--- Log opened Fri Dec 31 00:00:21 2004
--- Day changed Fri Dec 31 2004
00:00 < rxr> the signal would be that noisy that an normal digital voltmeter would show pure crap?
00:00 < jsaw> no noisy, but spikey.
00:00 < rxr> yeah
00:00 < rxr> hm
00:01 < jsaw> so, unless you have a rms voltmeter... it'll show crap indeed
00:01 < rxr> hm - I have an RMS at home - you believe it would display something useful?
00:01 < jsaw> if you have a diode and a capacitor...
00:01 < rxr> yeah - sure ...
00:01 < jsaw> rxr: it should.
00:02 < rxr> the question is whether the main power-suply is defect
00:02 < jsaw> but the striking part is the low voltage.
00:02 < rxr> or this downconverter board
00:02 < rxr> (or all :-()
00:02 < jsaw> the first thing I'd assume, that the thing dropped and you have some cold solder point.
00:02 < rxr> and the other question what the normal input range of the downconverted board should be - the pins are quite tiny - so it can not sustain a high current ...
00:03 < rxr> the case is quite unused and unscratchy
00:03 < jsaw> but high voltage...
00:04 < rxr> yep - this is why I wonder why they (apple) power it from the main switching power suply
00:04 < rxr> and do not feed Ac net power to it ..
00:05 < rxr> and I wonder why a high voltage is routed thru the whole logic board with the PPC cpu and such ...
00:05 < jsaw> I assume the switching coil is in the main supply, but technically part of the downconverter...
00:05 < rxr> well the downconverted also has many transformators, spulen (what the he en name) capacitors and so on
00:06 < rxr> enough parts to make a whole independant switching power suply ...
00:06 < jsaw> coil = spule
00:06 < rxr> no idea what apple soldered together there ...
00:06 < rxr> a yes ;-)
00:07 < jsaw> better get your oscilloscope back before you continue...
00:07 < rxr> well - if I can not get the power-suply operational any time soon I just plug the logic board onto an normal ATX power-suply ..
00:07 < rxr> and throw the CRT and co away ...
00:08 < rxr> (hopeing that the logic board is not defective as well ...)
00:08 < jonhbruce> repeats always die at build of binutils, always the message to check system clock and the binutils download time shows ok.
00:08 < jsaw> rxr: If this is an option, then I'd start re-soldering whatever looks not 100% okay.
00:09 < rxr> ok - /me in the background not interrupting build discussions ...
00:09 < rxr> thanks jsaw
00:09 < jsaw> jonhbruce: what tar version?
00:09 < jsaw> rxr: np
00:11 < jonhbruce> tar-1.14-4 on fc3 host
00:12 < mnemoc> rxr, jsaw: config.in thing is ready, do you want to review on ML or can i commit (1. implementation, 2. adaptions to current files)?
00:12 < mnemoc> it's very simple
00:13 < mnemoc> 3. drop %include which calls nested awk-s around Config
00:20 < jsaw> jonhbruce: I can't find a hint...
00:21 < mnemoc> damaged tar due to fedora's patch-flood?
00:22 < jsaw> ... maybe mnemoc's right, could you try a fresher tar?
00:23 < jonhbruce> something is setting an ls alias
00:24 < jsaw> oh.
00:30 < rxr> many sites sugest the flyback transformator get's often defective ...
00:31 < rxr> anyone experiece with such component defects?
00:31 < rxr> does those suck up a lot energy in those cases?
00:31 < rxr> or more likely tend to go into inf. ohm range?
00:31 < mnemoc> jsaw, rxr: no signal for my patch? ... i'll go home now... cu :'(
00:32 < jsaw> mnemoc: ups, got lost in the midst of debugging and electronics.
00:32 < jsaw> mnemoc: please send to ml...
00:32 < mnemoc> jsaw: ok
00:33 < mnemoc> jsaw: but please, reply something :)
00:33 < jsaw> :)
00:34 -!- nzg [~tschmidt@p508EAC0E.dip.t-dialin.net] has joined #t2
00:36 < mnemoc> hi nzg!
00:36 < nzg> hi mnemoc
00:36 < mnemoc> nice to see you around
00:37 < nzg> thx, you too.
00:38 < nzg> my boss definitely owes me
00:38 < mnemoc> why? :)
00:39 < nzg> i just wired this damn building just to get more phones in....
00:41 < nzg> so... now i drive home...
00:43 < nzg> k, i have still 1hr
00:44 < nzg> mnemoc: a friend of me is currently in chile
00:57 -!- nzg [~tschmidt@p508EAC0E.dip.t-dialin.net] has quit ["Verlassend"]
01:04 < jonhbruce> installed tar 1.15 but no change.
01:07 -!- nzg [~tschmidt@p508EAC0E.dip.t-dialin.net] has joined #t2
01:08 < jsaw> jonhbruce: I have no idea - and currently not more time. Maybe you can post a few details to the mailing list.
01:08 < rxr> re
01:09 < rxr> mnemoc: you can commit ... ;-)
01:09 < rxr> sorry overread the line
01:11 < jonhbruce> Thanks for all the suggestions, will post and sleep on it. cheers
01:11 < jsaw> cu jonhbruce
01:11 -!- jonhbruce [~john@host81-157-242-89.range81-157.btcentralplus.com] has quit ["Leaving"]
01:15 < mnemoc> rxr: i just mailed it :\
01:15 < mnemoc> nzg: where? why?
01:17 < nzg> where i don't knwo exactly somwhere near the atakarma(?)
01:17 < nzg> he was born in chile and his step-parents gave him this trip for his a-level
01:18 < mnemoc> atacama :)
01:19 < mnemoc> chile is great for vacations
01:19 < mnemoc> jsaw: sent
01:20 < nzg> once i've got the money, i going to visit chile ;-)
01:20 < mnemoc> =)
01:21 < mnemoc> if i get money someday i'll go to visit germany :)
01:23 < jsaw> ouhm:
01:23 < jsaw> root@volans:/scratch/src/t2# ./src.efsd.1104362748.5413.d9959e34/debug.sh
01:23 < jsaw> File size: 61026
01:23 < jsaw> debug-efsd:[TOOLCHAIN]# ls
01:23 < jsaw> . badfiles.txt debug.buildenv dependencies.debug fl_wrapper.rlog untar.txt
01:23 < jsaw> .. build.pid debug.hooks dependencies.txt fl_wrapper.wlog xsrcdir.txt
01:23 < jsaw> ERROR-LOG chroot.sh debug.sh efsd-2004-02-25 flist.txt
01:23 < jsaw> archdir cmd_wrapper.log debug_x.sh files.lst install_wrapper.log
01:23 < jsaw> debug-efsd:[TOOLCHAIN]#
01:23 < jsaw> no segfault
01:24 < jsaw> and that because I inserted
01:24 < jsaw> fprintf(stderr, "File size: %d\n", file_size); fflush(stderr);
01:24 < jsaw> into bash-3.0/builtins/evalfile.c
01:26 < mnemoc> jsaw: the first weirdness is on the prompt, your PWD is not TOOLCAHIN, it is TOOLCHAIN/src.foo....
01:26 < jsaw> cd somewhere, cd back, and voila:
01:26 < jsaw> debug-efsd:[src.efsd.1104362748.5413.d9959e34]#
01:26 < jsaw> harhar
01:26 < mnemoc> but that's the first sign of 'something wrong on bthis shell'
01:27 < rxr> the prompt might be defect due to my -rcfile migration ...
