#emc-devel | Logs for 2006-11-17

[01:43:45] <jmkasunich> jepler: how bad do you want libgtk2.0-dev?
[01:44:01] <jmkasunich> 74 upgraded, 123 newly installed, 66 to remove and 357 not upgraded.
[01:44:01] <jmkasunich> Need to get 77.1MB of archives.
[01:44:01] <jmkasunich> After unpacking 68.2MB disk space will be freed.
[01:49:02] <jmkasunich> that is just fscked up.... among other things it wants to install cupsys-common...what the hell does that have to do with GTK?
[01:49:35] <jmkasunich> ..... I just realized why
[01:49:47] <jmkasunich> remember - this is a 200MPz Pentium with 128M of RAM
[01:49:59] <jmkasunich> in order to make it work at all we needed to do a text only install
[01:50:13] <jmkasunich> so installing GTK means installing the entire Xserver
[01:50:23] <jmkasunich> ain't gonna happen, sorry
[01:51:05] <jmkasunich> the new VM based compile farm will be better
[01:51:23] <jmkasunich> and on that note, /me shuts down to install another 512M of RAM ;-)
[02:02:29] <jmkasunich> jmkasunich@ke-main-1006:/proc$ cat meminfo
[02:02:29] <jmkasunich> MemTotal: 906536 kB
[02:02:39] <jmkasunich> seems that should be a gig, not 900M
[02:29:23] <jepler> jmkasunich: just so that classicladder is tested on a system besides ubuntu ..
[02:29:31] <jepler> if it can't happen then so be it
[02:30:55] <jmkasunich> I'm gonna set up a VM for BDI-4.50 one of these days, hopefully soon
[02:32:03] <jmkasunich> of course, I need to revisit how I do farm stuff
[02:32:04] <jepler> at one time the default ubuntu kernel was limited to under some amount under 1 gig
[02:32:18] <jmkasunich> ahh...
[02:32:43] <jmkasunich> I was thinking maybe that count didn't include kernel memory, or maybe it didn't include memory that was reserved for RTAI
[02:32:44] <jepler> but that was way back on 'hoary hedgehog' ...
[02:33:01] <jepler> (the one before breezy? not sure)
[02:33:28] <jepler> in 2006 a 1GB RAM limitation will affect a lot more systems than it did in 2004
[02:33:34] <jmkasunich> yeah
[02:33:49] <jmkasunich> I just got 512M for $39.95 (after rebate)
[02:34:07] <jmkasunich> just got a 1G USB stick for $15.99
[02:34:12] <jmkasunich> mind-boggling
[02:36:07] <jepler> [17179570.612000] Memory: 498920k/514944k available (2115k kernel code, 15504k reserved, 595k data, 332k init, 0k highmem)
[02:36:11] <jepler> here's what this laptop says
[02:36:33] <jmkasunich> I haven't looked at dmesg yet
[02:37:02] <jmkasunich> [ 53.606110] Memory: 899616k/917504k available (1619k kernel code, 17268k reserved, 493k data, 220k init, 0k highmem)
[02:38:31] <jepler> that seems a bit low
[02:38:33] <jepler> I wonder why
[02:38:50] <jmk-vm02> this vm has 256M, dmesg reports 251140/262144K, proc/meminfo reports 256784
[02:39:26] <jepler> CONFIG_1GB=y
[02:39:34] <jepler> that's from /boot/config-*
[02:39:43] <jepler> I think that may mean you're limited to (somewhat less than) 1GB RAM
[02:39:49] <jmkasunich> I ran memtest86 (very briefly) when I rebooted, and it reported the full 1G
[02:40:02] <jmkasunich> yeah, that sounds plausible
[02:40:24] <jmkasunich> ISTR a RTAI limitation
[02:40:29] <jmkasunich> I'm running a RT kernel here
[02:40:54] <jmkasunich> where is that config flag? I want to see if the non-rt breezy kernel has it
[02:41:04] <jepler> /boot/config-(your kernel version)
[02:41:07] <jepler> this is the stock ubuntu kernel
[02:41:14] <jepler> dapper
[02:42:49] <jepler> bbl
[03:31:19] <jmk-vm03> is this normal when installing emc2-axis to breezy for the first time:
[03:31:20] <jmk-vm03> Setting up linux-image-2.6.12-magma (cjr12) ...
[03:31:20] <jmk-vm03> Hmm. The package shipped with a symbolic link /lib/modules/2.6.12-magma/source
[03:31:20] <jmk-vm03> However, I can not read it: No such file or directory
[03:31:20] <jmk-vm03> Therefore, I am deleting /lib/modules/2.6.12-magma/source
[03:38:10] <jepler> yeah I think so
[03:39:03] <jmkasunich> oh damn... I think I broke the VM
[03:39:26] <jmkasunich> I had copied the VM disk image from my non-rt breezy install
[03:39:37] <jmkasunich> that meant it had the same hostname as the other one
[03:39:43] <jmkasunich> so I changed etc/hostname
[03:39:56] <jmkasunich> then I rebooted (after installing the magma kernel)
[03:40:11] <jmkasunich> now sudo doesn't work, because the old name is still in hosts
[03:40:25] <jepler> weird
[03:40:28] <jmkasunich> help?
