T2 IRC Log: 2005-04-26

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 Tue Apr 26 00:00:43 2005
00:03 < valentin> wasn't there a gcc stl extention for getline which returns a std::string instead of dumping the output into a char* buffer ?
00:05 < jeru> good night all
00:06 -!- jeru [~jeru@p54BFF69C.dip.t-dialin.net] has quit [Remote closed the connection]
00:06 < valentin> gn8 jeru
00:06 < valentin> hey... now he is gone :/
00:34 < sepp> good night all
00:41 < valentin> gn8 sepp
01:41 < Mech> hey guys - anyone know where to find docs on the gnu system naming convention ? autoconf docs mention its 'cpu-vendor-os' but I thought it had more components ??
01:42 < Mech> (or is 'gnu-linux' considered 1 component ? )
01:50 < valentin> mech: i think so ...
01:51 < Mech> valentin: so if I build something with uclibc - Does that get put in there ? maybe i386-unknown-linux-uclibc ?
01:51 < valentin> yep
01:52 < Mech> ahh.. ok
01:52 < valentin> irc gnu stands for gnu libc
01:52 < valentin> it should be linux-gnu ?
01:52 < Mech> yeah - I got it backwards :-P
01:53 < Mech> when you add stuff to T2 - say heimdal support...
01:53 < valentin> mnemoc: are multiple search/nameserver entries allowed in one rocknet interface config ?
01:54 < Mech> which can be used in multiple packages, is it the responsiblity of the heimdal maintainer or the other package maintainers to update the 'extended' packages ?
01:57 < valentin> to be honest: in T2 maintainers are often silently ignored when it comes to updateing. Whoever sees an advantage/fix in some update, submits the patch. The maintainer is there to give feedback and he has the right to veto agains modifications in the package :)
01:58 < Mech> ahh thanks :-)
01:58 < valentin> mnemoc updates them all, anyway ....
01:59 < Mech> I'm having to build a 'mini-T2' for a project...
01:59 < valentin> what kind of project ?
01:59 < Mech> hopefully, after this, I can get back into T2 proper
01:59 < Mech> valentin: just a lil contract gig - not finalized yet, but I'm starting my background research/brainstorming/etc.
02:01 < Mech> two interesting differences with T2 will be database storage of build data, and a 'global configuration' step, where each package gets to look at the existing build parms and change them (for just itself, or globally)
02:01 < valentin> for an ip xxx.xxx.xxx.xxx, what are those xxx tripples called ?
02:02 < valentin> Mech: you can have the second point
02:03 < Mech> ??
02:03 < valentin> the global configuration - many packages insert their config option into the config
02:16 < Mech> sorry - phone
02:16 < Mech> anyway, with this 'config stage' - actual code in each recipe could be executed to evaluate the existing build parms and modify them
02:17 < Mech> and the 'finished' build parms would be stored in a database for future use
02:18 < Mech> it would be a very small system - it only has to support 1 particular 'mini-distro' but I thought it might be good for playing/testing stuff
02:18 < valentin> by now you can script in the package .conf files or in the targets
02:18 < Mech> ahh - didn't realize that... so I guess the database would be the only real difference
02:20 < Mech> the database would have all the packages in build order, so you can just walk the db from start to finish and dispatch jobs. Should cut down on memory usage versus keeping the whole build graph in memory (like OpenEmbedded does - python sucks over 300M for a build)
02:25 < valentin> what is the "build graph" ?
02:26 < Mech> the dependency graph for building packages - "busybox depends on glibc or uclibc"
02:26 < valentin> that is not 300M
02:27 < Mech> valentin: It explodes out - and I don't think python is very memory effiecient :-P OE brings my lil 2.6GHz celeron to its knees!
02:27 < valentin> in theory, a graph has at most quadratic number of edges
02:28 < Mech> I build a 'base image' with like 400 packages - it took like 8 hours
02:28 < valentin> hm - i see no reason why the dependency graph should slow things up
02:28 < valentin> and what is the pure build time ?
02:29 < Mech> not sure about that - but OE builds a full cross toolchain from scratch at the start of ever build - so that undoubtbly added a chunk of time
02:29 < Mech> I'm actually working in Lua instead of python
02:30 < Mech> Lua offered the advantages of being small, fast, and very easy to integrate with 'C' code - so time critical portions could be re-written in pure 'C' later
02:30 < valentin> we considered using lua for our config system in future versions
02:31 < valentin> by now most of our scripts are bash and binutils based
02:32 < valentin> but there is never a memory usage of 300 MB
02:32 < valentin> no idea what OE does
02:33 < valentin> normaly you seldom need more than 1000 Variables and a douzend sub-processes
02:34 < valentin> maybe OE holds the configuration states for each package in global memory all the time
02:34 < valentin> in t2 the config is evaluated each time a package is build
02:34 < Mech> I believe it does - its based on portage... I get the impression it reads everything into memory, and keeps it there
02:38 < valentin> i did not take a closer look at OE - but it seems to aim for a similar task like t2, only that it is far from perfect
02:39 < Mech> valentin: its pretty interesting - they have about 2500 packages available. the entire process is 'data driven' - configs (re)define vars to control the build
02:39 < valentin> and the people over there seem to develop software, too
02:39 < valentin> Mech: that is what we do
02:39 < Mech> makes it very hard to figure out where settings are coming from
02:40 < Mech> valentin: T2 seems to be a lil 'clearer' then OE - easier to find stuff
02:40 < Mech> yeah - a lot of the OE guys do embedded dev. for a living
02:40 < Mech> also, a lot of the OE recipes are 'broken' for 'wierd' combinations (uclibc, odd processors, etc)
02:41 < Mech> I had to fight just to get a i386 boot image :-P
02:41 < valentin> i would not bet every combination of hw/targets is working in t2, but at least the desktop target is well tested
02:42 < valentin> and t2 is not restricted to 'embedded'
02:43 < valentin> only thing is: we missed the marketing during the years
02:43 < Mech> yeah - for something like this a 'build farm' is pretty much a requirement - you really need to be able to do continous builds with multiple variations to see whats broken :-(
02:43 < Mech> a big task :-(
02:43 < valentin> esp. gentoo is well known to the world. it's ebuild hell...
02:44 < valentin> Mech: we aim for a bigger user base instead of a compile farm
02:45 < valentin> it is very late now - i'd like to take a nap
02:45 < Mech> valentin: yeah - big and varied would allow excercising most of the options..
02:45 < Mech> oh - g'nite :-)
02:45 < Mech> nice talking to you :-)
02:46 < valentin> have much fun - if there are questions, wait for someone to wake up or try our mailing list (hint:)
02:47 < valentin> cu
03:33 < mnemoc> re
03:42 < CIA-9> amery * r8525 /branches/2.1/package/powerpc/mol/ (pkg_linux24_post.conf pkg_linux24benh_post.conf): * merged 8520, fixed the mol postlinux conditionals to only build in stage > 1
03:50 < CIA-9> amery * r8526 /branches/2.1/package/scientific/comedi-driver/ (4 files): * merged 8521, fixed bug #6 - fixed comedi-driver postlinux to only build for stage >= 1
03:54 < CIA-9> amery * r8527 /branches/2.1/package/network/slmodem/ (6 files): * merged 8519, 8522, 8523 - fixed bug #7 - fixed slmodem to not install tools and not touch etc/modprobe.conf
04:03 < CIA-9> amery * r8528 /branches/2.1/package/network/slmodem/slmodem.desc: * reenabled slmodem, bug #7 is fixed and it doesn't show any problem further.
