T2 IRC Log: 2010-08-16

This is the log as captured by an IRC bot in the channel. The statements are those of the individual people and might not neccessarily reflect the policy and legal rules as set forth by the T2 SDE Project.

« prev | next »

--- Log opened Mon Aug 16 00:00:49 2010
02:17 -!- sparc-kly [~mubex@adsl-64-237-249-170.prtc.net] has quit [Read error: Connection reset by peer]
08:42 < CIA-32> aldas * r37278 /trunk/package/network/collectd/collectd.desc: * updated collectd (4.6.4 -> 4.10.1)
08:52 -!- Baldzius [~aldas@87-198-133-94.static.ptr.magnet.ie] has joined #t2
08:52 < Baldzius> moin
09:52 < CIA-32> aldas * r37279 /trunk/package/graphic/inkscape/inkscape.desc: * updated inkscape (0.47 -> 0.48.0)
10:03 -!- mjungwirth2 [~mjungwirt@213.174.234.68] has joined #t2
10:04 -!- mjungwirth [~mjungwirt@213.174.234.68] has quit [Ping timeout: 240 seconds]
10:20 < CIA-32> aldas * r37280 /trunk/package/security/libfwbuilder/libfwbuilder.desc:
10:20 < CIA-32> Roger Mason :
10:20 < CIA-32> * updated libfwbuilder (3.0.6 -> 4.1.0)
10:20 < CIA-32> aldas * r37281 /trunk/package/security/fwbuilder/fwbuilder.desc:
10:20 < CIA-32> Roger Mason :
10:20 < CIA-32> * updated fwbuilder (3.0.6 -> 4.1.0)
10:25 < CIA-32> aldas * r37282 /trunk/package/editors/vim/vim.desc: * updated vim (7.2 -> 7.3)
10:27 <@rxr> moin Baldzius
10:28 <@rxr> thanks for preserving the date and dl Url in the fwbuilder updates :-)!
10:28 < Baldzius> hey rxr
10:28 < Baldzius> yeah spotted that myself
10:28 < Baldzius> sending email
10:28 < Baldzius> have you seen my email?
10:28 <@rxr> yes
10:29 <@rxr> go ahead disable them, ...
10:29 < Baldzius> ok
10:29 <@rxr> I imagined one or the other built for me, but we can always re-enable when someone really needs it
10:30 < Baldzius> yep
10:30 < Baldzius> btw , I remember we talked about it but forgot, what's the story with fortran?
10:30 <@rxr> no idea
10:31 < Baldzius> there are 6 or 7 packages failing because of this
10:31 <@rxr> I just run a test setting F77 to gfortran instead of the f77
10:31 < Baldzius> looking for fotran and can't find
10:31 <@rxr> maybe they just changed their program name
10:31 < Baldzius> I think it's been a while, for couple of years or so
10:31 <@rxr> rogermason also used gfortran explicitly in some new scifi package
10:31 <@rxr> yeah, thing is we rarely have fortran stuff
10:32 <@rxr> and I guess fftw3 is mostly checking accidently in the configure :-)
10:32 <@rxr> not really needing it? ... maybe
10:33 <@rxr> if gfrotran does not work as a replacement we may need some other configure switches to gcc?
10:42 <@rxr> it's really depressing to see how few cents google adwords bring in :-(((
10:43 <@rxr> Baldzius: I should have fortran issue fixed in some minutes
10:43 <@rxr> looks like the only issue is the g77 symlink is no longer there
10:46 < Baldzius> kewl
10:47 <@rxr> oh, btw.
10:47 <@rxr> the dhcp fix shoudl be merged to 8.0
10:48 <@rxr> (theoretically even any older branch, but guess few care, use the old brancehs as it, rather cherry pick from latest trunk, branch, ...)
10:48 <@rxr> r37233
10:48 <@rxr> very surpsing that this was supposedly hidden since rocknet first appears soem decade ago in ROCK Linux, ...
10:48 <@rxr> guess few people use profiles, excessively
10:55 <@rxr> google webmaster toosl are also strange
10:55 <@rxr> totally lagging
10:55 < Baldzius> yeah noticed that, had an issue myself
10:55 < Baldzius> always used static ip
10:55 < Baldzius> I will merge that
10:55 <@rxr> the "Diagnostic" section lists super flat site load times, and the "Labs" still speak fo 2s load times, "faster than 67%" of the sites, ...
