvs come monday
I think we need to move, I just don't know where to
on one hand, I'm tempted to contribute to SF, after all I've been leeching off their service for two or three years now
on the other "for this service I'm supposed to pay money?"
I contribute to freenode because the service is good and they asked nicely
but sf is another story. I also would be more likely to pay them if there wasn't irritating advertising splattered everywhere on their site
some of the ads pop up over the links I want if I accidentally mouse over them
the new site layout is definitely a step backwards, it seems to be optimized for ads
that kind of crap bugs the heck out of me
its pretty clear the ubuntu installer is not SMP ;-)
one CPU is at 100%, the other at <10%
your qemu process runs on just one processor
now if I was running two qemus at once....
you could start another install if you have an iso
every so often the CPUs flip places
I've got several
BDI-2.20, TNG, and 4.27
and of course ubuntu
I also want to download BDI-4.38
maybe I'll start the 2.20 install
hope those old installs see the "network card" without too much dinking
I guess even the dos network client sees it if I use qemu -isa
it emulates a NE2000 card, that is as old as dirt
I'm kind of bummed - I spent $80 on this encoder and I don't know if I'll ever make a real part with this setup
you should be able to move the encoder to your lathe eventually
I looked into getting the minimum sherline parts I'd need to cnc my lathe (using my existing drivers) but it's like $300 ($500 if I get the good encoder ray is getting)
I'm not sure the sherline is worth that kind of money - it has serious size limitations (which is why I bought a bigger lathe, and it cost <$500)
what parts does that include? motors? or just motor mounts?
motor mounts and motors, that's it
I lied, the motor mounts alone are $200
plus $75 ea for two motors
not worth it for such a tiny machine
so I'm near $400 with no spindle encoder
you could reuse your $80 encoder
I'd pay that to cnc my big lathe, but it's probably too sloppy to be useful. it's pretty worn.
the $200 encoder ray bought has a hole in the middle - very easy to set up
I just realized that I didn't specify an image file when I started qemu
maybe something will turn up at workshop
oh hell, no hard disk?
it just got to "starting up the partitioner", this should be interesting
yep, no disks to partition
* jmk_cf starts over (again)
qemu -hda some-image-file -user-net
and -m 256 -cdrom <iso> -boot d
I got all those right, and forgot the most fundamental thing
your little lathe needs NEMA23 motors, right?
I have some, dunno if they are torquey enough or not
yes I'm pretty sure their mounts are nema23
I like the sherline motors because they're dual shaft
yeah, thats nice
I've wished for that so many times on my mill I'd be really hesitant to not put them on the lathe (that I'd probably want to use manually most of the time)
got two CPUs at 100% now ;-)
remember my unhappy state yesterday, when the box seemed fscked?
yes I saw you figured it all out late
man was I relieved
ohh, this ain't good....
it's a nice machine
the bdi-2 install uses X
it'll be fine
and the emulated X display is unreadable
usually it just works
maybe the X is too old
widget titles are OK, but the text in the text boxes is no good
sounds like a font problem or something
Simulate a standard VGA card with Bochs VBE extensions (default is
Cirrus Logic GD5446 PCI VGA)
or color depth - I saw somethign in the qemu docs recommendgin that the host have 16 bit color
I think this is 24
oh 24 is bad for some things
I bet you should use 16 or 32
can I change on the fly? or do I have to reboot or restart X?
you'll have to restart X to change color depth
* jmk_cf shuts down both emu's :-(
fix it in xorg.conf and start a second X server
easy for you to say
I'll help if it's not too late
the installs were just barely started
I'll fix it in xorg.conf and then restart x
(but help with xorg.conf would still be welcome)
just change DefaultColorDepth near the bottom of /etc/X11/xorg.conf
looks like I already did that
Device"Matrox Graphics, Inc. MGA 2164W [Millennium II]"
1$ xdpyinfo|grep "depth of root"
depth of root window: 24 planes
john@emccf:~/farm$ xdpyinfo | grep "depth of root"
depth of root window: 16 planes
the good news is I don't have to change it
hmm, guess that's not the problem
the bad news is you killed your installs already?