04:39 < Mech> has anyone played with AutoPackager ?
04:42 < mnemoc> not me
04:44 < Mech> mnemoc: http://autopackage.org/
04:44 < Mech> could be interesting - I'm looking for something with a smaller footprint then .debs
04:44 < Mech> debs store tons of state on the filesystem :-( ipkg could be good too
04:45 < mnemoc> we just tarball the flist
04:47 < Mech> mnemoc: yeah - that would be the ultimate in low overhead. But I'd like some basic uninstall capability and the ability to download from a web repository
04:47 < mnemoc> we have uninstall capability :)
04:47 < mnemoc> even safe-update capabilities
04:49 < Mech> 'safe-update' ?
04:49 < mnemoc> update creates .new files for eveyr modified file
04:50 < Mech> ahh
04:51 < Mech> mnemoc: what is the disk overhead like? I assume you have to store some sort of db/list of whats installed to handle un-install ?
04:52 < mnemoc> var/adm/*/$pkg does the trick
04:52 < mnemoc> plain files but easy to compress
04:52 < mnemoc> flist is file list
04:52 < mnemoc> md5sum the original checkusm of every file
04:52 < mnemoc> we may lzo them in future
04:53 < Mech> so say, 100 bytes per file (full path + md5sum ) ?
04:54 < Mech> that would be umm 100K for 1000 files ( aside from rounding error :-P )
04:54 < Mech> so 10K files per meg
04:54 < Mech> thats a lot better then the debian stuff
04:55 < Mech> is there support for un-install scripts ?
04:55 < mnemoc> let see linux26
04:55 < mnemoc> with all third-party modules ... (including nvidia and friends)
04:55 < mnemoc> # du var/adm/*/linux26
04:55 < mnemoc> 4 var/adm/cache/linux26
04:55 < mnemoc> 104 var/adm/cksums/linux26
04:55 < mnemoc> 12 var/adm/dep-debug/linux26
04:55 < mnemoc> 4 var/adm/dependencies/linux26
04:55 < mnemoc> 4 var/adm/descs/linux26
04:55 < mnemoc> 112 var/adm/flists/linux26
04:56 < mnemoc> 128 var/adm/md5sums/linux26
04:56 < mnemoc> 404 var/adm/olists/linux26
04:56 < mnemoc> 4 var/adm/packages/linux26
04:56 < mnemoc> flists, packages and md5sums are the only mandatory
04:58 < mnemoc> currently we have postinstall scripts.... but if postremove are needed they are not hard to add
04:58 < mnemoc> # wc -l var/adm/flists/linux26
04:58 < mnemoc> 1554 var/adm/flists/linux26
04:58 < mnemoc> linux26 has 1554 files
04:58 < Mech> and its 112 blocks - 66K ?
04:59 < Mech> err - 56K
04:59 < mnemoc> default du output is kilobyte afaik
05:00 < mnemoc> but those are plain text files.... highly compressables
05:01 < Mech> yeah - should get 4:1 easy on text
05:03 < mnemoc> but i think it's better to have var/adm mounted on a compressed filesystem instead of hacking var/adm compression
05:05 < Mech> I think ipkg actually skips the file list - I believe it can be setup to download the package and use that to control removing files
05:06 < Mech> interesting idea - saves in disk space, but you pay for it in network transfer
05:07 < Mech> I think the reasoning was you typically don't 'remove' software - your more likely to update it with a new version. So the file list from the new version can be used to remove the old version prior to update
05:08 < mnemoc> the file list of the new version say nothing about the old
05:08 < mnemoc> many package version their files
05:08 < Mech> yeah - or files may be added or removed.
05:09 < Mech> to be honest though - I haven't had a chance to play with ipkg yet. Just done a lil reading about it
05:09 < Mech> so I could be way off mark
05:09 < Mech> given that it was designed for stuff like ipaq's, I'm sure space will be a prime concern
05:11 -!- sepp [~sepp@p213.54.227.11.tisdip.tiscali.de] has quit [Read error: 60 (Operation timed out)]
05:14 < mnemoc> sometimes, even when space (or speed) is the prime concern, people tend to design using bloat eyes
05:14 < mnemoc> simplicity is the best way to design for space or speed (imho)
05:17 < mnemoc> cramfs can do the rest (just an example)
05:27 < CIA-9> amery * r8529 /trunk/package/ (python/wxpython/wxpython.desc x11/wxwidgets/wxwidgets.desc):
05:27 < CIA-9> * updated wxwidgets (2.5.3.1 -> 2.5.5.1)
05:27 < CIA-9> * updated wxpython (2.5.3.1 -> 2.5.5.1)
06:15 -!- keinek_ [~Keinek@201.254.13.95] has joined #t2
06:24 -!- keinek [~Keinek@201.254.12.98] has quit [Read error: 60 (Operation timed out)]
07:41 < rxr> re
07:42 < rxr> mnemoc: isn't 2.6 out ?
07:43 < mnemoc> not for wxpython
07:43 < mnemoc> and we use wxwidgets from wxpython
07:52 * mnemoc trying to build ghostscript with pdf_sec.ps injection disabled
07:52 < rxr> mnemoc: I can take a look into that totay
07:52 < rxr> today even
07:53 < mnemoc> i need to know if its a damn-weird side effect of a .ps, or ccache
07:54 < mnemoc> it failed
07:54 < rxr> mnemoc: I investigate in a minute, ok ?
07:54 < mnemoc> ok
07:54 < mnemoc> it's all yours
07:56 < rxr> maybe you could debug segfaulting binaries of cc-4.0.0 with -mcpu=ultrasparc ?
07:56 < rxr> just jokin ,-)
07:57 < mnemoc> no thanks :)
07:57 < rxr> ghostscript-8.15rc2 ? ?#
07:57 < mnemoc> yep
07:57 < rxr> well, ...
07:58 < mnemoc> the one which has been building for months
07:58 < rxr> okok
08:00 < mnemoc> weird... failed again with ccache disabled.... i'll file the issue and continue building
08:04 < rxr> mnemoc: please merge -^ ;-)
08:04 < CIA-9> rene * r8530 /trunk/package/develop/ccache/parse-config: * fixed chroot ccache directory creation
08:04 < mnemoc> ok
08:07 < CIA-9> amery * r8531 /branches/2.1/package/develop/ccache/parse-config: * merged 8530 - fixed chroot ccache directory creation
08:12 < rxr> hm - damn in debug.sh the loop mount is not done ?!
08:17 < rxr> | #define HAVE_MALLOC 0
08:17 < rxr> | #define malloc rpl_malloc
08:34 < rxr> mnemoc: with the above fixed gs builds
08:34 < rxr> I have to analyze why this happens ...
08:44 < rxr> damn - when I run it manually it does not happen :-(
08:47 < rxr> ah!
08:48 < rxr> the f*cking ghostsript configure crap thinks it is running in cross compilatation mode ...