10:56 <@rxr> aside that the stupid "Site perforamnce" estimate does only show "something" for our most prominent sites (e.g. http://t2-proejct.org)
10:56 <@rxr> still what I see does not correlate with either their own "Craw stats", nor my own tests, and is outdated, lagging like h*ll
10:57 <@rxr> these webmaster tools are one of the most broken things at Google
10:57 <@rxr> also the adwords management interface
10:57 <@rxr> it is so arcane, and that for years
10:58 <@rxr> wonder's why do nto beef up teir nearly only source of income, letting it rotting away, while spending all resources to beef up the things that do not bring a cent, are discontinued every few years, ...
10:58 <@rxr> if you can see this graph, that is the craw stat time of one of our sites:
10:58 <@rxr> https://www.google.com/chart?cht=lc&chc=wmc&chd=s:________________feASMIHRVcTJHWIazPFgILdToiSICCJECCCC5QPPOQMAPQOCVKCCCACMCCMEVCCCCOQQQQCCCCCCCCCDCCCCCCCCCC_________________&chds=0.0,61.0&chs=500x150&chma=70,0,0,0&chco=800000&chm=R,666666,0,0.25,0.251%7CR,666666,0,0.5,0.501%7CR,666666,0,0.75,0.751%7CB,eeeeee,0,0.0,0.0&chxt=y,x&chxr=0,0.0,100.0%7C1,0.0,100.0&chxl=0:%7C0%7C350%7C700%7C1:%7CMay%7CJun%7CJul%7CAug&chxp=0,0.0,50.0,100.0%7C1,12.5,37.5,62.5,87.5&chxs=0,666666,11.0,1.0%7C1,666666,11.0,0.0&sig=ZLRmfgZq1yVVt1hJBj7-qb379Eg
10:59 <@rxr> that is the outdated site performance graph:
10:59 <@rxr> https://www.google.com/chart?cht=lxy&chc=wmc&chd=s:A9,AA,A9,kk,A9,99,dfhjl,7uuxx&chds=0.0,1.0&chs=750x150&chco=000000ff,dd3c10ff,000000ff,6489d0ff&chls=0.0,1.0,0.0%7C1.0,1.0,2.0%7C0.0,1.0,0.0%7C4.0,1.0,0.0&chg=100.0,101.0,1.0,0.0&chm=tFAST,c0c0c0ff,0,0.0,12.0,0,vh:350.0:35.0%7CtSLOW,c0c0c0ff,2,0.0,12.0,0,vh:350.0:-24.0%7Cb,efffeac0,0,1.0,0.0%7Cb,ffefeac0,1,2.0,0.0&chxt=x,y,r&chxl=0:%7CMay%7CJun%7C1:%7C%7C0.5%7C1.0%7C1.5%7C2.0%7C2.5%7C2:%7C%7C%7C%7C%7C%7C&sig=YiVrnFsyfFEWuY7Sa64h437FF1Q
10:59 <@rxr> guess this cryptic graph Urls are not entirely cut'n paste friendly, hm?
11:00 < Baldzius> bad request
11:00 <@rxr> whatever
11:01 <@rxr> == 10:55:33 =[5]=> Building scientific/fftw3 [3.2.2 9.0-trunk].
11:01 <@rxr> == 08/16/10 11:01:40 =[5]=> Finished building package fftw3.
11:02 <@rxr> another funny thing is the "site performance" suggestion:
11:02 <@rxr> "Save up to .... 1 DNS lookup"
11:03 < rogermason> How do.
11:03 <@rxr> yeah, funny thing Gogole, it's youw own stupid adsense DNS lookup you are marking on my sides, what would need need another ext. reference for otherwise???