the bad news is I have no idea why the BDI screen is funky
are there any options at the initial bdi install screen?
iirc you can install rh62 in text mode
you type "text" or something like that at the lilo prompt
although that may simply defer the problem
but even if X never works, you're ok right?
i'll be really disapointed if there is no X
especially on the ubuntu one
I'm sure it'll work
rh62 will probably work too - maybe you'll have to dink a little
I bet rh62 entirely predates qemu
I hope it works (or can be made to work)
if my ubuntu install ever gets done installing its malayalam and gujarati fonts, I'll let you know
is it _still_ going?
it had stopped for a full 24 hours asking me what video mode I wanted in X
thats damned annoying, it should ask all questions up front
yep, but I see why it can't, it has to do it when it installs the packages (that have to be installed in dependency order)
but for X, if it would pick some resolution I could just change it later if I wanted.
I'm resisting the temptation to set up cvs using this copy of the emc cvs tree
a local repository?
no, I'd share it
I'm pretty sure I could host our cvs.
any old dsl like I have (and jepler has) is more reliable than sf.
bdi-2 is has already formatted the disk and is installing packages
ubunutu is still asking me what keyboard I have
yep no software ever gets faster over the years, does it
speaking of that, I'm installing ttf-tamil-fonts now, in case I ever learn to read tamil
oh, now telugu
I've never heard of these languages
(to be fair, I think it's cool when I see all the web pages displayed right, even though they're in another language)
ubuntu is very proud of their excellent i18n
yes they're sure serious about it
but it really slows down the install
you wouldn't mind if you only spoke telugu
we're spoiled because we speak 7-bit ascii, er, english
I'm gonna miss ascii
"just say no to unicode"
I'm partial to iso-8859, it gives you latin languages while still using one byte per character
and a few nice symbols
I just don't want to have to think about "wchar"
yep, it's pretty nasty
that's why 8859 is still nice.
this is one busy PC
2 cpus both solidly at 100%, 433M of ram in use, and 100KB/s network traffic (I'm downloading two more isos)
sounds like a good test
I ran memtest86 all night and day too...
but that's only one cpu
this is excersizing everything, NIC, disks, etc
I'm really glad I was able to get that extra 256M of ram
it will surely come in handy
speaking of "getting stuff", I now have a total of 11x18G drives and 4x9G ones
there are 2 18's in this box and I'm gonna keep at least 2 spares...
are they the plugin type or the normal scsi type?
I htink its 11, maybe 10
plugins, with the carriers
bdi-20 is halfway thru with installing packages, ubuntu is still detecting hardware, hasn't even started the partitioner yet
actually, more like 1/3 of the way thru
huh, wonder why it's so slow
hope it runs acceptably once it's installed
the debian install does seem to take a LONG time
tell me about it
the install on the host box seemed to take forever
I bet I know what it is: bz2
notice how the install is only one CD? :-)
the TNG install is gonna be interesting - it requires you to swap CDs half-way thru
no problem, you have to use the "console" to switch disks
network stats for this box: received 3.5GB today ;-)
(I scp'ed some isos over from my other box, then d/l'ed another one, and am doing two more now)
qemu ... -monitor stdio
yeah, I read something about that in the docs
I'll read again before I start the TNG install
I have to figure out how I'm actually gonna run the farm
today, all slots are alive and running all the time (but usually in the middle of a "sleep 3600")
I dunno if I should have them all alive, or invoke qemu one or two slots at a time
invoking qemu means that sleeping slots won't consume ram, but it also means lots of cycles wasted booting every time
you could use the qemu ram snapshot to stop and start them if you have to
[06:11:12] <cradek> http://timeguy.com/cradek-files/emc/toolpath.png
I guess toolpath hasn't bitrotted too much - it built and ran easily
I think I told you about this
it loads 3d models (dxf) and generates a contouring path over it (can consider various tool shapes)
maybe, but I think I forgot
also supports rotary axis for milling from several angles
that cylinder isn't in the model - it's a configurable area that's not milled away
[06:15:45] <cradek> http://timeguy.com/cradek-files/emc/toolpath2.png
yay, the ubuntu is finally writing to the disk (partition table)
the 2.20 is 3/4 done
wow, that's a shocking difference (but what I've also noticed on real hardware)
and freebsd is much faster than redhat
have you seen OSzoo?