08:48 < rxr> cross_compiling=maybe
08:48 < rxr> ??? what a assignment ...
08:48 < rxr> fazzy logic configure script ? *g*
08:49 < rxr> ah! there is the defect hiding ...
08:50 < rxr> 8429 amery var_append confopt "--with-drivers=ALL --with-x"
08:50 < rxr> it was you destroying it !!!
08:51 * rxr pulls mnemoc left ear
08:51 < rxr> mnemoc: test build running
08:51 < rxr> commit in some seconds
08:53 < rxr> == 08:52:37 =[5]=> Building graphic/ghostscript [8.15rc2 2.2.0-alpha].
08:53 < rxr> -> Compiler Cache Hits while pkg build: 100.00% (611 hits, 0 misses)
08:53 < rxr> == 04/26/05 08:55:01 =[5]=> Finished building package ghostscript.
08:53 < CIA-9> rene * r8532 /trunk/package/graphic/ghostscript/ghostscript.conf: * fixed ghostscript misconfiguration, regression caused by r8429
08:58 < CIA-9> rene * r8533 /branches/2.1/package/graphic/ghostscript/ghostscript.conf: * merged r8532 from trunk: fixed ghostscript regression
09:29 < rxr> mnemoc: spamassassin also get's perl to segfault with a ccache-less rebuild perl ...
09:42 < rxr> and locally ...
10:04 < rxr> mnemoc: perl dumps core when run from within a perl script
10:07 < rxr> hm
10:07 < rxr> second
10:12 < rxr> oh - maybe FPIC ?
10:35 -!- _jeru [~jeru@mail.oberlinhaus.de] has joined #t2
10:35 -!- _jeru is now known as jeru
10:35 < jeru> moin
10:36 < rxr> moin
10:43 < rxr> hm - maybe perl segfault is linux kernel arg $0 program name handling ...
10:45 < rxr> hm - no
10:59 < rxr> hm - no idea about the perl problem ...
11:05 < jeru> rxr: is ccache default size of 1000MB ok ?
11:05 < rxr> yes
11:05 < rxr> please remeber it is per package ...
11:07 < rxr> hm - partly run gcc testsuite on the sparc
11:07 < rxr> === gcc Summary ===
11:07 < rxr> # of expected passes 7583
11:07 < rxr> # of unexpected failures 5
11:07 < rxr> doesn't look too bad - and the unexpected are all timeouts as far as I have seen
11:11 < rxr> even with debugging symbols I do not get anything useful out of perl ...
11:11 < rxr> I leave it for now - too time wasting ...
11:47 < rxr> http://www.golem.de/0504/37720.html
11:55 < Mech> wow - 'distro build tools' keep turning up...
11:55 < Mech> http://www.gnnix.org/
11:55 < Mech> I feel for this guy - I've been in pretty similiar situations :-)
11:55 < Mech> he seems to have a text mode installer thats real easy to work with too - 'locknload'
12:01 < rxr> hm - somehow gcc-4.0 is really not that solid on sparc ... :-(((
12:31 < valentin> moin
12:34 < valentin> $100.000.... that is quite much compared to the price of 40 minimal 3Ghz systems
12:39 < rxr> hehe
12:39 < rxr> moin valentin
12:46 < Mech> hmm - scratchbox is going to use GAR for its build stuff ( http://www.lnx-bbc.org/garchitecture.html )
12:47 < Mech> It would be nice if everyone could decide on a 'standard' meta-data layout
12:47 < Mech> then T2, OE, Scratchbox, etc etc could share 1 meta data repo
12:48 < Mech> that would eliminate a _ton_ of redundant work
12:53 < rxr> valentin: what is your vote regarding how quickly we should try to adapt gcc-4.0 in 2.2 ?
12:54 < valentin> if the result is somehow stable (does not seem to be like that for sparc) i vote for quick adaption
12:55 < rxr> well - some stuff builds and runs fine on sparc
12:55 < rxr> e.g. even including vectorization ...
12:55 < rxr> but so far curl segfaults and bash does get a SIGBUS on startup ...
12:55 < valentin> how strict is 4.0 compared to 3.x ? that is, how many patches do we need for sources ?
12:55 < rxr> but only when ultrasparc optimizatoin are applied, e.g. not with v7 ...
12:56 < rxr> well - I just start a build - I expect a few packages to be patched
12:56 < rxr> some ugly historic extensions disapeared ...
12:56 < rxr> maybe I can hand out numbers soon
12:56 < rxr> btw. maybe the sparc problem is due to the old base system
12:56 < rxr> like glibc from stone age ...
12:57 < rxr> I try to start a system build there as soon as the athlon manages to produce a base system
12:59 < mtr_> hi rxr, valentin
13:00 < mtr_> do you have already fixed the glibc issues with gcc 4? "Please note that version 2.3.5 cannot be compiled using GCC 4": http://sources.redhat.com/ml/libc-announce/2005/msg00001.html
13:05 < rxr> well - it gets quite ugly right now anyway
13:05 < rxr> I'll delay this work even more ...
13:05 < rxr> for the log:
13:05 < rxr> D gcc-3.4.3-linkonce-1.patch
13:05 < rxr> D no-ld-error.patch
13:05 < rxr> D 3.4.3.patch
13:05 < rxr> D gcc-300-force-visibility.patch
13:05 < rxr> D 3.4.3-amd64.patch
13:05 < rxr> D no-install-zlib.patch
13:05 < valentin> with all the hype around 4.0 i expect upstream patches for such stuff soon
13:06 < rxr> with all the redhat'idi in glibc, i expect a public gcc-4 friendly glibc in about 2 years
13:07 < mtr_> ;-)
13:07 < CIA-9> rene * r8534 /trunk/package/base/gcc/apply-protector.sh: * "fixed" gcc to work without a protector patch
13:07 < rxr> M fixincl.patch
13:08 < mtr_> McGrath announced glibc 2.3.6 to appear soon, whatever this means
13:09 < valentin> if our transition to 2.2 takes as long as it did from 2.0 to 2.1 we have those two years ,)
13:12 < rxr> mtr_: there also where tags without a source tarball for over one year ...
13:32 < valentin> they are working with an air hammer outside. hope vibrations will not kill my server's harddrives...
13:34 < rxr> .oO
14:06 -!- mtr [~michael@Ha624.h.pppool.de] has joined #t2
14:24 -!- mtr_ [~michael@H912e.h.pppool.de] has quit [Read error: 110 (Connection timed out)]
14:57 < CIA-9> rene * r8535 /trunk/package/base/ (glibc/glibc.desc glibc32/glibc32.desc):
14:57 < CIA-9> * updated glibc (2.3.4 -> 2.3.5)
14:57 < CIA-9> * updated glibc32 (2.3.4 -> 2.3.5)
15:00 < rxr> we need a new config system - not only is it destructive, it is also dog slow ...