11:05 < CIA-32> rene * r37283 /trunk/package/develop/gcc/parse-config: * fixed gcc fortran support by using gfortran, instead of the no longer existing g77
11:06 < CIA-32> rene * r37284 /trunk/package/develop/gcc/wrappers.in: * fix gcc wrapper creation, wrap gfortran, instead of the no longer existing g77
11:06 <@rxr> Baldzius: you could merge these 2, if they work for you -^
11:06 < Baldzius> thanks, I will
11:06 <@rxr> and not this one: -v
11:06 < CIA-32> rene * r37285 /trunk/package/develop/gcc/gcc.conf: * removed an old-style, versioned kcc (kernel compiler) symlink
11:06 < Baldzius> ok
11:07 <@rxr> the third may break something, though I hope it does not
11:11 < CIA-32> rene * r37286 /trunk/package/perl/perl-dbi/perl-dbi.desc: * updated perl-dbi (1.59 -> 1.613)
11:11 < CIA-32> rene * r37287 /trunk/package/network/whois/whois.desc: * updated whois (5.0.6 -> 5.0.7)
11:11 < CIA-32> rene * r37288 /trunk/package/scientific/paraview/paraview.desc: * updated paraview (3.8.0 -> 3.8.1)
11:17 < Baldzius> I am going to revert autoconf last commit
11:17 < Baldzius> for me autoconf 2.66+fix works fine on 8.0
11:17 < Baldzius> I believe that should be the same for trunk
11:17 <@rxr> ack
11:17 < Baldzius> in worst case scenario we can revert back to 2.65 if complains keep coming
11:23 < CIA-32> rene * r37289 /trunk/package/develop/mpich2/mpich2.conf: * removed F77=gfortran hotfix quirk from mpich2, now properly fixed globally in T2
11:24 < Baldzius> rxr: http://article.gmane.org/gmane.comp.t2.devel/4697 you missed that one?
11:25 <@rxr> I was under the impression I applied it
11:26 <@rxr> ah, I have the patch indeed lingering in my nearly clean test build tree on another VM
11:26 <@rxr> strange
11:26 <@rxr> I thought I comitted it
11:26 <@rxr> will do now
11:27 < CIA-32> aldas * r37290 /trunk/package/base/autoconf/ (autoconf.desc hot-fix.patch): * reverted r37052, going back to autoconf 2.66+fix , 2.67 is strict and causes some havoc for us
11:28 < CIA-32> rene * r37291 /trunk/package/develop/ (apr/apr-config.patch.cross apr-util/cross-bdb.patch):
11:28 < CIA-32> "Dmitry S. Vasilchenko" :
11:28 < CIA-32> * further fixed up apr and apr-util to cross compile
11:29 < rogermason> Hello Baldzius. I'm having a bit of trouble getting cryptsetup working. On reboot: fsck.ext3: No such file or directory while trying to open /dev/mapper/home. /dev/mapper/home: The suberblock could not be read...
11:38 < Baldzius> did you get swap to work?
11:41 < rogermason> yes. And I mounted /dev/mapper/sda4 on /home/rmason manually and copied a file to it. Then I edited fstab and rebooted.
11:42 < Baldzius> but it looks for /dev/mapper/home
11:42 < Baldzius> seems like you have mixup with names
11:42 < rogermason> sorry, bad typing, I meant /dev/mapper/home
11:46 < Baldzius> I think I see a problem in my howto
11:46 < Baldzius> wonder how I missed it
11:46 < Baldzius> as did this on live system
11:46 < Baldzius> home partition entry is missing in crypttab
11:47 < Baldzius> it doesn't open device, that's why it's missing on boot
11:47 < Baldzius> but before reboot you could create filesystem on it? right?
11:47 < rogermason> yes
11:48 < Baldzius> give me couple of minutes
11:48 < Baldzius> I will try to recall the details
11:49 < Baldzius> too bad I wiped out my previous t2 instalation
11:50 < rogermason> Looking at crypttab, I'm guessing it should have 'home /dev/sda4 defaults'
11:54 < Baldzius> I am not sure about options, but yes
11:55 < Baldzius> you can remove defaults
11:55 < Baldzius> try without it
11:55 < Baldzius> it's for swap basically (out of box feature)
11:55 < Baldzius> when you can specify cipher you like
11:56 < Baldzius> this won't work for luks device
11:56 < Baldzius> so home and /device
11:56 < Baldzius> should do the trick
11:56 < rogermason> OK. Suppose I'll need to boot using usb key as / is mounted ro
12:02 < Baldzius> won't work if init doesn't remount / as rw
12:03 < Baldzius> or /dev in this case
12:06 < rogermason> Added 'home /dev/sda4' to /etc/crypttab but I still see the same error on reboot.