they have pre-installed qemu images for about 20 free OS's
including a couple bsd's and other non-linux
[06:18:03] <jmk_cf> http://free.oszoo.org/
the download page has the list
free, net, and open bsd
reactOS (whatever that is)
as well as mainstream stuff, debian, ubuntu, mandrake, etc
[06:20:20] <cradek> http://timeguy.com/cradek-files/emc/ubuntu-qemu.png
aside from requiring patience, it worked with no problems at all
ubuntu is just beginning to install the base system, 2.20 is done with all packages
any bets which one will have a working X first?
well, it tried
it probed the card, says its 1024x768x16bit
started X, but the dialog where you click to say "I can read this" had fscked text in it (as before)
now I have a choice of "go back and modify the X config" or "quit"
I'd like to choose "continue without X for now"
I wonder if "quit" will actually abort the install, or just abort X?
sorry, I don't remember
I can also go back and the next time it starts X I can click on the button that says "I can read this" even tho I can't
(I installed a lot of rh62 machines but it was a while ago)
I think I'll tell it I can read the display, and see what happens
I should be able to get to a tty and edit the startup to not start X if needed
great, now it seems hung
its back in a ncurses screen, but ignoring input
you still there?
it boots, gives a login prompt, but after a few seconds, proceeds to start X and wait for a graphical login
but the mouse pointer is a square black box, and much (not all) of the text us fscked up
gotta figure out how to get a text mode console
you can probably fix the mouse pointer with Option "sw_cursor"
no immediate guesses about the font problem
you can get to text mode by grabbing the keyboard (click the mouse) then ctrl-alt-F1
(waiting for it to boot)
that gets me a host computer text mode console
clicking in the qemu window, then ctrl-alt-f1, no joy
well reboot again and tell it runlevel 3
and how do I do that? I have to edit a file, and I need text mode to do it
no, just reboot, at the lilo prompt type linux 3
* jmk_cf is learning new things
you're learning OLD things
I remember (I think) what line in which file to change to make it use rl 3 every time
but didn;t know you could do it from lilo
it doesn't change the default, it just boots to a different runlevel
right, I understand
lilo is a one time thing
inittab changes make it stick
hmm, there are freebsd packages that set up chrooted cvs pserver
qemu should have a "--title" option that sets the window title
so when you have multiples running you kno which is which
ok, inittab changed to rl 3
where do I put that Option "sw_cursor"?
somewhere in XF86Config I think
in Section "Device" I think
hmm, it may depend on your "graphics card"
which Driver is it using?
I'm not very familiar with the format of this file
is the driver in the device section, the screen section, or somewhere else
there are two device sections, "Generic VGA" and "My Video Card"
maybe you should look at the log...?
I can't make heads or tails of this file, it has many similar sections
the screen section invokes the My Video Card device
seems like one is chosen at runtime
(--) SVGA: RIVA TNT: Using hardware cursor
where is the log usually located?
might be xdm-something or XF86something
none of the above
oh is this some kde abomination?
nfc where it logs
the machine I have is a normal redhat 6.2
I knew this stuff at one point
I think the X messages just went to the screen
and I did "startx >log"
X -probeonly -verbose >log 2>&1
yeah, its using hardware cursro
device ID "My Video Card"
qemu says it is cirrus logic GD5446 PCI VGA
XFree86 3.3.6 has a clgd5446 driver
this is 3.3.5 (but probably still has the driver)
so, how do I tell it to use that driver?