15:06 < Mech> nite guys :-)
15:06 < rxr> cu Mech
15:06 < Mech> well, actually its morning here - but I'm nocturnal :-P
15:06 < Mech> bye rxr :-)
15:06 -!- Mech [~chatzilla@dsl081-073-058.sfo1.dsl.speakeasy.net] has quit ["Chatzilla 0.9.66 [Mozilla rv:1.7.3/20041001]"]
15:09 < jeru> ok ... have to move too ... cu later
15:11 -!- jeru [~jeru@mail.oberlinhaus.de] has quit [Remote closed the connection]
15:24 -!- tilix [~ilia@212.116.139.141] has joined #t2
15:28 -!- tilix [~ilia@212.116.139.141] has quit [Remote closed the connection]
16:40 -!- rxr_ [~rene@e178182199.adsl.alicedsl.de] has joined #t2
16:40 -!- Topic for #t2: T2 | 2.1.0-beta4 RELEASED | The next generation of System Development Environments (SDE) | http://www.t2-project.org/
16:40 -!- Topic set by menomc [] [Sat Apr 16 00:33:34 2005]
16:40 [Users #t2]
16:40 [ _Ragnar_] [ mnemoc] [ nzg] [ rxr_ ] [ valentin]
16:40 [ CIA-9 ] [ mtr ] [ rxr] [ sparc-kly]
16:40 -!- Irssi: #t2: Total of 9 nicks [0 ops, 0 halfops, 0 voices, 9 normal]
16:40 -!- Channel #t2 created Sun Aug 8 21:15:33 2004
16:40 -!- [freenode-info] help freenode weed out clonebots, please register your IRC nick and auto-identify: http://freenode.net/faq.shtml#nicksetup
16:41 -!- Irssi: Join to #t2 was synced in 11 secs
16:53 -!- _jeru [~jeru@p54BFBC79.dip.t-dialin.net] has joined #t2
16:53 < _jeru> re hi
16:54 -!- rxr [~rene@e178139069.adsl.alicedsl.de] has quit [Read error: 110 (Connection timed out)]
16:57 -!- keinek_ [~Keinek@201.254.13.91] has joined #t2
16:58 -!- keinek [~Keinek@201.254.13.91] has joined #t2
16:59 < keinek> hi
16:59 < _jeru> hi keinek
16:59 < keinek> where is that keinek_?
17:00 < keinek> i was all night offline
17:00 < keinek> :/
17:00 < _jeru> this was displayed before you joined ->
17:00 < _jeru> keinek_ (~Keinek@201.254.13.91) has joined channel #t2
17:00 < _jeru> keinek (~Keinek@201.254.13.91) has joined channel #t2
17:43 < mnemoc> hi
17:43 < _jeru> hi mnemoc :)
17:44 < mnemoc> rxr_: my line felt again while we were talking :(
17:44 < mnemoc> hi _jeru
17:44 < _jeru> mnemoc: voip ?
17:45 < mnemoc> i don't have audio here :\
17:45 < mnemoc> talking=irc :)
17:45 < rxr_> mnemoc: no problem
17:45 < rxr_> ghostscript is fixed
17:45 < rxr_> but I stopped investigating Perl
17:45 < rxr_> no idea ...
17:45 < rxr_> it crashes executing some variabler content (./configure.pl or so)
17:46 -!- You're now known as rxr
17:46 < rxr> I think it is not too urgent
17:46 < rxr> let's rest it some days ... if you have not time to track that ugly thing ...
17:46 < rxr> mnemoc: maybe we should test out voip some day
17:46 < rxr> could be a nice "communication device" ,-)
17:46 < mnemoc> rxr: i only changed that (ghostscript) for confopt="$confopt .." to var_append... nothing else
17:47 < rxr> you forgot the " " seperator ;-)
17:47 < _jeru> rxr: I have some pending packages in my working copy for voip services ;)
17:47 < mnemoc> rxr: oh
17:47 < rxr> that --build=i386-bla-blus--the
17:47 < rxr> that --build=i386-bla-blus--with-drivers=all or so ..
17:47 < mnemoc> damn
17:47 < rxr> and thus did not build all drivers and made ghostscript think it is cross copiling
17:47 < rxr> since host != build ....
17:48 < rxr> was fun tracking
17:48 < rxr> however perl is not fun ... if anyone is interested ... ,-)
17:55 < mnemoc> no thanks :)
17:57 < rxr> too bad
17:58 < mnemoc> _jeru: when will www2 be back? i want to clock ghostscript bug
17:58 < mnemoc> _jeru: s/clock/close/
17:58 < mnemoc> .oO( that was a weird typo )o
17:59 < _jeru> mom
18:01 < _jeru> mnemoc: tracker is back ... but still the weird behavior (no nav for tracker atm) ... I will write a small static portlet to "simulate" the menu.
18:01 < mnemoc> _jeru: thnkas :) add thing commands on URL is 'unconfortable'
18:02 < mnemoc> s/thing//
18:02 < _jeru> mnemoc: ack
18:02 < mnemoc> 503 Service Unavailable
18:02 < _jeru> https:// ?
18:03 < mnemoc> oops
18:03 < _jeru> mnemoc: you thought it was offline ?
18:03 < mnemoc> yep
18:03 < mnemoc> [034] - Live Kernel Configuration Panel 0.2.1 ??
18:04 < _jeru> mnemoc: and we lost the state of the tracker issues due to our copying
18:05 < mnemoc> _jeru: it's databases gets corrupted very easily.... followup counter is always out of sync, and i have to reindex every time
18:06 < _jeru> hmmm ... damn
18:06 < mnemoc> i'll add a follow up to 11 now
18:08 < mnemoc> http://trasto.hopto.org/~wipeout/penguin.pdf
18:08 < keinek_> ups... i have two x-chat open
18:08 -!- keinek_ [~Keinek@201.254.13.91] has quit ["Abandonando"]
18:08 < keinek> i am a stupid.
18:08 < mnemoc> *G*
18:08 < keinek> :D
18:08 < mnemoc> ack
18:18 < valentin> hi mnemoc
18:18 < mnemoc> hi valentin
18:18 < _jeru> hi valentin
18:18 < valentin> i have tasty grapes from chile here
18:18 < mnemoc> :D
18:18 < mnemoc> EU removed the over fees imposed over chilean salmon yesterday
18:19 < valentin> nice - after ruining all chilean salmon-farmers ?
18:20 < mnemoc> they were close
18:24 < CIA-9> amery * r8536 /trunk/package/graphic/ghostscript/ghostscript.desc: * updated ghostscript (8.15rc2 -> 8.15rc3)
18:25 < mnemoc> does norway belong to EU?
18:25 < mnemoc> s/belong to/is part of/
18:25 < mnemoc> .oO( member? what's the proper word? )o
18:26 < valentin> i would say member, but i am not sure
18:27 < valentin> and no, norway does not belong
18:27 < mnemoc> norway has lot of money invested on salmon-farms here
18:28 < rxr> hehe
18:28 < rxr> #t2 the politicans and world trade channel
18:28 < valentin> mnemoc:
18:28 < valentin> http://en.wikipedia.org/wiki/European_union#Member_states_and_successive_enlargements
18:29 < valentin> rxr: what about my tea offer ?
18:29 < rxr> consuming tea this week over at you ?
18:29 < rxr> did you name a day ?
18:29 < valentin> mnemoc: so "member states" is the correct term
18:29 < valentin> rxr: end of week, see query
18:29 < rxr> oh
18:29 < valentin> pref. friday
18:29 * valentin -> noodles
18:29 < rxr> which query ...