01:27 < mnemoc> yes, it was an effect of that
01:27 < mnemoc> but it's weord :)
01:28 < rxr> we endured the bash bug so long now - I think we can wait the few more days until we fix it, report it upstream, or whatever ...
01:28 < rxr> I think we should work with the bash 3.0 in favour of the very old bash2 code base ...
01:29 < jsaw> the working directory mismatch is because PWD gets assigned
01:29 < jsaw> I removed it from debug.buildenv, and everything is alright
01:29 * mnemoc but the lack of ./debug.sh makes harder to fix packages
01:29 < mnemoc> jsaw: :D
01:30 < jsaw> ok, now removing patch, and trying this again.
01:30 < jsaw> <- starts thinking there's something wrong with read-only variables
01:31 < rxr> hm - interesting potn
01:31 * rxr in bed in some secons
01:31 < rxr> since I'll be off to hannover early tomorrow ...
01:34 < mnemoc> what's on hannover?
01:38 < jsaw> It's the PWD assignment.
01:39 < mnemoc> what?
01:39 < mnemoc> the whole problem is due to PWD assignation?
01:39 < jsaw> try to remove it and the run debug.sh
01:41 < jsaw> hm. no. on one path it works on another not.
01:42 < rxr> it is random hashtable corruption
01:42 < rxr> but maybe jsaw is right and it is due to assigning build-ins or so ...
01:43 < jsaw> one thing about the PWD. The source says, if it is found on startup, the wd is adapted according to PWD. This is not the case.
01:44 < mnemoc> .oO( i was not too lost on screaming about $PS1 )o
01:53 < jsaw> I removed as much actually read-only variables as possible. Now it works again.
01:56 < mnemoc> we may need a cleanup rutine after dumping env then
01:56 < rxr> bash should actually never core-dump ...
01:57 < jsaw> right. but...
01:57 < mnemoc> but?
01:57 < jsaw> but we may solve the problem faster with a work-around then finding the problem with bash internals...
01:58 < rxr> yes -s ure
01:58 < rxr> I just wanted to remind that even if the removal of those read-onlies workarouns the problem we should report it to the bash folkgs ...
01:58 < rxr> folks even
01:58 < rxr> arrgh
01:58 < mnemoc> we can work, and you can report the bug upstream :)
01:58 < rxr> yes sure
01:59 < rxr> imagine you do a lot of stuff and mistype a var and boom
01:59 < rxr> kiss your history good bye ...
01:59 < rxr> jsaw: but your observation might be still random
01:59 < jsaw> yep.
02:00 < jsaw> I'll send the list of variables I removed to ML.
02:00 < rxr> when I took a look the day before yesterday the heisenbug apeared and reapeared removing one complex function definition after the other
02:00 < mnemoc> jsaw: can i assume you agree the design of the rework of config.in sth?
02:00 < rxr> disapeared the first one should have been
02:00 < rxr> mnemoc: ack btw.
02:01 < mnemoc> rxr: one positive vote :)
02:02 < jsaw> mnemoc: two
02:02 < mnemoc> great
02:03 < mnemoc> wait the two commits before re-update
02:05 < CIA-9> mnemoc * r5327 /trunk/scripts/ (Config config.func config.in): * reworked config.in loading on Config
02:05 < CIA-9> rene * r5328 /trunk/package/network/coda/coda.desc:
02:05 < CIA-9> * updated coda (6.0.3 -> 6.0.8) includes sub-packages updates:
02:05 < CIA-9> lwp (1.10 -> 1.12), rvm (1.8 -> 1.10) and rpc2 (1.20 -> 1.25)
02:08 < rxr> gcc:
02:08 < rxr> >The 4.0 compiler will be enough
02:08 < rxr> >> faster than 3.4 that we'll brag about it in the release notes,
02:08 < rxr> Let's not do that, it's only true for C++.
02:08 < rxr> For C we are slower as usual ;-)
02:08 < mnemoc> two more commits i will need
02:09 < mnemoc> "slower as usual"?
02:10 < rxr> yeah - the optimizing gets complexer in each release
02:10 < CIA-9> rene * r5329 /trunk/package/ (filesystem/coda/ network/coda/): * moved network/coda into filesystem/
02:10 < rxr> and the C++ parser faster and generating less "noise" read temporary stuff and co ...
02:10 < rxr> plus the C++ frontend was quite slow in the past
02:11 < rxr> it is not such a wonder they can speed it up here and there ..
02:11 < rxr> hm - the rock IRC channel is quite silent these days - I browsed the logs in the afternoon ...
02:12 < rxr> I especially like our commit graph:
02:12 < rxr> http://svn.exactcode.de/svn.png
02:13 < mnemoc> *click*
02:15 < CIA-9> jsaw * r5330 /trunk/scripts/Build-Pkg: * fix an argument list overflow of /bin/cat in Build-Pkg
02:16 < jsaw> ./scripts/Config: line 94: fg: no job control
02:16 < CIA-9> mnemoc * r5331 /trunk/ (5 files in 4 dirs): * addapted subconfig.in loading
02:16 < jsaw> <- anybody seen this
02:16 < jsaw> (while ./scripts/Config -oldconfig)
02:16 < mnemoc> rxr: why the overflow on the top of the graphic
02:17 < rxr> because I force the autoscaling to not full-scale the tiny projects would suffer too much ... ;-)
02:18 < jsaw> somebody please try "./scripts/Config -oldconfig"
02:18 * rxr is not at :HEAD and wants sleep ...
02:18 < rxr> ccache of e.g. mozilla based stuff looks way better now ...
02:19 < rxr> yep - seems to work quite nice now ...
02:20 < jsaw> mnemoc: there's something wrong with the Config.
02:20 < jsaw> a lot of variables have disappeared.
02:20 < jsaw> e.g ROCKCFG_OUTPUT_TERM_COLOR
02:21 < jsaw> or ROCKCFG_INIT_STYLE
02:21 < mnemoc> wait for the text patch
02:21 < mnemoc> the renames
02:22 < jsaw> okay. I wait...
02:23 < CIA-9> mnemoc * r5332 /trunk/ (18 files in 9 dirs): * renamed config.in files to be expert and noexpert on new schema
02:24 < mnemoc> now you can diff outputs :)
02:25 < mnemoc> errr..
02:25 < mnemoc> bug
02:25 < jsaw> Creating configuration script.
02:25 < jsaw> T2 2.1.0-beta3 configuration ...
02:25 < jsaw> ./scripts/Config: line 32: misc/output/*/config.in: No such file or directory
02:25 < jsaw> ./scripts/Config: line 94: src/default/subconfig-gcc.in: No such file or directory
02:25 < jsaw> ./scripts/Config: line 36: [: missing `]'
02:25 < jsaw> ./scripts/Config: line 59: [: missing `]'
02:25 < jsaw> New config written to config/default/*.
02:25 < jsaw> Cleaning up. Configuration finished.
02:25 < CIA-9> rene * r5333 /trunk/package/develop/ccache/ (libtool-options.patch parse-config): (log message trimmed)
02:25 < CIA-9> * various ccache fixes:
02:25 < CIA-9> * fixes my privous libtool-options.patch
02:25 < CIA-9> * enabled CCACHE_UNIFY so the comments and whitespaces are removed
02:25 < CIA-9> and unified in the preprocessed source (this e.g. fixes to cache
02:25 < CIA-9> packages like mozilla and co where the *config.h includes absolute
02:25 < CIA-9> path names that contains the temporary T2 build dir src.* and thus
02:26 < mnemoc> yes
02:26 < rxr> "got cached result for" just filling my thunderbird rebuild logs ..