12:07 < Baldzius> it asks you a password?
12:09 < rogermason> No. fsck.ext3: No such file or directory while trying to open /dev/mapper/home
12:10 < rogermason> /dev/mapper/home: The superblock could not be read or does not describe a correct ext2 filesystem...
12:10 < Baldzius> ls -la /dev/mapper
12:11 < rogermason> home is not listed
12:13 < Baldzius> try manually opening it
12:13 < Baldzius> cryptsetup luksOpen /dev/sda4 home
12:14 < CIA-32> aldas * r37292 /trunk/package/network/zabbix/zabbix.desc: * updated zabbix (1.8.2 -> 1.8.3)
12:15 < Baldzius> if your luks _is_ sda4
12:15 < rogermason> asked for passphrase, got 'Key slot 0 unlocked.'
12:15 < Baldzius> ls /dev/mapper
12:15 < rogermason> home is listed
12:15 < Baldzius> check crypttab
12:15 < Baldzius> make sure you have no extra spaces of tabs
12:16 < Baldzius> make it "home /dev/sda4"
12:16 < Baldzius> only one space in between
12:16 < Baldzius> or typo
12:16 < Baldzius> damn it should work
12:16 < Baldzius> I will get this stuff for myself this weekend
12:17 < rogermason> Will need a few minutes to reboot to usb key.
12:28 < rogermason> No I'm still getting the same error.
12:39 < Baldzius> sorry, out of ideas
12:39 < Baldzius> I will check myself over weekend
12:39 < Baldzius> it should be something very simple
12:39 < Baldzius> as your luks device is fine
12:40 < rogermason> OK, thanks. I'll revert to unencrypted.
12:40 < Baldzius> you told me swap works?
12:40 < Baldzius> if swap works, luks should work too
12:41 < Baldzius> it should say "Found encrypted disks, activating volumes."
12:41 < Baldzius> and activate swap and ask for passphrase for you luks device
12:41 < rogermason> Yes. Swap works.
12:44 < Baldzius> ok , I might know what's the problem
12:44 < Baldzius> you device might not be luks
12:44 < Baldzius> open device again
12:45 < Baldzius> and do cryptsetup status home
12:45 -!- T-One [~T-One@194.183.141.238] has joined #t2
12:46 < Baldzius> also recreate filesystem mke2fs -j if you are using ext3
12:47 <@rxr> re
12:50 < Baldzius> or you have no /dev/console on boot
12:51 < rogermason> Baldzius: do I need to do cryptsetup -c aes-cbc-essiv:sha256 -y -s 256 luksFormat /dev/sda4?
12:51 < Baldzius> you did this before?
12:53 < rogermason> yes, as per your how-to.
12:53 < Baldzius> then no
12:54 < rogermason> But since then I have reformatted the partition.
12:54 < Baldzius> it doesn't matter once you get it initialized
12:54 < Baldzius> oh you mean you formated /dev/sda4?
12:54 < rogermason> yes
12:54 < Baldzius> :)
12:55 < Baldzius> then luks is gone
12:55 < Baldzius> then you need to do that command again
12:55 < Baldzius> open the device
12:56 < Baldzius> format it
12:56 < Baldzius> close it
12:56 < Baldzius> make sure it exist on crypttab
12:56 < Baldzius> fstab
12:56 < Baldzius> reboot
12:56 < Baldzius> before closing you can check status
12:56 < rogermason> open the device means?
12:57 < Baldzius> cryptsetup luksOpen /dev/sda4 home
12:59 < Baldzius> howto should be accurate anyway
13:00 < Baldzius> except missing home entry for home in crypttab
13:10 < rogermason> No. It is still failing in the same way. Time for breakfast :)
13:17 < Baldzius> crypt_setup isLuks /dev/sda4
13:18 < Baldzius> lunch
13:50 <@rxr> loving xcode: Internal error occurred while creating dependency graph: _registerUndoObject:: NSUndoManager 0x20a9f2cc0 is in invalid state, must begin a group before registering undo
13:51 <@rxr> guess time to reboot the monster that grew bigger the Emacs, ...