Option "sw_cursor" (542x/3x/46/6x)
This disables use of the hardware cursor provided by the chip.
it should probe - do you know it is using the wrong one?
the log says the device is My Video Card
it doesn't say anything about what driver its using (or I didn't look close enough)
(--) SVGA: chipset: RIVA TNT
does it have something like that?
PCI: Cirrus Logic GD5446 rev 0
that means I should be asleep!
but neigher of us is allowed to sleep until my X is working
it's using 16 bit color right?
does it detect 2MB of ram?
sorry guys, I'm after 5 hours of sleep ;)
can't see the monitor right
going back ;)
you should leave soon too
do I just stick the Option sw_cursor right in the Device section?
I think so
Config Error: Option sw_cursor
there are also options no_bitblt no_imageblt no_mmio that might help
Corrupted text in terminal window.
This has been reported on non-standard video implementations.
Use the "no_bitblt" option.
to test I just drop the -probeonly option from the cmd line, right?
that'll only give you a blank screen
can I still pass -verbose?
I think so
cursor seems better
no dialogs or widgets yet....
we're assuming it's a video driver problem when it could be a kde problem?
as in "kde's anti-aliased fonts are screwed up"
no, its fine now
just with sw_cursor?
I put in all of those options
here, beat it with this stick
I'll try commenting them out one at a time and see what happens
no_bitblt might be the keeper according to the source
and sw_cursor, right?
not sure if they might be the same problem
commented out no_imageblt and no_mmio, testing
very slow startup
when I think about qemu being software that pretends it's a computer, I'm impressed by the speed
slow but workable KDE
now I'll comment out sw_cursor
got the black box
is today the day the shut down the internet for cleaning?
I'm assuming that sw_cursor is _not_ what fixed the text problem, so I'm not gonna bother commenting it out to test
something else besides SF is busted?
nah, just an old 4/1 joke
well, looks like BDI-2.20b works just fine in qemu
cool, I bet the rest will be easier.
ubuntu is 34% of the way thru the base system install, its gonna be hours and hours before its done
time for bed
I bet bdi4 will be the same (it's similar debian stuff)
so you might have a lot of hours of installing left.
yep, talk to you later this weekend sometime
probably be installing all weekend
when I installed ubuntu in qemu it was hours and hours
same for bdi
Hi alex. Sorry I'm slow today.
no problem, me too :)
Off working on halconfig.
just woke up, and it's 20:30
I've got tabs working.
It looks pretty good.
I really wish sf cvs was working.
seems svn is down too, same machine?
their new versioning system that's running since january
Do you think we need to switch to that and away from cvs?
no, it offers some limited functionality more than cvs
but the switching is hard to do (for the users)
Good. I didn't want to have to stumble through a new software.
most our CVS users can't use CVS fully, including me
so trying something completely new will set us back for a while (that's why we didn't bring it up)
Uh yea you're talking to one right now.
I think we all know that
I wonder about moving cvs to dreamhost?
rayh: I think we can move cvs to somewhere else, then set up a cron job to commit changes to SF too
Okay. That would work for me.
but I would want to talk that through with people who know more :)
I could probably find money to buy deamhost cvs service.
there are free alternatives out there I think
I'm guessing, perhaps a bit paranoid, that sf's parent company is applying downtime as a tool to get support or to move some projects to their commercial service.
rayh: that would be terrible, but it's not excluded
Just the electricity to host 112k projects and 750k + users has to be significant.
rayh: not that significant
that's only a few racks
750k users is not THAT much, probably they have something like 20-30 servers
It's significant by my meager cashflow.
rayh: I know of a hosting company in germany, they have the most % of the market share in europe
1&1 I think they are called
I really don't want us to get caught on a "private" server.
those people have something like 1.5 Million kWh / month electricity consumption
that is a LOT
A small hydro plant or nuke?
heh, regular power
but they need to announce when doing test going offline
because the extra power available, when they are offline, brings the whole county down :D
I did a bit of elec work for a coal fired plant here with 6 10M generators and 2 20M.
yeah, they have a huge generator too
about 6000l/hour diesel :D
consumption I mean
Pay that bill and 1.5 eu per l
yeah, or even supply that much diesel
does 1&1 host free projects?
don't think so
but they are pretty cheap, anyways that was just a comparison to SF
Makes the sf bunch seem small.
they have 30000 servers (I just read)
brb start cooking the potatos.
how do you cook them?