02:27 < rxr> I guess I have a nearly 100% compiled from cache thunderbird in the morning
02:27 < rxr> I consider ccache fully operational now - at least I have fixed all packages that failed to cache now ...
02:27 -!- nzg [~tschmidt@p508EAC0E.dip.t-dialin.net] has quit ["Verlassend"]
02:28 < mnemoc> aaarg.... 15/90 link to a 64kbps link... sucks
02:28 < rxr> if you notice one failing to built up nearly 100% out of the cache for consecutive rebuilds please tell me
02:28 < rxr> so n8 all - have fun and cu in the early morning I hope
02:28 < jsaw> cu rxr
02:32 < mnemoc> almost done
02:32 < mnemoc> cu rxr
02:35 < CIA-9> mnemoc * r5334 /trunk/ (5 files in 5 dirs): * fix types on last config.in addaptions
02:35 < mnemoc> remove -x on just Commited scripts/Config :\
02:37 * mnemoc diffing
02:38 < jsaw> bbl
02:38 < mnemoc> they are identical :)
02:38 * mnemoc happy
02:39 < mnemoc> cu jsaw
02:46 -!- rxr_ [~rene@p213.54.199.60.tisdip.tiscali.de] has joined #t2
02:46 -!- Topic for #t2: T2 | 2.1.0-beta2 RELEASED | The next generation of System Development Enviroments (SDE) | http://www.exactcode.de/t2
02:46 -!- Topic set by mnemoc [] [Tue Dec 28 14:55:44 2004]
02:46 [Users #t2]
02:46 [ _Ragnar_] [ daja77] [ mnemoc] [ rxr ] [ sparc-kly]
02:46 [ CIA-9 ] [ jsaw ] [ N0V4K ] [ rxr_] [ valentin ]
02:46 -!- Irssi: #t2: Total of 10 nicks [0 ops, 0 halfops, 0 voices, 10 normal]
02:46 -!- Channel #t2 created Sun Aug 8 21:15:33 2004
02:46 -!- Irssi: Join to #t2 was synced in 11 secs
02:48 < CIA-9> mnemoc * r5335 /trunk/ (15 files in 8 dirs): * renamed config.hlp files accordingly
02:48 < mnemoc> finally
02:49 < mnemoc> it's not good to depend on an stolen and poor signal when you are fixing core brokeness
03:03 -!- rxr [~rene@p213.54.196.223.tisdip.tiscali.de] has quit [Read error: 110 (Connection timed out)]
03:23 -!- sparc-kly [~mubex@66.50.161.111] has quit [Read error: 104 (Connection reset by peer)]
03:25 -!- sparc-kly [~sparc@66.50.161.111] has joined #t2
03:38 < CIA-9> mnemoc * r5336 /trunk/package/ (3 files in 3 dirs): * added checksum for scons xvattr and nvidia driver
06:33 -!- sparc-kly [~sparc@66.50.161.111] has quit [Read error: 104 (Connection reset by peer)]
06:34 -!- sparc-kly [~sparc@64.237.130.154] has joined #t2
06:45 -!- N0V4K [~N0V4K@246.Red-80-33-149.pooles.rima-tde.net] has left #t2 ["Adios"]
08:42 < rxr_> moin
08:42 -!- You're now known as rxr
08:43 < rxr> he ccache finally works good:
08:43 < rxr> first halfly cached mozilla build:
08:43 < rxr> err thunderbird even:
08:46 < rxr> from a build where ot aborted due to full disk:
08:46 < rxr> -> Compiler Cache Hits while pkg build: 96.79% (2382 hits, 79 misses)
08:46 < rxr> a rebuild:
08:46 < rxr> -> Compiler Cache Hits while pkg build: 99.43% (2447 hits, 14 misses)
08:47 < rxr> buildinf firefox thereafter:
08:47 < rxr> -> Compiler Cache Hits while pkg build: 73.03% (1690 hits, 624 misses)
08:47 < rxr> and then mozilla:
08:47 < rxr> -> Compiler Cache Hits while pkg build: 61.98% (1702 hits, 1044 misses)
08:47 < rxr> isn't that great?
08:47 < rxr> previously it was cache missing >90% ....
12:15 -!- freddy [~junix@pD9E77CD1.dip.t-dialin.net] has joined #t2
12:15 < freddy> hi @ll
12:59 < mnemoc> rehi
13:00 < freddy> hi mnemoc
13:00 < mnemoc> hi freddy
13:00 < mnemoc> rxr: have ccache people answered to your improvements?
13:10 * mnemoc needs more harddisk :( generic-full-builddir+ccache is damn big
13:34 < rxr> re
13:34 < rxr> I have not mailt the patches yet
13:34 < rxr> and maybe the libtool one it not needed anymore with the unification enabled ...
13:36 -!- junix_ [~junix@pD9E777D8.dip.t-dialin.net] has joined #t2
13:39 < rxr> hi junix_
13:42 < rxr> hey - we are on http://symlink.ch
13:44 -!- freddy [~junix@pD9E77CD1.dip.t-dialin.net] has quit [Read error: 60 (Operation timed out)]
13:48 < jsaw> wow
13:48 < jsaw> hi all
13:49 < rxr> what - we have tcpdump 3.8.3 since 2004-04-23 and it is on freshmeat, today?
13:52 < rxr> http://www.kde-apps.org/poll/index.php <- ogg vorbis leading
13:55 < jsaw> harhar
13:55 < jsaw> gcc -o conftest -I/usr/include/freetype2 -g -O2 -lz -lm conftest.c -lpng >&5
13:55 < jsaw> /usr/lib/gcc/i586-unknown-linux-gnu/3.4.3/../../../libpng.so: undefined reference to `deflate'
13:56 < jsaw> when will ppl learn that libs have to be appended after the files to compile
13:57 < rxr> never
13:58 < rxr> I wonder how they can release it - it seems to just work on other patched compilers .
13:58 < rxr> ..
13:58 < jsaw> :/
14:01 < rxr> btw such a cached thunderbird builds in 45 minutes on my 800Mhz iBook
14:02 < rxr> normally it takes nearly 2h IIRC
14:02 < rxr> and in this 45 minutes most time is spend in configure and our post-install information gathering ... :-(
14:02 < jsaw> :)
14:02 < rxr> next I can polish distcc ;-)
14:03 < rxr> as well as cross builds ...
14:03 < jsaw> I thought about the copyright tags. What's if we simply keep Commercial and add "Restricted" to make qmail and mnemoc :) happy
14:04 < rxr> btw we could hack a progress bar estimation by couting the gcc calls and estimate a progress bar value based on it - if we like to have such an feature ...
14:05 < rxr> it only fails to work for packages that do not call a compiler often ...
14:05 < rxr> like header or data pacakges ..
14:05 < jsaw> we can actually add open calls to source files
14:05 < jsaw> or call wrapped in fl_wrapper generally
14:05 < rxr> another possibility to display a percentage according to the chached build log line count ...
14:06 < jsaw> not everybody has the space to use ccache
14:06 < rxr> this is not ccache dependant ...
14:06 < jsaw> but a counter for every wrapped call should be easier...
14:06 < rxr> I mean the t2 build log ...
14:06 < jsaw> o, cached, not ccached.