13:59 < rogermason> Baldzius: cryptsetup isLuks /dev/sda4 returns nothing
14:02 < Baldzius> cryptsetup status home
14:03 < rogermason> 'home is inactive'
14:05 < Baldzius> cryptsetup luksOpen /dev/sda4 home
14:05 < Baldzius> and then status again
14:07 < rogermason> I'm asked for passphrase. status says its active and lists details.
14:08 < Baldzius> if you say that swap works so it means crypttab works
14:08 < Baldzius> as swap is not luks so it doesn't ask for passphrase
14:08 < rogermason> status on swap gives similar details.
14:09 < rogermason> Is there some chance something messed up in kernel?
14:09 < Baldzius> when you boot, does it says "Found encrypted disks, activating volumes."
14:13 < rogermason> Yes. Then 'Checking file systems'. It is the fs check on /dev/mapper/home that fails.
14:16 < Baldzius> but it doesn't ask you a passphrase ?
14:20 < rogermason> No.
14:20 < rogermason> Got to go. back later. Thanks.
14:22 < Baldzius> then I'd say you don't have /dev/console at boot
14:36 < Baldzius> rxr: what's the rule if I want to add a package created from 3rd party?
14:36 < Baldzius> let's say opensde
14:36 < Baldzius> they have 2 packages
14:36 < Baldzius> they are quite big and I don't have time to package them myself
14:37 < Baldzius> needed for latest fecking kde
14:37 <@rxr> well, theoretically you could just copy them, leaving their copyright line, like we have in most packages since ROCK Linux, ...
14:37 <@rxr> or I create them from scratch :-)
14:37 <@rxr> which packages are that ?
14:39 < Baldzius> http://git.opensde.net/opensde/package-nopast/commit/?id=6a7e43d0ecbb33d147af264175925cc4661d2cd8
14:39 < Baldzius> http://git.opensde.net/opensde/package-nopast/commit/?id=026b2abb153c6039a6280365cfae5d17237fcf0e
14:40 <@rxr> hm
14:40 <@rxr> we have docbook xml files, now ?
14:41 <@rxr> http://www.t2-project.org/packages/docbookx.html or so ?
14:41 < Baldzius> hm
14:41 < Baldzius> indeed
14:41 <@rxr> I vaguely remeber seeing this xml fluffery xmlcatalog handling in some of those packages
14:42 < Baldzius> but I think kdelibs somehow can't find it
14:42 < Baldzius> I will take a closer looks later
14:42 <@rxr> maybe priority? or just the latest version missing
14:42 <@rxr> ?
14:49 < Baldzius> yeah, will check, thanks
14:52 <@rxr> our graph is skyrocking, ... http://svn.exactcode.de/big.png
15:00 < Baldzius> 40 merges waiting from 8.0 -> trunk
15:05 < CIA-32> aldas * r37293 /branches/8.0/package/develop/gcc/wrappers.in:
15:05 < CIA-32> * merged r37284 from trunk to branches/8.0:
15:05 < CIA-32> fix gcc wrapper creation, wrap gfortran, instead of the no longer existing g77
15:06 < CIA-32> aldas * r37294 /branches/8.0/package/develop/gcc/parse-config:
15:06 < CIA-32> * merged r37283 from trunk to branches/8.0:
15:06 < CIA-32> fixed gcc fortran support by using gfortran, instead of the no longer existing g77
15:08 < CIA-32> aldas * r37295 /branches/8.0/package/network/dhcp/rocknet_dhcp.sh:
15:08 < CIA-32> * merged r37233 from trunk to branches/8.0:
15:08 < CIA-32> fixed dhcp netconf module for the case of multiple dhcp network
15:08 < CIA-32> configs and the to-be-activated interface having another name
15:08 < CIA-32> then the one specified last, by not using an global, likely
15:08 < CIA-32> clobbered interface variabel reference - a wonder noone noticed,
15:08 < CIA-32> complained before, ...