18:30 < rxr> the last thing in query was on Apr 23 ...
18:30 < rxr> friday sound nice
18:30 < rxr> have to finish some work including the lingering gcc article first ...
18:32 < mnemoc> rxr: as far as we don't bash for free, i see no problem on discussing misc topics :)
18:32 < rxr> mnemoc: btw. glibc maintenance update into 2.1 ?
18:33 < mnemoc> rxr: let me finish this ref first :)
18:33 < mnemoc> rxr: file a merge or package request on the tracker :)
18:34 < mnemoc> rxr: i'll force us to use that damn tool
18:34 < rxr> why damn? why force ?
18:35 < mnemoc> because we are not used to use that kind of stuff
18:35 < mnemoc> when you merged ghostscript you didn't tell what bug was it related :)
18:46 < rxr> forgot that you filled a ticket for it ...
18:46 < mnemoc> rxr: what do you think about issue trackers? should we write some kind of policy?
18:52 < rxr> issue trackers - or the one we use ?
18:52 < mnemoc> both
18:53 < rxr> ouhm - if you have spare time - policies and ducuments are always welcome
18:54 < mnemoc> but i want to know your thoughts about those tools
18:56 -!- sparc-kly [~sparc@64.237.140.181] has quit [Read error: 54 (Connection reset by peer)]
19:04 < rxr> ouhm
19:04 < rxr> so you want me to write a paper so that you get an overview?
19:05 < valentin> rxr: you see my query ?
19:05 < rxr> ack
19:06 < valentin> :)
19:07 < mnemoc> rxr: nah... just some comments :) but if you want to sell an article about much better :)
19:19 < rxr> not really - I just wanted to know wath you expect me to do ,-)
19:20 < mnemoc> i want to hear you :)
19:23 < rxr> issue tracking is nice and a must
19:23 < rxr> if the plone thing scales I have no idea
19:24 < rxr> looked barely into ito
19:24 < rxr> how happy are you mnemoc and valentin with the plone site ?
19:26 -!- Cypl0dE [~mm@196-49-112.adsl.terra.cl] has joined #t2
19:26 < rxr> btw. I find the defaul font size a bit small
19:26 < rxr> hi Cypl0dE
19:27 < mnemoc> rxr: yep, i have to ctrl + every time :(
19:28 < rxr> my setups are often more sensible to this, since I usually drive my displays at a higer DPI
19:28 -!- Cypl0dE [~mm@196-49-112.adsl.terra.cl] has quit [Client Quit]
19:28 < mnemoc> i'm quite happy with plone thing as concept.... but the mess done to implement i18n is disturbing
19:29 < rxr> but when you also say this the font might indeed a bit tiny by default
19:29 < _jeru> rxr: I will increase it ;)
19:30 < mnemoc> _jeru: why manager has a home?
19:31 < _jeru> mnemoc: it's default
19:32 < rxr> I think we need at least a big design work to make the look attractive
19:32 < rxr> but I can take that part some time later
19:32 < rxr> (if necessary)
19:33 -!- Cypl0dE [~mm@196-49-112.adsl.terra.cl] has joined #t2
19:33 < _jeru> rxr: I'm preparing a nice skin already
19:33 < rxr> oh - cool
19:34 < mnemoc> but menu (portlets) problem is a bit more importand :)
19:34 < mnemoc> important*
19:34 < _jeru> mnemoc: ack
19:35 < rxr> i18n I guess
19:35 < rxr> of course functionality is important - I just worry to present T2 in a better light than the prevous preleminary site
19:35 < mnemoc> to get i18n working on navigation menu we had to update some core stuff
19:38 < rxr> right now I'm not too interested in the details ...
19:38 < mnemoc> :)
19:38 < rxr> mnemoc: you might consider pulling the gnome update into 2.1
19:38 < rxr> of course after your current ref build or so ...
19:38 < mnemoc> does it really build and work?
19:38 < mnemoc> i wanted to finish one ref before adding important rebuilds
19:45 < mnemoc> rxr: i think i'll remove wxpython24
19:45 -!- minto [~chatzilla@82-217-66-7.cable.quicknet.nl] has joined #t2
19:45 < minto> Hi all
19:46 < valentin> hi minto
19:46 < valentin> rxr: i like plone more than drupal from what i've seen so far
19:46 < minto> valentin: how is the father to be? And how is the mother?
19:46 < valentin> minto: thanks. good.
19:47 < valentin> how is your son ?
19:47 < minto> 18 months
19:48 < mnemoc> .oO( fahter to be.... uhm.... i wanted to name 2.1.0 as jsaw's daughter... what will i do with valentin's?? )o
19:48 < rxr> valentin: name 2.2 after her
19:48 < rxr> mnemoc: why remove wx24 ?
19:48 < valentin> what is with lars daughter ? *g*
19:48 < mnemoc> no wx24, wxpy24
19:49 < mnemoc> outch
19:49 < mnemoc> to many children :\
19:49 < valentin> btw minto: you are the great exception. the only one with a son...
19:49 < mnemoc> to many women surounding T2
19:49 < rxr> then hand the names for 2.1.0, 2.1.1 and 2.1.2 ...
19:49 < rxr> when we have soo many *grrr*
19:49 < minto> :-)
19:50 < _Ragnar_> *lol*
19:50 < rxr> or just queue the names up when we will miss them later ...
19:50 < mnemoc> ok, and rene's for 2.2.0
19:50 < rxr> or we release major versions more quickly ...
19:50 < rxr> mnemoc: I doubt that
19:50 < mnemoc> _Ragnar_: can you add yourself to about/team ?
19:50 < mnemoc> rxr: let see :)
19:50 < valentin> rxr and susann are young and healthy, so....
19:51 < mnemoc> "The T2 Generation"
19:51 < rxr> valentin: I and healthy ...
19:51 < minto> mnemoc: How can I test the swig .24 stuff?
19:51 < mnemoc> minto: svk?
19:52 < mnemoc> svk does extensive usage on SVN::
19:52 < mnemoc> of*
19:52 < _Ragnar_> mnemoc: how? where? what?:)
19:52 < mnemoc> _Ragnar_: https://www2.t2-project.org
19:52 < minto> mnemoc: ack
19:52 < valentin> btw: why https ?
19:52 < mnemoc> register and edit ... i hope
19:52 < mnemoc> ask jeru :(
19:53 < rxr> oh my god - one of my C++ tests runs ten times slower with -Os than -O2
19:53 < mnemoc> pff
19:53 < rxr> and 15 times slower than -O2 -funroll-loops
19:54 < _Ragnar_> where do i register?
19:54 < rxr> well - I know why I changed the optmization default from Os to O2
19:54 < minto> mnemoc: We have not svk package in T2
19:54 < valentin> rxr: funrolls speed up this much ?
19:54 < rxr> valentin: I just bench
19:55 < rxr> I just do a openbench rerun with 3.4.3 and 4.0.0 with way better selected optimization comparision
19:55 < rxr> I can publish gzip and tramp3d soon
19:55 < rxr> the rest will run longer
19:55 < rxr> valentin: unroll-loops is not always that much faster
19:56 < mnemoc> minto: nope... svk is a bunch of around 40 perl packages
20:12 < _jeru> I've to do some maintenance on plone
20:12 < mnemoc> ok
20:12 < _jeru> thx :)
20:13 < valentin> what will be the future plan, i.e when to replace www by www2 ?