14:06 < rxr> altough counts might be less overhead compared to linecouting a huge file
14:07 < rxr> I need to lear to hit the key a bit more ..
14:07 < rxr> so many chars missing in the last days ..
14:07 < jsaw> hehe
14:07 < rxr> maybe I'm nore sensitive to the keyboard now that some juice is damping the keys ...
14:07 < jsaw> as long as the code does not miss 'em
14:07 < rxr> need to clean it again on some water tap ...
14:07 < jsaw> what about the copyright tag?
14:08 < rxr> yeah - we need to do it soon ...
14:08 < jsaw> 14:05 < jsaw> I thought about the copyright tags. What's if we simply keep
14:08 < jsaw> Commercial and add "Restricted" to make qmail and mnemoc :) happy
14:08 < rxr> I now hack on this scons build thing to not bypass the gcc wrapper, so that I can commit the blender update and then I can do the copyright implanter script overhaul ...
14:08 < jsaw> And for both, "Commercial" and "Restricted" the check/warning is issued
14:09 < rxr> well - this one is not my primary focus ... - sound ok - but I do not use commercial / restricted licenses too often
14:09 < jsaw> no, I'm at it. Only asking for the tags.
14:09 < rxr> so I'm not the first choice voter for that topic ...
14:09 < rxr> I meant the ROCK-*-NOTE vs. T2-*-NOTE
14:09 < rxr> or are you at that, too?
14:09 < jsaw> ah so.
14:09 < jsaw> no.
14:10 < jsaw> ac_link='$CC -o conftest$ac_exeext $CFLAGS $CPPFLAGS $LDFLAGS conftest.$ac_ext $LIBS >&5'
14:10 < rxr> better discuss the [L] tag usage with people needing it - I only need it for the M$-core-fonts and the nvidia and ati binary X driver
14:10 < jsaw> It's autoconf that is wrong...
14:10 < rxr> I'll most probably not ues many other non-free packages often - if ever
14:11 < jsaw> xv, the wincodecs and a few other.
14:11 < jsaw> I have afpl-ghostscript and some others in queue
14:11 < rxr> I do not utilize xv - neither the w32codes since I nearly never us x86 for desktops these days ...
14:12 < rxr> or xanim -...
14:12 < jsaw> I only wanted to state that we have it.
14:12 < rxr> but xanim and xv are both packages I see no real value in ...
14:12 < jsaw> And for companies it might be interesting too.
14:12 < jsaw> rxr: that's a different question...
14:13 < rxr> yeah - but I do not know all the license details that would bring me into a position to judge for the needed abstractions ...
14:13 < jsaw> okay.
14:13 < jsaw> I'll do the code and post it.
14:14 < rxr> your simplificatin sounds good - either impement it and wait until s.th. does not fit - or wait a bit until mnemoc can tell which packages do not fit into the scheme ...
14:14 < rxr> jups ;-)
14:14 < rxr> so /me on scon s...
14:15 < jsaw> ./scripts/Config -oldconfig ; ./scripts/Create-ErrList -newdelete
14:16 < jsaw> <- seems not to work anymore
14:16 < rxr> uhm
14:17 < jsaw> oh, shouted to soon. Misplaced a patch...
14:20 < jsaw> no, I did not shout too soon. :(
14:23 < jsaw> wtf?
14:23 < jsaw> grep '^(ROCK Linux|T2) Package Source Checksum: '
14:23 < jsaw> does not match?
14:24 < jsaw> but
14:24 < jsaw> grep '^\(ROCK Linux\|T2\) Package Source Checksum: '
14:24 < jsaw> is okay?
14:24 < jsaw> huh?
14:26 < jsaw> ah, MNEMOC
14:28 < CIA-9> jsaw * r5337 /trunk/scripts/Create-ErrList:
14:28 < CIA-9> * fix scripts/Create-ErrList -newdelete (from previous r5326):
14:28 < CIA-9> (|) must be escaped for grep
14:28 < jsaw> oh, there are more refs
14:31 < CIA-9> jsaw * r5338 /trunk/scripts/ (Create-UpdList Emerge-Pkg): * more scripts/ fixes: escaping (|) in grep
14:51 < mnemoc> re
14:51 < mnemoc> jsaw: sorry :\
14:53 < mnemoc> why that didn't failed on my tests here?
14:55 < jsaw> hmm. what's your grep version?
14:55 < jsaw> grep (GNU grep) 2.5.1
14:56 < mnemoc> the same
14:56 < mnemoc> :/
14:58 < mnemoc> jsaw: is Config working good for you?
15:00 < jsaw> mnemoc: yes
15:00 < jsaw> brb
15:02 < mnemoc> jsaw: found!, i used grep -e on my tests
15:02 < CIA-9> rene * r5339 /trunk/package/develop/scons/environment.patch:
15:02 < CIA-9> * fixed scons to inherit the build environment to have the PATH to
15:02 < CIA-9> the wrappers as well as their configuration
15:04 < CIA-9> rene * r5340 /trunk/package/multimedia/blender/ (blender.conf compile.patch non-static-python-extras.patch):
15:04 < CIA-9> * migrated blender to use the scons build system instead of the not
15:04 < CIA-9> so well maintianed NaN Makefiles that need more and more hacks ...
15:05 < mnemoc> hi rxr
15:08 < rxr> hi mnemoc
15:09 < mnemoc> rxr: i think ccache is not an expert option, what do you thing? what about others like distcc?
15:10 < jsaw> mnemoc: ah, that's why
15:11 < jsaw> gotta leave. cu later!
15:17 < rxr> mnemoc: yep - both normal user options
15:17 < mnemoc> rxr: ok
15:42 < rxr> blender build with fixed sconcs so that ccache is used:
15:42 < rxr> -> Compiler Cache Hits while pkg build: 8.77% (54 hits, 562 misses)
15:42 < mnemoc> uhm
15:42 < rxr> now building blender 2.36 to commit the update
15:42 < rxr> let's see how much a update caches ...
15:43 < mnemoc> that 8.77% was the a rebuild with a different tool?
15:43 < rxr> nope - first build cache
15:44 < rxr> due to compiling simillar stuff than other packages - e.g. for tests or so ..
15:44 < mnemoc> then it's good
15:44 < rxr> currently I do not log the cache misses or hits - so I can not tell which files cached
15:44 < rxr> but if you are that interested I can reenable logging - clear the cache and so build it from scratch
15:44 < mnemoc> no need
15:45 < rxr> == 15:43:06 =[9]=> Building multimedia/blender [2.36 2.1.0-beta3].
15:45 < rxr> so /me taking a short nap - so tired
15:48 < CIA-9> mnemoc * r5341 /trunk/scripts/config.in: * adapted scripts/config.in's workflow described on top
15:49 -!- madtux [~mike@200.91.101.98] has joined #t2
15:49 < madtux> hello.
15:52 < mnemoc> hi mightytux
16:24 < mnemoc> rocklinux decided to use $extraversion in packages starting from 1 (as redhat/fedora does)
16:25 < mnemoc> The rene repository will vanish in two steps: First step will be
16:25 < mnemoc> integration of some packages into base. Then a call for maintainers will
16:25 < mnemoc> be sent out to the mailing list.
16:25 < mnemoc> Assigned to: clifford and fake (first step)
17:17 < rxr> so - re
17:18 < rxr> -> Compiler Cache Hits while pkg build: 26.62% (168 hits, 463 misses)
17:20 < CIA-9> rene * r5342 /trunk/package/multimedia/blender/blender.desc: * updated blender (2.34 -> 2.36)
17:22 < mnemoc> rxr: have you seen this awk preprosesor we have on Config ?