15:25 < Baldzius> for kdelibs is just a path problem, nothing major
15:32 <@rxr> :-)
15:33 < Baldzius> not sure how to make it dynamic atm
15:33 < Baldzius> will set is as static for now
15:34 < Baldzius> +var_append extraconfopt " " '-DDOCBOOKXML_CURRENTDTD_DIR=$root/usr/share/xml/docbook/docbook-xml-dtd-4.2'
15:34 < Baldzius> +var_append extraconfopt " " '-DDOCBOOKXSL_DIR=$root/usr/share/xml/docbook/docbook-xsl-1.75.2'
15:34 < Baldzius> I don't worry about xml
15:34 < Baldzius> but xsl path will change in the future
15:34 < Baldzius> even though chance is small
15:46 <@rxr> I think this catalog registration stuff should register it in some /etc/...catalog file so that it can be found by docbook tools
15:48 < Baldzius> stupid cmake doesn't look for that
15:48 < Baldzius> have some hardcoded paths and that's it
15:52 <@rxr> I'm not a particular lover of cmake either, ...
16:19 -!- rconstruct [~cgwork@bl16-166-202.dsl.telepac.pt] has joined #t2
16:19 < rconstruct> hi
16:19 < rconstruct> rxr: there are some major problems with xorg...
16:19 < rconstruct> well, not that major, but major enough to cause big headaches
16:20 < rconstruct> some fonts went installed into /usr/X11/lib/X11/fonts/X11
16:20 < rconstruct> they should be in /usr/X11/lib/X11/fonts
16:20 < rconstruct> not fonts/X11/
16:20 < rconstruct> the xorg file autoconfigures itself to add /usr/X11/lib/X11/fonts/fontstypes_here
16:20 < rconstruct> so there are several important font types that don't get installed in correct type
16:21 < rconstruct> this means that those fonts pkgs need their path fixed, so that fonts.scale and fonts.dir (and fc-cache) are generated accordingly
16:21 < rconstruct> otherwise when using UTF8 in terminals and such, you'll get some strange text, non-AA like (at least in xterm with fluxbox, blackbox and such)
16:21 < rconstruct> also, there's a major problem with the nvidia driver package
16:22 < rconstruct> new kernels introduced nouveau, the reverse engineered OSS driver, totally incompatible with nvidia driver
16:22 -!- arete74 [~arete74@net-93-144-136-68.cust.dsl.teletu.it] has joined #t2
16:22 < rconstruct> so if kernel is enabled with nouveau, the user cannot install nvidia binary drivers
16:22 < rconstruct> and vice-versa
16:27 -!- mqueiros_ [~mqueiros@a81-84-220-211.cpe.netcabo.pt] has joined #t2
16:36 -!- Baldzius [~aldas@87-198-133-94.static.ptr.magnet.ie] has quit [Quit: Leaving]
16:43 -!- rconstruct [~cgwork@bl16-166-202.dsl.telepac.pt] has quit [Ping timeout: 276 seconds]
16:50 -!- rconstruct [~cgwork@bl16-166-202.dsl.telepac.pt] has joined #t2
17:01 < rogermason> baldius said: -then I'd say you don't have /dev/console at boot'. So, how do i get /dev/console? fiddle with udev?
17:08 < rogermason> I just checked: /dev/console exists
17:15 -!- arete74 [~arete74@net-93-144-136-68.cust.dsl.teletu.it] has quit [Ping timeout: 240 seconds]
17:28 -!- arete74 [~arete74@net-93-144-166-179.cust.dsl.teletu.it] has joined #t2
17:47 -!- arete74 [~arete74@net-93-144-166-179.cust.dsl.teletu.it] has quit [Quit: Leaving]
17:48 -!- arete74 [~arete74@net-93-144-166-179.cust.dsl.teletu.it] has joined #t2
17:52 -!- arete74 [~arete74@net-93-144-166-179.cust.dsl.teletu.it] has quit [Client Quit]
18:15 < rconstruct> ok, hmm, gcl is a problem
18:15 < rconstruct> http://www.mail-archive.com/gcl-devel@gnu.org/msg01890.html
18:15 <@rxr> hm
18:15 <@rxr> The Debugger has exited due to signal 15 (SIGTERM).The Debugger has exited due to signal 15 (SIGTERM).