Customers:5.08 million customers on paid services
Domains:5.3 million domain names are registered and handled by 1&1
Connectivity:25,000 MBit external Carrier Class Connectivity
Our Own Backbone
that really makes SF look small :D
boiled and mashed today with sausage gravy.
Boy I guess that makes most of the server farms I've seen seem tiny.
rayh: told you they are big :)
anyways, had some nice potatos today, boiled till they are mashy, but not mashed
with some oil & spring onions
mixed together, those are great (even when cold)
Oh yea. I'm drooling.
Do you know who built stepper_inch_net.hal and it's ini?
Just tried it and it fails.
I suspect #loadrt motmod (not loaded by loadrt, no args saved)
Yep that's it.
never seen stepper_inch_net.hal
bet it's yours ;)
Oh Maybe it is mine.
Darn I should keep my mouth shut.
rayh: you're under friends here, so it won't matter :)
Well I've got the ini file down to 90 lines.
oh, that's a change ;)
And the net list seems to work great.
I'll need to add some stuff to halconfig in order to handle stuff like thread timing and such.
does iocontrol use the ini's emcio cycle_time?
I'm thinking so because it's a user space thing.
yes, I think so
and halui probably will use the same one
btu here's my thinking: if we want to make the ini's short, we should put the defaults into the code (some defaults are already there)
and take the stuff out of the ini, let users who want to change the cycle time (not sure there are any) add the stuff back
Right. What I did was build the netlist from halcmd save.
and then removed all of the variables that were now hard coded into .hal.
Now I'm working on ways to get access in halconfig for all of those.
I can see halui using the same loop timing.
* alex_joni is happy ;)
just added another thing to linuxcnc.org :D
speaking of things on linuxcnc.org, it sure is hard to find stuff
well...for some reason I got the old linuxcnc homepage for a while last night (in browser cache maybe)
probably DNS cache
all of the links on it were busted
(as expected I guess)
but trying to find the same data on the new site wasn't easy
its down 2 or 3 levels instead of directly accessible
and the BDI page seems to be gone completely
the BDI page was gone for a month or so
paul took it down on the old server aswell
using his ftp access
he set up a forwarding to his new home
I guess thats his right...
I was looking for isos
he has some mirrors on his new home
I got what I needed from wildrice and your mirror
jmk_cf: seriously now, I spend quite a time on this new page stuff, gotta let me know what you don't like
so I can make it better
it just seems harder to navigate
its hard to describe in detail
depends on what you expect
you might be used to some things..
thats probably the problem - I expect the old site
I think it's a bit clearer structured for the new users
I wonder how google will do with it?
they will eventually update links
woot! finally (after three tries) I got a good download of BDI-Live_rc46
but it takes a month or two
will google be able to link deep into the site,or does the CMS make that complicated?
whats with all the copies of you on the other channel?
it's an interface out of www.linuxcnc.org
setting it properly up now
I mean client out of the page
for users that don't have IRC clients installed, or don't know how to use them
so you can IRC with just a web browser?
yes, but you need java
I wonder if my work computer has java
if I can talk on IRC during the day, I'm doomed ;-)
try it out : http://www.linuxcnc.org/index.php?option=com_mospjirc&Itemid=8&lang=en
it wants me to download a plugin
it should look like this: http://www.robcon.ro/emc/irc.linuxcnc.org.jpg
I only get a message about a certificate
but probably you don't have java on your pc :)
I hope not ;-)
I'll try it from work some day
it's under Contact Us
from here I'll use a real IRC client
jmk_cf: I know YOU will use a real IRC client, so will I
but I thought this might help new users, who just want to contact us
18521 john 25 0 281m 273m 2404 R 97.0 36.1 355:31.53 qemu
can you com to board for a bit?
rayh: look at http://www.robcon.ro/emc/irc.linuxcnc.org.jpg
143kB it will load eventually :(
Got the first couple lines.