17:22 < mnemoc> src/rockconfig.awk
17:23 -!- madtux [~mike@200.91.101.98] has quit ["leaving"]
17:25 < rxr> do not remember off hand ...
17:25 < rxr> what does this mean (rock bof): as soon as
17:25 < rxr> full control over the websites has been regained.
17:25 < rxr> ?
17:26 < mnemoc> no idea
17:26 < rxr> what the heck, are they crazzy:
17:26 < rxr> Frontend needs a major overhaul. For now, create a default
17:26 < rxr> /etc/network/config file and make network configuration call $EDITOR
17:26 < rxr> /etc/network/config.
17:26 < rxr> don't tell me this is an improvement ...
17:26 < mnemoc> *G*
17:26 < rxr> huh:
17:26 < rxr> - - fix install procedure
17:26 < rxr>
17:26 < rxr> It's not broken, so we don't fix it. It needs big documentation, though.
17:27 < rxr> Assigned to: esden and blindy
17:27 < rxr> ????
17:27 < mnemoc> "ok ok ok.... the shit sucks but does the job, so it's not _broken_"
17:29 < rxr> I should not read the CCC stuff on the rock list - somehow it makes my stomarch turn over
17:29 < mnemoc> it's kind of funny :)
17:31 < rxr> the rock list is quite spam flooded ...
17:32 < mnemoc> uhm?
17:32 < mnemoc> i have never received spam from there
17:33 < rxr> take a look into lurker
17:33 < rxr> it is quite full
17:33 < rxr> softeare there viagra here ...
17:33 < rxr> maybe your server drops it ..
17:34 < mnemoc> i guess
17:36 < rxr> blender rebuild:
17:36 < rxr> -> Compiler Cache Hits while pkg build: 99.68% (629 hits, 2 misses)
17:36 < mnemoc> :D
17:36 < mnemoc> rxr: can i remove %include crap?
17:37 < mnemoc> this awk preprocesor looks quite stupid to me
17:38 < rxr> parly cached blender build: 27min
17:38 < rxr> mnemoc: I take a look in some minutes++
17:39 < mnemoc> rxr: and the rebuild?
17:39 < rxr> fully caced blender rebuild: 10min
17:39 < rxr> cached even ...
17:39 < mnemoc> OT: which is the file with the country names?
17:39 < rxr> but the first build has 26.62% hits already so the real speedup is even bigger
17:40 < mnemoc> oh
17:40 < rxr> usr/share/misc/countries
17:40 < mnemoc> what about firefox and thunderbird after mozilla?
17:40 < mnemoc> thanks
17:40 < rxr> I posted in the morning stats that parly show this ...
17:41 < rxr> firefox or thunderbird after a mozilla build should cache in the range of 60-70% ...
17:41 < rxr> hopefully the rck people do not pick up the ccahe and scon improvements that fast ...
17:41 < rxr> maybe they never notice ...
17:41 < mnemoc> :D
17:42 < mnemoc> daja77 is listening :)
17:42 < rxr> do not type his anme then ;-)
17:42 < rxr> name even
17:42 < mnemoc> oops
17:42 < mnemoc> daja*
17:42 < rxr> this is why I intentionally mistyped the affected peaces ...
17:42 < mnemoc> :(
17:43 < rxr> ;-)
17:43 < rxr> we could update gentoo 0.11.54
17:44 < rxr> but hopefully it is not daja just sneaking for code ...
17:45 < mnemoc> :)
17:45 < rxr> I heared the hacky live cd stuff in rock was some success on the congress
17:46 < rxr> I should spend some time in my semester holidays to add a really engeneered livecd to t2 ...
17:46 < mnemoc> valentin?
17:46 < rxr> yeah - valentin was tehre ...
17:46 < mnemoc> and a decent bootdisk
17:46 < rxr> there even
17:46 < rxr> you mean real boot_disk_, not CD?
17:46 < rxr> without freedos?
17:46 < rxr> the livecd is really a thing people want and is cool to give away ...
17:47 < rxr> but the rock one we have and the updated rock one quite stink ...
17:47 < mnemoc> rxr: both, and maybe busybox instead of kiss
17:48 < mnemoc> rxr: udhcp, ....
17:48 < rxr> yeah - we should polish it ..
17:48 < mnemoc> ash script instead of hardcoded linuxrc :)
17:50 < rxr> not sure if the ash script would scale - but we can try
17:50 < mnemoc> ash script + c++ helpers? :)
17:51 * rxr showering after my evening nap ..
18:02 < valentin> mnemoc: ?
18:02 < mnemoc> valentin: hi
18:02 < valentin> hi
18:03 < valentin> just saw this yellow line in irssi
18:03 < mnemoc> :)
18:03 < valentin> have some stuff to do, we have guests for new year celebration
18:03 < mnemoc> rene told rock livecd was a success on 21c3, and i thought you had told him
18:03 < mnemoc> valentin: sure
18:03 < valentin> yep
18:04 < valentin> they arive in about 116 minutes
18:04 < mnemoc> go go go
18:07 < valentin> i only have to get rid of some old dust (esp the on on the hardware :))
18:07 < mnemoc> :D
18:10 < valentin> btw: anyone here in 2005 already ?
18:10 < mnemoc> 2pm here
18:12 < valentin> dark here in germany, and many morons are out there on the streat shooting heavy firework
18:12 < valentin> many ppl here have illegal firework from poland ....
18:13 < mnemoc> there are 'official firework events' or neither?
18:13 < valentin> yes, and there is legal firework you can buy in the stores
18:13 < mnemoc> aa... then common people can handle fireworks
18:14 < valentin> but some explosions are quite heavy - 500 meters away and everything shaking here - do not think that are legal bombs
18:14 < valentin> i fear my window will break :(
18:14 < mnemoc> oh, that heavy
18:15 < valentin> yep - sounds like artillery :(
18:15 < valentin> not the loud sharp bang of german crackers, but more like WOOOOOOUUUM
18:15 < valentin> they sell such stuff in poland
18:16 < mnemoc> here only military artillers can handle fireworks legally, and main cities do public events under military supervision
18:16 < mnemoc> of course there are ilegal stuff for sale
18:16 < mnemoc> with about 200-300 burned children per year
18:17 < rxr> here normally childrens do nto get burned ... :-/
18:17 < rxr> normally ...
18:17 -!- junix_ is now known as freddy
18:17 < freddy> hi rxr
18:18 < rxr> hi freddy
18:18 < valentin> the stuff you can legaly buy here in germany will not kill you, but if you have bad luck you can loose an eye or so
18:18 < freddy> juten rutsch
18:18 < rxr> how are you ?
18:18 < valentin> hi freddy
18:18 < mnemoc> stupid parents let children play with ilegal stuff :\
18:18 < freddy> hi valentin
18:18 < rxr> get well into the new year, too freddy - I hope you do attend some party or so in berlin?
18:19 < freddy> i think russian come over oder
18:19 < freddy> i m in berlin
18:19 < freddy> man geht das hier ab
18:20 -!- sparc-kly [~sparc@64.237.130.154] has quit [Read error: 54 (Connection reset by peer)]
18:23 < valentin> yep
18:23 < rxr> freddy: sorry to interrupt partying ... some work stuff: I do not think we have microcontoller on monday (we would have had the las mon before the holidays ...) so we should do this atmel stuff next week, hm?