18:16 <@rxr> fun, good that is not t2 :-)
18:16 < rconstruct> debian and redhat/fedora ppl mention having to disable brk randomization in kernel, patching some src files, and configure script
18:16 < rconstruct> well, i get the linking error they have
18:16 < rconstruct> gcl devs mention it's fixed in 2.6.8
18:16 < rconstruct> but hmm no 2.6.8 yet, so it'll be a cvs co
18:17 < rconstruct> debian/fedora ppl probably have a 2.6.8pre package, from a cvs co, since there's no 2.6.8pre in gcl site
18:17 < rconstruct> i'm trying a couple of things
18:17 < rconstruct> the nvidia binary drivers issue, won't work with nouveau, the user needs to build a new kernel
18:17 < rconstruct> and remove DRI, etc...
18:18 < koan> I noticed that the nouveau driver doesn't build here
18:18 < rconstruct> of course he can use nouveau, but in which case, nvidia pkg should check for nouveau and refuse to install
18:18 < rconstruct> hmm, worked here
18:18 < rconstruct> built, installed, x starts and everything
18:19 < rconstruct> it's just not enough for the use i give to the machine though, i need the binary drivers (unfortunately, hopefully nouveau will continue improving)
18:21 < rconstruct> http://lists.gnu.org/archive/html/gcl-devel/2010-08/msg00000.html
18:21 < rconstruct> for gcl
18:21 < rconstruct> seems it's serious enough to warrant a new release
18:25 <@rxr> rconstruct: can you import some fixes to get this work?
18:25 <@rxr> for nouveau it's probably best to have it working out of the box, as few need the high-performance binary-only-one
18:26 <@rxr> guess it's best to just rebuld your kernel without hte nouveau bits, ...
18:26 <@rxr> --typos of this French name that I always misspell, ...
18:26 < rconstruct> well, just built 2.6.35.1 with some evil things
18:26 < rconstruct> lol
18:26 < rconstruct> you don't want to know... really...
18:26 <@rxr> for gcl or nvidia?
18:26 < rconstruct> nvidia
18:26 < rconstruct> but saw some interesting "features" there
18:27 < rconstruct> experimental...
18:27 < rconstruct> couldn't resist
18:27 < rconstruct> lol
18:27 <@rxr> heh, yeah, I'm not too interesting in either Nvidia's binary junk nor LISP ...
18:27 < rconstruct> well, i don't have a choice with nvidia ...
18:27 < rconstruct> need 3d hw acceleration for work
18:27 <@rxr> can imagine
18:27 < rconstruct> until the nouveau people, or ati OSS people get decent performance drivers
18:27 < rconstruct> ati ppl are further ahead with this though
18:28 <@rxr> just rebuild the kernel with the CONFIG_NOUVEAOU not set, and you're there
18:28 < rconstruct> thanks to AMD/ATI for disclosing the docs
18:28 < rconstruct> hmm, well i built it by hand
18:28 < rconstruct> make menuconfig, etc...
18:28 <@rxr> or convince the nvidia guys that it would be nice if their binary-only blob would play nice with vanilla kernel configs :-)
18:28 <@rxr> no need to build by hand
18:28 <@rxr> just add this one thing to a t2 kernel config
18:28 < rconstruct> taht's the issue, you don't want my config
18:28 < rconstruct> trust me
18:28 < rconstruct> lol
18:28 < rconstruct> it's not going to behave with other machines
18:29 < rconstruct> lol
18:29 < rconstruct> i enabled some embebbing options, made gcc "uninline" functions marked inline
18:29 < rconstruct> stripped symbols generated from assembler files
18:29 < rconstruct> removed everything except the essential for this machine
18:30 < rconstruct> and enabled some other things for this alone, some amdfam10 features
18:30 < rconstruct> enabled HPET as well, seems amdfam10 also has error correction for memory
18:30 < rconstruct> reporting....
18:31 < rconstruct> and so on
18:31 < rconstruct> i'm just used to build it by hand since 2.2 days
18:31 < rconstruct> gcl... well need gcl for maxima/wxmaxima
18:32 < rconstruct> btw, t2 boots fast on a SSD
18:32 < rconstruct> you don't see the kernel boot, you just see a blur and the login prompt
18:32 < rconstruct> wondering if i made the right choice in fs though, using ext2
18:32 < rconstruct> noop schedular for it
18:33 < rconstruct> noatime/nodiratime
18:33 < rconstruct> *scheduler
18:33 < rconstruct> i don't trust ext4
18:33 < rconstruct> i'll try it in 10 years or so
18:33 < rconstruct> once other people find all the bugs, after loosing their data
18:33 < rconstruct> lol
18:38 <@rxr> i use ext4 on my SSD IIRC
18:38 < rconstruct> i'm very conservative with fs
18:38 < rconstruct> already lost data, for whatever reasons, twice
18:38 < rconstruct> once on xfs, once on reiser3
18:39 < rconstruct> precious data
18:39 < rconstruct> now i have it mirrored in 5, 6 hdds, plus one external
18:39 < rconstruct> just in case
18:39 <@rxr> yeah Im not an ext* fan, used reiser3 in the past, too bad Hand killed his wife, ...