[03:40:40] <jepler> hit ESC to get the grub boot prompt and choose recovery mode
[03:40:42] <jmkasunich> how is one _supposed_ do change the hostname on a box?
[03:40:47] <jepler> but then after that I'm not sure what you do
[03:40:59] <jmkasunich> well, I can change /etc/hostname back to what it was at least
[03:41:11] <jmkasunich> where's cradek when he's needed
[03:41:36] <jepler> does sudo -k / -K remove the outdated info?
[03:43:07] <jmkasunich> I'll try it
[03:43:14] <jmkasunich> grub didn't respond to esc
[03:43:26] <jmkasunich> dunno if thats a VM pecularity or what
[03:44:02] <jepler> sudo: unable to lookup mybox via gethostbyname()
[03:44:03] <jepler> ^^ this?
[03:45:12] <jepler> if that's the error, then the fix is to put an entry for your new hostname in /etc/hosts while in recovery mode
[03:45:21] <jmkasunich> yes, thats the error
[03:45:55] <jmkasunich> so I just have to figure out how to get into recovery mode
[03:46:01] <jepler> hit ESC at the grub boot prompt
[03:46:08] <jmkasunich> I did
[03:46:21] <jmkasunich> it didn't go to the menu, just booted as if I didn't do anything
[03:46:30] <jmkasunich> dunno if vmware filtered the esc or something
[03:46:41] <jepler> click inside the vmware window so it gets keypresses
[03:46:54] <jepler> otherwise they don't go to the VM they go to the vmware viewer application
[03:47:08] <jepler> and hit it 2 or 3 times for good measure
[03:47:19] <jmkasunich> trying again
[03:48:45] <jmkasunich> heh, that was the trick
[03:48:48] <jmkasunich> thanks
[03:49:35] <jepler> hope the /etc/hosts thing works too
[03:50:05] <jmkasunich> I suspect it will
[03:50:28] <jmkasunich> cause the old name was in there, I can see how that would confuse things
[03:52:01] <jmkasunich> I greped in /etc for the old name
[03:52:23] <jmkasunich> it was in hosts, and in a few ssh keys
[03:52:32] <jmkasunich> I will delete and regenerate the keys later
[03:52:49] <jmkasunich> didn't see it anywhere else, so I bet this will work...
[03:53:28] <jmkasunich> it seems happier
[03:54:32] <jmkasunich> yep, working
[03:54:34] <jmkasunich> thanks
[03:57:44] <jmkasunich> goodnight jeff
[04:06:30] <jepler> goodnight
[15:00:32] <jepler> I wonder if I should make building on a 64-bit system (even sim) require a special configure flag, which notes that it's experimental and may not work properly
[15:03:44] <SWPadnos> are you seeing many problems?
[15:04:22] <cradek> back to civilization swp?
[15:04:26] <SWPadnos> I don't really see any with sim (on a 64-bit machine)
[15:04:35] <SWPadnos> temporarily. we're leaving tomorrow night :)
[15:04:52] <jepler> I suspect that the whiner's latest report is on a 64-bit machine
[15:05:11] <SWPadnos> at least this time we'll be in the US, so my phone will work, and we'll have semi-reliable internet access (vs. quasi-reliable)
[15:05:15] <cradek> oh really! that's a new twist.
[15:05:20] <SWPadnos> hmmm - interesting
[15:07:19] <jepler> though I see that halvcp also works fine on my 64-bit system
[15:07:47] <jepler> Nov 17 11:27:38 localhost kernel: halvcp[16136]: segfault at 00002ad868280008
[15:07:48] <jepler> rip 00002ad86531d6ca rsp 00007fff458a5d80 error 6
[15:07:58] <jepler> this is the kind of message I've only seen on my 64-bit systems -- the addresses are way too big
[15:08:02] <jepler> for 32-bits
[15:08:12] <SWPadnos> true enough
[15:08:20] <alex_joni> bbl
[15:08:35] <cradek> he may be using an odd compiler too, last I knew he still used 2.9x for everything
[15:08:44] <jepler> "hal/halmodule.cc:430: error: invalid conversion from 'int (*)(PyObject*)'
[15:08:44] <jepler> to 'Py_ssize_t (*)(PyObject*)'
[15:09:18] <jepler> this is a change in python2.5 -- they made some changes to allow containers to have more than 2^31 items, but it's got some compatability problems at the source level.
[15:09:53] <jepler> I have NFC about the xgettext configure error but I'm sure it's another "weird version" problem
[15:11:31] <SWPadnos> hmmm - I wonder if a realtively polite response to his email would help. something along the lines of "it lkooks like you're using a 64-bit system with python 2.5, have you tried XX and YY instead?"