18:23 < valentin> i cannot open my win anymore
18:23 < valentin> too much smoke
18:23 < valentin> and it is 18:24 ...
18:24 < freddy> yeap
18:24 < freddy> @rxr query
18:33 < mnemoc> there was an asault here
18:34 < mnemoc> bbl
18:40 < CIA-9> rene * r5343 /trunk/scripts/config.in: * substituted ROCK Config ID with Config ID
18:40 < mnemoc> re
18:42 < rxr> mnemoc: I svn up'ed, I still have all below the expert option switch, is this normal or do I need to debug it ?
18:43 < mnemoc> yes
18:43 < mnemoc> i just ported what was there before
18:43 < mnemoc> no modifications yet
18:45 < mnemoc> i wanted to drop %include first.... but i can postpone that desire
18:45 < mnemoc> i'm mounting a second mirror o chile btw
18:45 < mnemoc> on*
18:46 < rxr> hm - the file I can crash our bash3 with does not crash the freebsd amd64 box ...
18:46 < rxr> freddy: are there some builds logs? or where are those ports lying so I can take a look how it (bash) is built?
18:48 < freddy> re
18:55 < mnemoc> rxr: http://t2.geeks.cl/mirror/2.1/
18:55 < rxr> freebsd does pass --without-bash-malloc as wee do - but does not seg fault
18:56 < rxr> new url?
18:56 < mnemoc> aditional
18:56 < rxr> another box ?
18:56 < mnemoc> yes
18:56 < mnemoc> better bw, i hope
18:56 < rxr> cool
18:57 < rxr> freddy: is the old x86 nexus still available ...
18:57 < freddy> yeap
18:57 < freddy> .1
18:58 < rxr> ah - ;-)
18:59 < CIA-9> rene * r5344 /trunk/package/perl/perl/perl.desc:
18:59 < CIA-9> * updated perl (5.8.3 -> 5.8.6), took over maintainance and updated
18:59 < CIA-9> the package description
19:02 < mnemoc> rxr: the old on my box got a new location: http://www.geeks.cl/t2-mirror/2.1/ ... but the one in use still valid, for now.
19:04 < rxr> the other^Wnew box is also hosted in chile?
19:05 < mnemoc> yes
19:07 < rxr> hm - on the x86 freebsd box the bash does not segfault ...
19:07 < rxr> maybe it is a bug in our glibc's malloc?
19:07 < rxr> :-(
19:07 < mnemoc> bad randomly-choosen-day?
19:07 < rxr> yeah - maybe
19:08 < rxr> they have a 2.3 branch now - maybe I should give it a try ..
19:08 < rxr> mnemoc: you do not use rock 2.1 ?
19:08 < mnemoc> nope
19:08 < rxr> 2.0 ?
19:08 < mnemoc> i have lot of 2.0 machines :\
19:09 < mnemoc> welll not lot, just 9
19:09 < rxr> I read (in lurker) th merged bash 3.0 into stable?
19:09 < rxr> do you have bash3 on a 2.0 box ?
19:09 < mnemoc> they wanted
19:09 < mnemoc> nope
19:09 < mnemoc> i have 2.01
19:09 < mnemoc> 2.0.1
19:09 < rxr> hm - ok - I ggo experimental
19:09 < mnemoc> plus updates
19:09 < rxr> I Emerge it on a remote box with old 1.7 rock ...
19:10 < mnemoc> :D
19:13 < mnemoc> rxr: on http://t2.geeks.cl/mirror/ i'll have copies of your snapshots
19:15 < rxr> do you setup some cron job or so to mirror ?
19:17 < mnemoc> to mirror the download? rsync from my box which is updated by hand
19:17 < mnemoc> mirror/ will have stuff on dl.ec.de/t2 grabbed by hand for now
19:17 < mnemoc> mi i plan to cron an update.sh to do that things
19:17 < mnemoc> my plan*
19:18 < mnemoc> err... i plan*
19:20 < rxr> and the source snapshots ?
19:20 < rxr> damn uni net ...
19:21 < mnemoc> i'm doing it by hand
19:21 < mnemoc> i have never used wget to mirror :)
19:22 < rxr> there are some integrated suits to handle ftp / http mirroring ...
19:22 < rxr> we even have some in t2 (but I never used them myself, since I only push stuff to my own server via rsync and to not need to pull from other sites ...)
19:24 < mnemoc> as i keep my local download perfectly up2date i don't need to pull stuff from yours :)
19:25 < mnemoc> the snapshots is another topic
19:26 < mnemoc> i think i'll tweak something on cron
19:33 < rxr> the glibc homepage is a hoke
19:33 < rxr> joke
19:33 < rxr> matches the release planning and policy - or more the lack of
19:33 < rxr> the page does not even list a cvs location ...
19:34 < mnemoc> aproa
19:34 < rxr> ah - ok - resources does ...
19:34 < mnemoc> :p
19:34 < mnemoc> ctrl-S
19:34 < rxr> that is a feature, you know ?
19:35 < mnemoc> yes :\ great feature
19:35 < rxr> it is scroll lock or so
19:35 < rxr> at least I now finally know what scroll lock is about ...
19:35 < mnemoc> yes
19:35 < mnemoc> :D
19:35 < rxr> since my first days I wondered what it should do and what people would need it for ...
19:36 < mnemoc> can i silence curl with arguments?
19:36 < rxr> most probably ...
19:36 < rxr> hm
19:37 < rxr> on my rmote box with this libc:
19:37 < rxr> Compiled by GNU CC version 3.2.2.
19:37 < rxr> Compiled on a Linux 2.4.20 system on 2003-02-06.
19:37 < rxr> bash does not crash on the file that crahes my iBook ...
19:37 < rxr> I double check if the file crashes my athlon at home ...
19:39 < mnemoc> good, the two guys and the client shooted are ok
19:40 < rxr> hm - that file does not crash my home athlons box...
19:40 < rxr> well - but somehow I get the feeling that it might be the glibc
19:40 < rxr> e.g. I also did not saw an indication that rock 2.1 suffers this bash crashes ...
19:41 < mnemoc> daja said they have bash3 as bash3 (optional thing)
19:41 < mnemoc> a pkgfork
19:43 < rxr> glibc rocks - take this commit message:
19:43 < rxr> Updated to fedora-glibc-20041213T2323
19:43 < rxr> and it includes code changes in malloc.c
19:44 < mnemoc> *G*
19:44 < rxr> I hate redhat and fedora - and especially drepper
19:44 < rxr> and this one - december 11th:
19:44 < rxr> _int_realloc): Add checks for corrupted memory.
19:44 < rxr> (_int_free): Make clear message are result of free() calls.
19:44 < rxr> and this is another one:
19:44 < rxr> Updated to fedora-glibc-20041210T0634
19:45 < rxr> are they crazzy ...
19:45 < rxr> glibc is the crappiest project arround
19:45 < freddy> hehe
19:45 < rxr> I never saw such a chaos, crap, and company driver enforcements ...
19:45 < rxr> every time I have to touch glibc I dream about a glibc free system ...
19:46 < rxr> and I really do nto get why the bsd folks ported the crappy beast over to freebsd and co kernels ...
19:46 < mnemoc> rxr: http://t2.geeks.cl/update.sh <--- this is grabing your snapshots
19:47 < rxr> is t2. very new? I get an unknown host ehre ..