18:39 <@rxr> Hans even
18:39 < rconstruct> eheheh, the orphaned inodes...
18:39 < rconstruct> there were lots of bad jokes by that time
18:39 < rconstruct> fs/murder jokes
18:39 < rconstruct> lol
18:40 < rconstruct> bad taste, funny nevertheless
18:40 < rconstruct> lol
18:40 < rconstruct> but about the ssd, well, was trying to minimize writes
18:40 < rconstruct> i placed /var, /tmp, /usr/src in hdds
18:41 < rconstruct> i know ext4 has a trim option
18:41 < rconstruct> but hmm, the ssd controller should do all the wear&levelling/ECC anyway
18:41 < rconstruct> not sure it has a cache this ssd though, forgot
18:41 < koan> I don't think writes are still such an issue with modern SSDs
18:42 <@rxr> I trust on that, too
18:42 < koan> except if you're running your machine 24/7
18:42 < rconstruct> well...
18:42 < rconstruct> not 24/7
18:43 <@rxr> have mac osx and windows 7 on the SSD along t2 and all are not using TRIM as this ATA commdand is not going thru the Nokia Booklet's PATA-to-SATA bridge :-(
18:43 < rconstruct> but hmm... it's up frequently
18:43 < rconstruct> windows 7?
18:43 < rconstruct> o_O
18:43 < rconstruct> you have to use that?
18:43 < rconstruct> work?
18:48 <@rxr> not often
18:48 <@rxr> but sometimes clients bug me to test something with win7
18:48 <@rxr> and as this booklet already came with it, ...
18:49 < rconstruct> i have a blessed life in that respect
18:49 < rconstruct> haven't seen windows since nt 3.51
18:49 < rconstruct> and when family asks me to "fix their computers" i tell them i don't know, i use other "exotic systems"
18:49 < rconstruct> never again....
18:50 < rconstruct> lost a work in windows, thanks to blaster32
18:50 < rconstruct> lost the work, the face, the client, the $
18:50 < rconstruct> not going to bet that in redmond's incompetence ever again
18:51 < rconstruct> quick question, anything against rc.local in t2?
19:03 <@rxr> patch to add it welcome
19:03 <@rxr> in the meantime you could add stuff to /etc/conf/kernel, ...
19:05 < rconstruct> ok, i thought t2 supported rc.local, will take a look at this
19:05 <@rxr> yeah, somehow we never had it in the whole ROCK histor
19:05 <@rxr> y
19:05 <@rxr> but for compatiblity it may be nice to add
19:30 < rconstruct> k, back 2 work, i'll b around
20:32 -!- rconstruct [~cgwork@bl16-166-202.dsl.telepac.pt] has quit [Quit: leaving]
21:12 < CIA-32> aldas * r37296 /trunk/package/develop/qjson/ (. qjson.cache qjson.desc): * added qjson (0.7.1) - A library for mapping JSON data to QVariant objects
21:12 < CIA-32> aldas * r37297 /trunk/package/develop/libdbusmenu-qt/ (. libdbusmenu-qt.desc libdbusmenu-qtcache): * added libdbusmenu-qt (0.3.0) - A library providing Qt implementation of DBusMenu specification
21:13 < CIA-32> aldas * r37298 /trunk/package/kde/kdelibs/kdelibs.cache: * updated kdelibs cache to include qjson and libdbusmenu-qt
21:15 < CIA-32> aldas * r37299 /trunk/package/kde/kdelibs/kde-conf.in: * help kdelibs to find docbookx files, hardcoded atm temporary, needs to be fixed later
--- Log closed Tue Aug 17 00:00:49 2010