[15:12:02] <SWPadnos> and the standard "please provide us with the details of your system config ..."
[15:12:14] <jepler> I dunno, I'm in capable of writing such a response. I know he is perfectly capable of giving a good problem report (including disclosing that he's using this kind of a weird system) so I conclude he's just being an ass.
[15:12:20] <cradek> SWPadnos: in the past, polite responses only cause more rants
[15:12:57] <SWPadnos> well, that's fine - it only serves to point out to those who don't realize he's being an ass, or it gives us the information
[15:13:12] <cradek> he's also capable of fixing some or all of these simple problems and sending patches, but it's clear his goal isn't to get them fixed.
[15:13:22] <SWPadnos> that's obvious
[15:13:54] <SWPadnos> well, to us at least
[15:14:02] <skunkworks> I can't believe he is such a good programmer for being 6 years old.
[15:14:11] <SWPadnos> heh
[15:14:31] <cradek> he's a smart guy but something has gone really wrong with him. maybe he's not healthy or something.
[15:14:55] <jepler> *sigh*
[15:14:55] <cradek> maybe one of these days he'll be back to the old paul.
[15:15:17] <jepler> when I tire of working on emc2, I hope I don't stick around and crap on everyone else
[15:15:32] <SWPadnos> me too :)
[15:16:36] <SWPadnos> that is - I hope you don't stick around to crap on people when you get tired of working on emc2 ;)
[15:16:56] <cradek> ha
[15:38:50] <SWPadnos> hi Ray
[15:39:19] <rayh> HI Steven
[15:59:02] <SWPadnos> couldn't stay away, huh :)
[16:00:24] <rayh> Nope. Seems my ISP has decided the way to minimize bandwidth is to cut off any connection that had downloaded 500k
[16:00:53] <SWPadnos> yeah - I'm sure your dial-up is a real bandwidth hog on their backbone connection :)
[16:01:30] <rayh> You bet. All those EMC2 updates.
[16:01:43] <SWPadnos> it's the 27KBaud that gets them
[16:02:17] <SWPadnos> just wait - without Net Neutrality, you'll need to pay more to prioritize that 3KBytes/second
[16:03:06] <rayh> I think I pay now cause folk see the dribble and decide it's not worth supporting.
[16:03:14] <SWPadnos> heh
[16:03:27] <rayh> Stuff like mapquest and such.
[16:03:33] <SWPadnos> oh - you meant the small amount of data. (I thought you were drooling ;) )
[16:03:36] <rayh> Big nfw.
[16:03:58] <rayh> I am. But just can't get there from this low spot.
[16:04:00] <SWPadnos> heh
[16:05:59] <rayh> the microwave guy promised me spring, but I don't think I can hit that tower either. Even though it's 1.5 miles away.
[16:06:38] <SWPadnos> hmmm - where is he relative to your house? toward Crystal Falls?
[16:07:04] <SWPadnos> err - where is the tower ... ?
[16:07:20] <rayh> It's east and north a bit.
[16:07:36] <rayh> Cell tower and I don't get cell here either.
[16:07:50] <rayh> Probably still need a 120 foot tower here.
[16:08:01] <SWPadnos> oh, so kind of a bit further down the main road from Swante Aho road? (away from CF "downtown")
[16:08:11] <rayh> right.
[16:08:28] <rayh> about a half mile north of Swante Aho.
[16:08:42] <SWPadnos> ah
[16:08:49] <rayh> I can see it from the ridge east of me.
[16:09:01] <SWPadnos> ok, but not from your rooftop ...
[16:09:20] <rayh> Suppose I could 'doze a path through the woods.
[16:09:42] <SWPadnos> well, for the shorter run, you can just lay some cable
[16:10:11] <SWPadnos> standard ethernet can go 100M, and you can go further if you run through the calculations (ie, 200M should be OK for a single run to a switch)
[16:11:04] <SWPadnos> and you can get fiber that goes kilometers
[16:11:15] <SWPadnos> of course, if you don't own the land, there could be problems ;)
[16:11:28] <rayh> Might talk to the neighbor about a common access point.
[16:12:43] <SWPadnos> yeah - you're kind of lucky in that there doesn't seem to be much interest from the phone/cable providers to run high speed connections to your area
[16:13:20] <SWPadnos> that's the big risk for "neighborhood" projects - you spend a few $k for equipment, then the cable company comes in and nobody wants to pay their share any more
[16:13:23] <rayh> the cable guys tell me they need 10 customers per mile.
[16:13:48] <SWPadnos> oh, well you have 10 miles per customer ...
[16:14:02] <rayh> About that.
[18:16:33] <Lerneaen_Hydra_> Lerneaen_Hydra_ is now known as Lerneaen_Hydra
[19:53:05] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[20:38:40] <Jymmmm> Jymmmm is now known as Jymmm