20:13 < rxr> suer
20:13 < rxr> when www2 is ready
20:13 < valentin> so we will start to migrate the missing content of www biger Christ kommt mir das Kotzen wenn ich die die derzeitige
20:13 < valentin> damn
20:14 < valentin> so we will start to migrate the missing content of www -> www2 soon ...
20:14 < rxr> was kotzt ...
20:14 < mnemoc> we need to solve menu problem and properly set the roles and permissions before open it
20:14 < valentin> what is about svn.exactcode ?
20:14 < rxr> what is about that ?
20:14 < valentin> rxr: someone on heise.de kotzt
20:14 < mnemoc> and @t2-project.org forwards :p
20:15 < rxr> valentin: ?
20:15 < rxr> mnemoc: you want a mail alias ?
20:15 < rxr> valentin: I thought we keep svn.exactcode.de
20:15 < mnemoc> exactcode.de is _the_ T2 sponsor
20:15 < rxr> no transition - no server work - keeping our already complex enough svn setup not too complex ...
20:15 < valentin> rxr: that was my question
20:16 < rxr> mnemoc: adding mail forwards is easy
20:16 < rxr> just currently I have no virtual domains set up in the jailed exim
20:16 < valentin> will the .org site stay at jeru's machine ?
20:16 < rxr> that is alejandro@t2-project.org would also work for alejandro@exactcode.de
20:16 < rxr> which is not what we want ...
20:16 < mnemoc> amery
20:16 < mnemoc> not alejandro please
20:16 < rxr> mnemoc: no problem ,-)
20:17 < rxr> I just wanted to point out that I need to majorly reconfigure our mail system for that
20:17 < rxr> and I have a spam update pending
20:17 < rxr> spamassassin and friends update of course ...
20:19 < valentin> rxr: what happens when you have dhcp and ip/gw keywords in your config ? you allways get dhcp ?
20:19 < rxr> that is unso
20:19 < rxr> unsupported and undefined
20:19 < rxr> you will get a mix
20:20 < rxr> I think DHCP first and ip/gw evaluation later on
20:20 < valentin> hm ... then i have to "erase" those lines from the config when the user clicks on dhcp
20:20 < mnemoc> but dhcp will overwrite ip/gw later
20:20 < mnemoc> do you plan to hardcode those restrictions or write some kind of schema?
20:21 < valentin> actually i just wanted to delete the address items from the config if you enable dhcp
20:21 < mnemoc> hardcoded
20:21 < valentin> only thing is: you loose your old config then
20:21 < mnemoc> what about dhcp, udhcp and dhcpcd ?
20:21 < rxr> mnemoc: valenting is talking about the gui frontend only I think
20:21 < valentin> mnemoc: for the fontend, yes
20:22 < valentin> rxr: well ... i planed to reuse the classes later for some more complex setup tools
20:22 < mnemoc> a plugable schema for validating options would be naice :)
20:22 < mnemoc> nice*
20:22 < rxr> btw eaier to read vectorization support matrix:
20:22 -!- Cypl0dE [~mm@196-49-112.adsl.terra.cl] has quit [Read error: 110 (Connection timed out)]
20:22 < rxr> valentin: yeah - Config++ would be welcome for 2.2
20:22 < rxr> Int Float
20:22 < rxr> 4 8 16 8 16
20:22 < rxr> Alpha: X
20:22 < rxr> ARM: X X X X X
20:22 < rxr> i386: X X X X X
20:22 < rxr> ia64: X X
20:22 < rxr> mips: X
20:22 < rxr> powerpc: X X X X
20:22 < rxr> supeh: X X X X X
20:22 < rxr> Sparc: X X
20:23 < rxr> damn - I guess with the unintended tabs it looks imperfect in your IRC clients ...
20:23 < valentin> rxr: please s/\t/ / or so
20:23 < rxr> valentin: it was xemacs - no tabs intended - damn beast ...
20:23 < valentin> mnemoc: elaborate on your scheme
20:24 < rxr> I guess he means some ruleset to define which combinations make sense
20:24 < rxr> or belong together
20:24 < valentin> is the user supposed to edit that rules ?
20:24 < mnemoc> valentin: xml does it to see what is valid and what is not
20:25 < mnemoc> valentin: we have mod_foo.sh for the procedures.... we may have some 'language' to define de usage and inject it together
20:25 < mnemoc> valentin: so you can validate the syntax and consistence of the config file without hardcoding anything
20:26 < rxr> .o(the network might get a totally different script language in 2.2 or later)
20:26 < rxr> the current bash code is too error prone and slow
20:26 < rxr> evaluating the profiles on my s
20:26 < rxr> 270Mhz UltraSPARC with just 7 or so interfaces takes over 3 (!!!) seconds
20:27 < rxr> I think this is not reasonable for such a tiny parsing work
20:27 < mnemoc> rxr: network 'engine' or config file?
20:27 < valentin> mnemoc: i am coding a config editor, not the rocknet self
20:27 < valentin> and in the config frontend, you have to do some hardcoding if you want clever grouped dialogs
20:28 < valentin> or you do it all the very generic way in some intermediate language like lua
20:28 < mnemoc> using schemas you don't need to
20:28 < valentin> mnemoc: ack, but for that first version that is too complicated. i want more or less simple classes now
20:29 < mnemoc> :)
20:29 < rxr> and s.th. working in some days
20:29 < valentin> and something with save undo
20:29 < mnemoc> but please, don't remove the posibility of migrating a machine from 2.1 to 2.2
20:30 < valentin> the config file formats should be mainly preserved - the tools not
20:30 < mnemoc> no problem
20:30 < rxr> I do not plan any config change for 2.2
20:31 < mnemoc> nice
20:33 < rxr> remember: release early release often
20:33 < rxr> avoid senseless config api or ui chanfes
20:33 < rxr> changes
20:34 < mnemoc> ACK
20:34 < valentin> rxr: config change is ?
20:34 < rxr> file format change
20:35 < valentin> rxr rather does such uncomplicated changes like updating glibc/gcc/... :)
20:36 < rxr> you just install a system with an updated gcc/glibc no hassle involved
20:36 < mnemoc> foo-%
20:36 < rxr> no idea yet
20:38 < rxr> aaaaaaaarrrrrrrrrrrgggggggggghhhhhhhhhhhh
20:38 < rxr> my brother is bothering me with his gimp problems
20:39 < rxr> and now he went off the phone to the buzzing egg cooker ... :-((
20:39 < mnemoc> tell him to install photoshop on osX :)
20:39 < rxr> brother comes from bothering - right?
20:39 < mnemoc> i think so
20:40 < valentin> what does the code keyword do ?
20:40 < rxr> code ??
20:40 < mnemoc> source an script iirc
20:40 < valentin> public_code
20:40 < mnemoc> or run an script....
20:40 < rxr> uh?
20:40 < mnemoc> minute
20:40 < rxr> no idea
20:41 -!- sparc-kly [~sparc@64.237.138.134] has joined #t2
20:41 < valentin> mnemoc: we have on_up/on_down ?