* alex_joni447 likes it :)
Interesting. You could see the current IRC right from linuxncnc.
rayh: and also talk, like I am doing now
rayh: the only downside is the need of java
Yes. That would be a very nice addition to the site.
it's already there
as you can see, I am using it :)
java on the home machine or on the server?
on the home machine
what's the url?
Could joomla test for that and put up an error message.
if it doesn't find what it needs on the home machine.
the browser gives you the proper error if you don't have what it needs
'Ah okay. I say we put it in then.
er you put it in then.
rayh: I already did :)
you can get to it from the Contact Us page
I mean, I already added because I needed to test
would have removed it if it were bad
ubuntu firefox won't automatically install java
cradek: too bad, guess I need to keep my eyes open for one that doesn't use java
how about flash ? :D
manual install it says.
flash would probably be fine (but I don't know if flash can make network connections or not)
seems there is some Flash IRC client, but it needs a gateway to work
[19:56:50] <alex_joni447> http://www.wdmedia.org/WFIC/index-en.html
You know I'm pretty certain that the ubuntu users will not need the link to IRC from the web site.
rayh: I bet so too
well that's true
freenode is on the default list of irc networks
they do have an IRC client in ubuntu
they'll just use xchat
it may even be *the* default
cradek: on ubuntu?
I used xchat. you guys have helped me through it on the actual machine.
I find the linuxcnc IRC interesting for cases like me at work, or on the road, where all I have access to is somebody elses PC
jmkasunich: hope it will be of some use
jmkasunich: that's what putty is for (ssh to a real computer)
1) my computers are firewalled to keep outsiders from getting into them
2) work is firewalled to keep me from getting out
1) opening ssh is relatively safe
you need a firewall with ssh
3) I may be forbidden from sticking putty on the work box
jmkasunich: you don't need to stick it
3) you can run it right off the putty web site (no install needed)
I often run it out of the browser
obviously I have lots to learn
I've done that a lot of times on public PCs
putty is really nice if you're stuck with windows
google & search putty & feeling lucky
4) I don't have a static IP
it doesn't change often (if ever) but...
at home? you can use dyndns
does it change often? I don't have one either, but it stays the same for years
* jmkasunich 's head spins
I just want to write code, not become a sysop
jmkasunich: don't worry ;)
let others bother :D
* alex_joni loves iptables
I just set up a fake ftp server with DNAT
well, opening a hole in my NAT firewall for ssh, and setting up keys or whatever is needed to let me log in from elsewhere, that is sysop stuff that only I can do
depends on the NAT
at home I have one of those wireless & ethernet firewall & router
besides, I turn my computer off when I'm not home (except on weekends)
with http interface, makes it easy for the basic user to adapt
does your isp give you a shell account?
there's also shell.sf.net which you can use anytime (no sysoping required)
IRC on the linuxcnc website will let me talk on the channel without my home computer even being on
no shell account
checking to see if I can irc from shell.sf.net...