19:47 < rxr> here even
19:48 < mnemoc> yes, i created just when i told you
19:48 < rxr> ah ...
19:48 < mnemoc> but here it resolved some seconds later
19:48 < rxr> hm - half of the glibc commits are in a fedora-branch ...
19:48 < mnemoc> o_O
19:48 < rxr> might need some mintues -> hours to propagate I guess ...
19:49 < rxr> mnemoc: you own geeks.cl ?
19:49 < mnemoc> i use my ISP as resolver
19:49 < mnemoc> yes
19:49 < mnemoc> geeks.cl innernet.cl and geekpreview.org are mine
19:49 < mnemoc> all abandoned :\
19:50 < mnemoc> some day i will do something else than mail
19:50 < rxr> ,-)
19:52 < mnemoc> INFO: Setting proxy to proxy.gtdinternet.com:8080.
19:52 < mnemoc> INFO: Auto-detecting best mirror ...
19:52 < mnemoc> INFO: Downloading mirror-list from exactcode.de.
19:52 < mnemoc> INFO: Testing ... ok 24263.606
19:52 < mnemoc> INFO: Testing ... ok 21636.184
19:52 < mnemoc> INFO: Testing ... ok 21188.490
19:52 < mnemoc> INFO: Testing ... ok 48117.474
19:52 < mnemoc> INFO: Using mirror <http://t2.geeks.cl/mirror/2.1>
19:52 < mnemoc> www.geeks.cl is on my lan but i did it using ISP's proxy
19:53 < mnemoc> so the test is kind of honest
19:56 < rxr> cool ;-)
19:56 < rxr> I test this one now
19:56 < rxr> cvs -z4 -Q -d :pserver:anoncvs@sources.redhat.com:/cvs/glibc co -P -r glibc-2_3-branch -D 2004-12-24 libc
19:56 < mnemoc> glibc-xmas2004.tar.bz2 :)
19:58 < rxr> yeah - let's what they present us ...
20:00 < mnemoc> if it work can you push it into nexus? cvs co-s are expensive
20:01 < rxr> hm - the branch checkout does not seem to work good ...
20:01 < rxr> yes, sure
20:01 < rxr> if I comit it into trunk it will be on the mirrors, soon
20:01 < rxr> I can even trigger a manual sync if you are awake at that time and want to test
20:01 < mnemoc> i prefer, _before_ :)
20:01 < rxr> ic
20:01 < mnemoc> 16:00 here :)
20:02 < rxr> ok - then I push it manually everywhere ...
20:02 < mnemoc> at office until 20:00
20:02 < rxr> ok - as soon as I have a useful checkout that basically builds I put it online ...
20:02 < mnemoc> 20:30 to be realistic
20:02 < rxr> 20:02 here ...
20:02 < mnemoc> ok
20:02 < rxr> well - here is evening meal in some minutes ...
20:02 < rxr> and the brahcn checkout does not yet yield data ...
20:02 < rxr> damn cvs stuff ...
20:02 < rxr> so it might be past 20:30 ...
20:02 < mnemoc> you will be on 1005 and /me at the office :\
20:03 < mnemoc> 2005*
20:03 < rxr> ok - not it yielded the files
20:03 < rxr> it was in quiet mode ...
20:03 < rxr> when will you be at the office?
20:03 < mnemoc> arg, hddtemp.db was updated
20:04 < mnemoc> in four hours
20:04 < rxr> damn file ...
20:04 < mnemoc> i hate unversioned crap
20:04 < rxr> yep
20:04 < rxr> we should ick the author ...
20:04 < rxr> kick even ..
20:05 < rxr> will you be in the office in the night?
20:05 < mnemoc> he is not the only one
20:05 < rxr> sure not
20:05 < mnemoc> i work until 20:00, i.e. now+4:00
20:05 * rxr is typing a bit as well ...
20:05 < rxr> ah ;-)
20:06 < mnemoc> + some hugs, 20:30
20:13 < CIA-9> mnemoc * r5345 /trunk/package/base/hddtemp/hddtemp.desc: * updated hddtemp.db (2004-12-19 -> 2004-12-26)
20:13 -!- freddy [~junix@pD9E777D8.dip.t-dialin.net] has quit [Read error: 54 (Connection reset by peer)]
20:13 -!- freddy [~junix@pD9E777D8.dip.t-dialin.net] has joined #t2
20:30 < rxr> mnemoc: shouldn't be the file on the mirror already?
20:30 < mnemoc> yes
20:30 < mnemoc> but i was downloading directly
20:31 < rxr> huh?
20:31 < rxr> !> make[1]: *** [resolv/subdir_install] Segmentation fault ..
20:32 < mnemoc> err
20:32 < rxr> core: ELF 32-bit LSB core file Intel 80386, version 1 (SYSV), SVR4-style, SVR4-style, from 'make'
20:32 < CIA-9> mnemoc * r5346 /trunk/scripts/Download: * Introduced download/Me as a $mirror_name to skip during mirror selection
20:35 < rxr> aiii
20:35 < rxr> !> patching file Makeconfig
20:35 < rxr> !> Hunk #1 FAILED at 522.
20:35 < rxr> !> 1 out of 1 hunk FAILED -- saving rejects to file Makecon ..
20:35 < rxr> this glibc is fun ...
20:36 -!- junix_ [~junix@pD95D0BB8.dip.t-dialin.net] has joined #t2
20:36 < mnemoc> :D
20:38 -!- freddy [~junix@pD9E777D8.dip.t-dialin.net] has quit [Read error: 60 (Operation timed out)]
20:46 < rxr> ok - with two glasses of wine my alcohol level reached the level to get back to glibc
20:46 < mnemoc> =)
20:46 < mnemoc> chilean wine? :)
20:46 < rxr> no - german
20:46 < rxr> and two sweet for my taste ..
20:46 < mnemoc> :'
20:47 < rxr> but my father likes those
20:49 < mnemoc> we are exporing champagne to germany too
20:50 < rxr> ;-)
20:50 < rxr> youare are in the office of this winery and more?
20:50 < mnemoc> we=chilean :)
20:51 < rxr> mnemoc: you seem one of the only t2 devel avail here right now ..
20:51 < rxr> I'll update the apache and subversion software on exactcode rigth now
20:51 < rxr> see note on the list
20:51 < rxr> I hope it does not interrupt your ...
20:51 < rxr> you even
20:52 < rxr> I guess it is the unnoticed update of the year at this time ..
20:52 < mnemoc> ok
20:52 < mnemoc> fsfs?
20:57 < rxr> nope
20:57 < rxr> just normal update
20:57 < rxr> fsfs is to experimental for my taste on production systems
20:57 < rxr> I know the bdb backend and it works ...
20:57 < rxr> so I stay for it - normal update procedure ...
20:57 < rxr> and the backward diff does not seem to be a bdb bottlenbecks
20:58 < rxr> bottleneck even
20:58 < rxr> e.g. the load on the box is pretty low - it seems some version incompatibility or so ..
20:58 < rxr> I review in detail if it still persists after the 1.0.6 -> 1.1.2 update ..
20:58 < mnemoc> don't forget to restart your apache after that :)
21:00 < rxr> == 12/31/04 20:59:19 =[9]=> Finished building package apache.
21:00 < rxr> yep - sure
21:09 < rxr> ouhm
21:09 < rxr> == 12/31/04 21:09:17 =[9]=> Aborted building package subversion.
21:09 < rxr> but just the python bindings ...