20:41 < mnemoc> yes, excute anything
20:42 < mnemoc> on_up/on_down ?
20:42 < mnemoc> there is run_up and run_down
20:43 < mnemoc> to execute anything on that action
20:43 < mnemoc> code is similar but you say when
20:43 < mnemoc> a hack
20:43 < valentin> ok
20:44 < valentin> i do not like hacks ...
20:44 < mnemoc> neither do i
20:44 < valentin> consider an ip address xxx.xxx.xxx.xxx/bits <- what is that bits part called ?
20:45 < valentin> we do not have a real subnet mask keyword, do we ?
20:45 < mnemoc> 8,16,24,32 usually
20:45 < mnemoc> no, only bits
20:45 < valentin> mnemoc: i know, but what it the name ?
20:45 < mnemoc> ip like bits
20:45 < rxr> the official name is in the t2 handbook
20:45 < mnemoc> iproute2 likes bits
20:46 < mnemoc> "netmask in bits" ?
20:46 < rxr> CIR CIF or whatever ..
20:46 < valentin> netmask-bits
20:46 < mnemoc> :)
20:46 < keinek> hey rxr
20:47 < keinek> mac os X is pretty?
20:47 < keinek> is easy?
20:47 < rxr> \item ip ip-address/netmask-bits
20:47 < rxr>
20:47 < rxr> Set the given ip {\bf and} netmask in CIDR notation (e.g. 192.168.5.1/24)
20:47 < rxr> hey - my memory is n
20:47 < rxr> ot too bad - CIDR it was ,-)
20:48 < keinek> good
20:48 < keinek> is easy if i know windows or linux for design?
20:48 < keinek> ex. adobe...
20:48 < keinek> is the same thing?
20:49 < valentin> keinek: what do you mean with easy ?
20:49 < valentin> easy to use / configure / develop applications ?
20:50 < mnemoc> develop drivers? :)
20:51 < rxr> osx stinks majorly
20:51 < rxr> welll - not as much as windows - but developing under linux is way mroe relaxing
20:51 < valentin> rxr: depends ...
20:51 < rxr> the only thing OS X is halfly good with is ht GUI
20:51 < valentin> yep
20:52 < rxr> and that only halfly ...
20:52 < valentin> i use linux on my apple if there is no special reason to boot into OSX - this should tell keinek all....
20:52 < rxr> well - same here - of course ...
20:53 < rxr> and the special reason is only for money earning development purposes - not usability or need or so ...
20:54 < keinek> velentin easy is i come from linux or windows.
20:55 < keinek> velentin easy is if i come from linux or windows.
20:55 < rxr> some stuff is different and quite counter intutive ...
20:55 < rxr> ejecting disks by dragging the disk icon into the trash can and the like ...
20:55 < valentin> hehe
20:55 < rxr> also you have normally no right mouse button and use ctrl for all
20:56 < valentin> on the other hand i like the one-menu only concept
20:56 < rxr> and the cut'n paste is apple key rather than alt
20:56 < rxr> plus a lot of other pseudo intuitive things, that are not intuitive at all
20:56 < rxr> e.g. we had to read a lof of Interface Builder manuals to get some of the srange drag and drop
20:57 < rxr> concepts we never would have imagined ...
20:57 < keinek> ok.. in one week is like all day for use?
20:57 < rxr> of course it is graphical and colorful and when you got used to it it works ...
20:58 < keinek> :)
20:58 < keinek> is good know that
20:58 < rxr> for a lot of OS X bashing consult the t2 logs ...
20:58 < rxr> IRC logs
20:58 < valentin> keinek: if you need to choose between M$ or apple, please take apple
20:58 < rxr> http://dl.exactcode.de/irclogs/t2/
20:58 < keinek> i asking because if someone tell me for work with that in a job, so i dont say "no" :)
20:58 < keinek> sure
20:58 < keinek> apple is more stable
20:59 < rxr> yes osx is one of the better platforms
20:59 < keinek> just for design
20:59 < valentin> at least 10.3 - forgot about 10.2 stability
20:59 < rxr> osx is also way more colorful and better to use
20:59 < rxr> way stricter human interface guidelines and applications than Windows has
21:00 < mnemoc> rxr: http://t2.geeks.cl/regressions/stable/regressions.html look this mm failing chain
21:02 < rxr> mnemoc: saw s.th. simillar over here
21:02 < mnemoc> there was any change on LD reciently?
21:02 < rxr> however I have no idea when this gnoem *mm stuff broke
21:02 < rxr> hm no
21:02 < rxr> or maybe - well
21:02 < mnemoc> my last reference didn't have this problem
21:02 < rxr> there is this visibility hidden gcc patch right?
21:03 < rxr> but that is older
21:03 < rxr> as is the optional -Wld,--as-needed thing - I hope you have not enabledf
21:03 < mnemoc> yes, older
21:03 < mnemoc> my config file is there :)
21:03 < mnemoc> i haven't enabled anything odd
21:03 < mnemoc> just extra pkgsel
21:04 < mnemoc> and xtrace
21:04 < rxr> export SDECFG_LD_AS_NEEDED='0'
21:04 < valentin> how many search/nameserver entrys can i have in one interface section ?
21:04 < rxr> glibc defines 3 in total
21:04 < rxr> 3 in /etc/resolve.conf
21:04 < rxr> but I have not looked at the code, if it really has a hardcoded array of 3 or so ...
21:04 < rxr> valentin: add two lines to the GUI
21:04 < rxr> primary and secondary nameserver entries are standard in GUIs ...
21:05 < valentin> and search ?
21:05 < mnemoc> valentin: do you really want to implement those restrictions?
21:05 < valentin> rxr: i need to handle any existing (valid) config file
21:05 < rxr> valentin: maybe search is not limited ...
21:05 < valentin> mnemoc: nope
21:05 < rxr> but only one keyword IIRC
21:05 < rxr> have not looked at the glibc parser for search, too
21:05 < mnemoc> i have two here
21:06 < rxr> hm - maybe list the first two in a text entry field and the others in a expert section as full, space sepearted list
21:09 < mnemoc> i think so
21:13 < rxr> valentin: does RTL in GCC for register transfer language - or for s.th. else ?
21:22 < rxr> I think the above ...
21:24 < mnemoc> wtf? gtk-sharp2 failed
21:24 < rxr> .oO
21:25 < rxr> RTL Representation Most of the work of the compiler is done on an intermediate representation called register transfer language
21:25 < rxr> ok - ack
21:26 < rxr> hey cool
21:26 < rxr> in tramp3d g++-4.0.0 was able to vectorize a loop instantiated by std::vector line 216 ,-)
21:29 < mnemoc> rxr: will you add my mirrors for 2.2.0-alpha?
21:31 < mnemoc> great sf.net is dead again
21:31 < valentin> rxr: stand ?
21:32 < valentin> ok
21:34 < rxr> valentin: stand what ?
21:35 < valentin> forget it
21:36 < rxr> you must not confuse me ...
21:36 < valentin> you did confuse me with your question 21:16 < rxr> valentin: does RTL in GCC for register transfer language - or for s.th. else ?
21:40 < rxr> glibc-2.3.5 still building for sparc32
21:40 < rxr> my "thesun" box is up right now if you feel the need to test s.th. on a sparc ...