I don't know why we're even having this discussion
lol, because we digress as always
I pointed out a reason why I might use alex's new IRC, and you guys are trying to convince me I don't have to
the main issues is SF's CVS right now :(
not an issue to talk about
but it's there
I looked into setting up a cvs server on my machine, but it's not at all trivial to do it securely
1) move the entire project to some other hosting site (Gna, etc)
2) have a backup repository somewhare and figure out how to sync them
3) move just cvs, not the entire project
2&3) might be a cron job that commits to SF
I mean it checks out changes on the new hosting, and checks them in to SF as well
to keep them synched
moving cvs (vs. a backup cvs) still leaves us with a single point of failure
you can't fix that with cvs - that's its architecture
the best way would be an rsync on the repository
but SF doesn't offer that support
I loathe switching, but maybe we should consider subversion, which is more distributed
cradek: seems svn is down aswell.. some reported
I can envision a script that does an up from one server, copies all the files (not the CVS dirs) to another checkout tree, and commits on that tree
but I think that doesn't stop all work
jmkasunich: it's not that simple: think add/delete
jmkasunich: btw, news update :)
there's a v2_0_branch now
I'd rather not learn subversion.
and TESTING has been moved to that branch
add/delete could probably be handled, but branching would be too much
I've got more than enough on my plate for the next year.
rayh: I'm the same way, but still I admit it seems to offer important benefits
how is svn more distributed? I thought there was still a central repository?
I just think we need to find cvs hosting that doesn't suck.
I don't doubt but what that's true.
jmkasunich: I'm NO expert, but I think you can pull/push from anyone's repository and sync everything later
I never touched svn, but then again I haven't really touched CVS before emc either
* cradek waves his hands
cradek: how about that stuff that jepler put up a while ago?
he's copying FROM sf cvs TO his svn
and I've never looked at it
ok, so he did it one way ;)
maybe the same could be inversed
then we move to HIS svn :D
sadly I don't want to learn svn either, even though I'm sure it's better
yeah, I can understand why
dreamhost has a ruby setup if there were a ruby irc.
seems there are a lot of source code hosting companies out there
btw, I probably will leave for a year or more (got a scholarship), not sure I will be able to keep up with emc then :(
Congrats. Where you going?
$25/month for each 5 users, plus $25/month for each 100Megs
rayh: just a bad april joke :)
$10/month or $100/year for open source projects, unlimited users
Darn. I had an april fool joke once. Bacterial mennengitis.
jmkasunich: 100MB disk space sounds kinda scarse
and "optional anonymous access" sounds like additional money
our cvs tree is > 100MB already
I'm sure you can get more space, dunno about cost
dreamhost has CVS too
so we could setup cvs.linuxcnc.org
seems like lots of places have it
I hear dreamhost doesn't have pserver
I just sent a note asking for details.
if we can find a way to do anoncvs, I think we should immediately start the move.
True. Or if they have levels of developer access like testing v committing.
I think we need true anon (though not necessarily pserver)
dreamhost gives you shell access right?
[20:34:30] <jmkasunich> http://www.mythic-beasts.com/support/topic_cvs.html
read down to vCVS
seems to be sw that lets you host multiple user cvs on a shell acct (including anon)
source is available, so you don't neccessarily need to go with that company
(or I could be confused)
[20:35:54] <jmkasunich> http://www.ex-parrot.com/~chris/software.html#vscvs
thats the sw
[Global Notice] Hmmm. You know, I could do an April Fool's Day joke, but maybe instead we should make this a group effort. Let's all just go to irc.oftc.net channel #oftc, and tell them that freenode has shut down and has repointed its servers to OFTC, and ask if they're going to be adding more facilities to handle the load. 8)
[Global Notice] Okay guys, thanks for the fun, let's let those poor guys on OFTC off the hook. :)
qemu has spent 491 minutes on the ubuntu install (and that is only after the reboot, it spent a load before)
[Global Notice] Okay guys, give the folks on OFTC a rest, it was a good joke, time to go home. :)
lol, I followed lilo's prank ;)
man that place was flooded :D
rayh is now known as rayh-away
[Global Notice] Hi all. On behalf of my self and all 200 or so people who participated in that April Fool's prank on OFTC, please accept our apology. I know that when you're the person pranked, it doesn't seem all that funny. So, apologies and bygones!
[Global Notice] Erm, assuming anyone from OFTC is still here. 8)