21:41 < minto> mnemoc: is there another way to test swig stuff. Installing svk involves to much work.
21:41 < mnemoc> minto: no idea :(
21:42 < mnemoc> minto: but cpan should do it for you
21:42 * minto doesn't know how to use CPAN
21:43 < mnemoc> cpan
21:43 < mnemoc> install svk
21:43 < minto> Does T2 have a cpan package>
21:43 < minto> ?
21:44 < minto> Uhmm, part of perl
21:44 < mnemoc> :)
21:50 -!- jsaw [~jsaw@volans.mpimf-heidelberg.mpg.de] has joined #t2
21:55 < rxr> hi jsaw
21:59 < mnemoc> hi jsaw
22:04 < CIA-9> amery * r8537 /trunk/package/gnome2/gtk-sharp2/ (gtk-sharp2.conf gtk-sharp2.desc): * updated gtk-sharp2 (r41822 -> 1.9.3)
22:09 < CIA-9> amery * r8538 /branches/2.1/package/gnome2/gtk-sharp2/ (gtk-sharp2.conf gtk-sharp2.desc): * merged 8537 from trunk - updated gtk-sharp2 (r41822 -> 1.9.3)
22:09 < rxr> mnemoc: does it build ?=
22:10 < rxr> http://exactcode.de/rene/hidden/gcc-article/2005-gcc-4.0/
22:10 < rxr> http://exactcode.de/rene/hidden/gcc-article/2005-gcc-4.0/stat2-rt.png
22:10 < rxr> Os is very bad with C++ and 4.0.0
22:10 < rxr> I guess not much inlining of the heavy tempalte code happening
22:10 < mnemoc> rxr: yes
22:11 < rxr> lu == -funroll-loops
22:11 < mnemoc> rxr: r41822 was failing on aclocal (wth?) but 1.9.3 is from april 12 and builds
22:11 < rxr> ofp == -fommit-frame-pointer
22:11 < rxr> -vect == vectorization applied
22:12 < rxr> as you can see 4.0 tempalte based C++ code can get twice as fast as with the already good 3.4.3 ... !
22:12 < mnemoc> your 'hidden' is now known by google monster
22:12 < rxr> but hidden does not index the dir
22:13 < rxr> that is the hidden subparts stay hidden ...
22:13 < mnemoc> ic
22:13 < mnemoc> what do you use for graphics?
22:15 < rxr> ploticus
22:15 < rxr> wow - -funit-at-a-time can also give a speedup ...
22:15 < rxr> uaat not yet in the graphic
22:27 < jsaw> hi guys
22:28 < rxr> hey jsaw !
22:28 < jsaw> hi rxr
22:29 < jsaw> have to leave again, going home, cu later
22:49 < jsaw> back home :)
22:52 < valentin> jsaw: wow - that was fast
22:52 < jsaw> hehe
22:52 < jsaw> hi valentin
22:52 < valentin> hi :)
22:53 < jsaw> Currently saving energy by using the car...
22:53 < jsaw> well, my energy :)
22:53 < valentin> hehe
22:54 < valentin> here in berlin using public transportation does not save much energy
22:54 < valentin> the city trains / subways are powered at 750 V
22:55 < valentin> do not ask me how much power is transformed into heat
22:55 < jsaw> tsss, yeah
22:55 < jsaw> how are the descendants?
22:56 < valentin> plural ?
22:56 < jsaw> who knows?
22:56 < valentin> i need to bring my daughter to bed now - together with my girl :)
22:56 < jsaw> (no, only used as idiom)
22:56 < valentin> jsaw: i saw the ultrasonic
22:57 < jsaw> :)
22:57 < valentin> my second child would be very good in hiding
22:57 < jsaw> exciting, isn't it
22:57 < valentin> so - have to leave now. maybe back later - maybe not
22:57 < valentin> yes, very exciting :)
22:57 < jsaw> cu valentin
22:57 < valentin> cu later
23:00 < rxr> mnemoc: the glibc built on sparc32 with v8 optimization
23:02 < rxr> are you sure the subway is powered by 750 V ?
23:02 < rxr> I somehow doubt that ...
23:06 -!- minto [~chatzilla@82-217-66-7.cable.quicknet.nl] has quit [Read error: 110 (Connection timed out)]
23:16 -!- sparc-kly [~sparc@64.237.138.134] has quit ["Leaving"]
23:18 < mnemoc> re
23:25 < rxr> hi
23:27 -!- Man0l0 [~mm@71-87-28.dial.terra.cl] has joined #t2
23:28 < mnemoc> hi
23:30 * rxr shower and then starting the crawling into bed
23:34 -!- Man0l0 [~mm@71-87-28.dial.terra.cl] has left #t2 []
23:44 -!- Mech [~chatzilla@dsl081-073-058.sfo1.dsl.speakeasy.net] has joined #t2
23:44 < Mech> whee!
23:44 < rxr> ~http://gcc.fyxm.net/summit/2003/mudflap.pdf
23:45 < mnemoc> Mech: hi Mr. iPkg
23:45 < Mech> mnemoc: Hey :-)
23:45 < Mech> mnemoc: you know - I still haven't played with iPkg ...
23:46 < Mech> mnemoc: right now, I'm trying to figure out how to handle build options for my lil toy system
23:46 < Mech> mnemoc: Just looking at curl - there's about 20 options that ./configure can take - krb libs, ssl libs, etc.
23:47 < mnemoc> Mech: what you tried T2's embeded target?
23:47 < mnemoc> have*
23:47 < mnemoc> rxr: interesting paper
23:47 < Mech> mnemoc: I played with it a lil...
23:47 < rxr> mnemoc: the gcc 4.0 feature I got curious about
23:48 < Mech> so I'm trying to figure out what to google for - there's gonna be an explosion in the number configuration options, and I'm trying to find research on it...
23:48 < Mech> I figured somebody working with weather data, or maybe physics data must have run across this problem before
23:49 < Mech> unfortunately, I can't find the 'right' keywords to feed google :-(
23:53 < valentin> re
23:56 < Mech> Hmm - it still bugs me scratchbox is gonna use GAR format for its recipes/meta-data :-( There's already several formats in use for this data - T2, OE, etc etc :-(
23:56 < mnemoc> rxr: the graphic effect you added to Emerge-Pkg makes me smile everytime i see it :D
23:56 < valentin> rxr: Technisch handelt es sich um zwei verschiedene Bahnnetze. Beide Netze benutzen die Normalspur (1435 mm Gleisbreite) und fahren mit Gleichstrom mit einer Spannung von 750 Volt.
23:56 < Mech> too bad they can't share the meta-data - it would lighten the load on everyone and make the meta-data more reliable
23:56 * mnemoc kicks valentin
23:57 < rxr> DC? Have they been drining too much ?!?
23:58 < rxr> mnemoc: good that I made you more happy ,-)
23:58 < mnemoc> rxr: thanks :)
23:59 < rxr> with updated .cache files with my reduced dependency noise Emerge will rock even more 8-)
23:59 < mnemoc> will you want my -rc1 cache files?
23:59 < rxr> yes
--- Log closed Wed Apr 27 00:00:06 2005