21:11 < rxr> svn checkout works
21:12 < mnemoc> =)
21:12 < rxr> Powered by Subversion version 1.1.2 (r12471).
21:12 < rxr> so ;-)
21:12 < rxr> we are online again
21:13 < rxr> == 12/31/04 21:11:49 =[0]=> Finished building package glibc.
21:14 < mnemoc> thx
21:15 < mnemoc> let's see how it performece
21:17 < rxr> hm - it can not win prizes performance wise ...
21:17 < rxr> :-(
21:17 < rxr> seems I have some new task for my todo ...
21:18 < rxr> the server is mostly idle during backward diffs ...
21:18 < rxr> Cpu(s): 0.3% us, 0.3% sy, 0.0% ni, 98.0% id, 1.3% wa, 0.0% hi, 0.0% si
21:18 < rxr> but I think it is way faster now ...
21:18 < mnemoc> i want that stats here :)
21:18 < rxr> time svn diff -r3883:3882
21:19 < rxr> real 0m6.300s
21:19 < rxr> user 0m0.030s
21:19 < rxr> sys 0m0.010s
21:19 < rxr> I think it was more in the 20s range before
21:19 < rxr> wasn't it ?
21:19 < mnemoc> w8
21:19 < mnemoc> it has 1m4s here...
21:19 < mnemoc> was*
21:20 < mnemoc> and now it is:
21:20 < mnemoc> real 0m12.974s
21:20 < mnemoc> user 0m0.040s
21:20 < mnemoc> sys 0m0.030s
21:21 < rxr> that looks good ,-)
21:21 < mnemoc> yeah :)
21:21 < mnemoc> very reasonable
21:22 < rxr> locally it is way faster
21:22 < rxr> the 6s was on my home athlon
21:22 < rxr> on the server locally it is this now:
21:22 < rxr> real 0m1.330s
21:22 < rxr> user 0m0.064s
21:22 < rxr> sys 0m0.018s
21:23 < mnemoc> that's cheating :)
21:24 -!- junix_ [~junix@pD95D0BB8.dip.t-dialin.net] has quit [Read error: 104 (Connection reset by peer)]
21:24 -!- junix_ [~junix@pD95D0BB8.dip.t-dialin.net] has joined #t2
21:24 < rxr> http://ec-outpost.exactcode.de:/glibc.patch
21:24 < rxr> http://ec-outpost.exactcode.de:/glibc-2.3.4-2004-12-24.tar.bz2
21:24 < rxr> --typos
21:25 < rxr> -:
21:25 < rxr> s/exactcode.de/dyndns.org
21:25 < rxr> damn
21:25 < mnemoc> glibc.patch?
21:25 < rxr> http://ec-outpost.dyndns.org/glibc.patch
21:25 < mnemoc> what about a version on it? :)
21:26 < rxr> http://ec-outpost.dyndns.org/glibc-2.3.4-2004-12-24.tar.bz2
21:26 < rxr> the glibc.patch is relateive to trunk
21:26 < rxr> not to be added to base/glibc/ ...
21:26 < rxr> patch -p0 < ... in t2-trunk/
21:27 < rxr> only barly tested
21:27 < mnemoc> :D
21:27 < rxr> handle with care
21:27 < mnemoc> i have a sandbox tree :)
21:27 < rxr> glibc.patch 100% 2156 2.1KB/s 00:00
21:27 < rxr> glibc-2.3.4-2004-12-24.tar.bz2 100% 13MB 3.3MB/s 00:04
21:27 < mnemoc> 13MB?
21:30 < mnemoc> i remembered glibc bigger
21:32 < rxr> -rw-r--r-- 1 root root 13674452 Nov 23 09:54 glibc-2.3.4-2004-08-01.tar.bz2
21:32 < rxr> uncompressed it is big ...
21:33 < mnemoc> where did you rsync-ed it?
21:33 < mnemoc> 3.3MB/s is nice :)
21:34 < mnemoc> # lftpget http://ec-outpost.dyndns.org/glibc-2.3.4-2004-12-24.tar.bz2
21:34 < mnemoc> `/glibc-2.3.4-2004-12-24.tar.bz2' at 409968 (2%) 1.6K/s eta:86m [Receiving data]
21:43 < rxr> on a local ethernet of course ...
21:44 < rxr> other locations I get such transfer rates is inter university DFN netowrk links ...
21:44 < rxr> s/is/are/
21:44 < rxr> when you download you grab it from my home box
21:44 < rxr> and the line is already saturated with my own download ...
21:44 < rxr> you should get a boost when my finished
21:45 < rxr> 85% [==============================> ] 11,837,124 8.47K/s ETA 03:13
21:52 < rxr> 100%[====================================>] 13,841,317 11.46K/s ETA 00:00
21:52 < rxr> now it should be faster over there ..
21:52 < mnemoc> new url?
21:52 < rxr> nope - no new url
21:53 < mnemoc> # lftpget http://ec-outpost.dyndns.org/glibc-2.3.4-2004-12-24.tar.bz2
21:53 < mnemoc> `/glibc-2.3.4-2004-12-24.tar.bz2' at 409968 (2%) 1.6K/s eta:86m [Receiving data]
21:53 < mnemoc> err
21:53 < mnemoc> # lftpget -c http://ec-outpost.dyndns.org/glibc-2.3.4-2004-12-24.tar.bz2
21:54 < mnemoc> `/glibc-2.3.4-2004-12-24.tar.bz2' at 6221124 (44%) 12.0K/s eta:10m [Receiving data]
21:54 < mnemoc> that :)
21:55 < CIA-9> rene * r5347 /trunk/scripts/Build-Tools:
21:55 < CIA-9> * only consider regular files during the updated/modified files search
21:55 < CIA-9> e.g. not the directory modification time ...
22:01 < mnemoc> good change
22:01 < mnemoc> ehm... no
22:01 < mnemoc> bad change
22:02 < mnemoc> if you remove or rename a file (patch)
22:02 < rxr> when you rename a file the new file will be newer
22:02 < rxr> and a patch does not affect the config options
22:03 < rxr> it even does: ! -name '*.patch*'
22:03 < mnemoc> if you remove a config.in :)
22:03 < mnemoc> ok ok you win
22:03 < rxr> then the new file (using svn mv) will have a newer date
22:03 < rxr> the only case is that it might cause "problems" is when you rm a config*
22:04 < rxr> but it would only affect the processing if other existing configure or .conf still use the rm'ed CFG option
22:04 < rxr> which would then still be present with the old value as saved in config/$cfg/config
22:04 < rxr> instead of beeing undefined
22:05 < rxr> but I think it is a reasonable prize for less config and tool reprocessing
22:05 < rxr> isn't it?
22:05 < mnemoc> yes
22:05 < rxr> I find the continous Tool rebuilds quite annoying ...
22:05 < mnemoc> if you remove a config.in :) ok ok you win
22:05 < mnemoc> small price bug benefit
22:06 < mnemoc> big*
22:06 < rxr> e.g. when you svn up and due to meta updates e.g. due to file modifcations or addions - whatever - that do not further interfere the config processing modify the dentry you get a tool rebuild
22:07 < rxr> and this is quite often the case ...
22:08 < mnemoc> yes
22:08 < rxr> and we need to lock the tools rebuild
22:09 < rxr> especially during the historic rock^Wt2 cluster build the nodes rm -rf the tools's timestamp and get into a tools rebuild orgy
22:09 < mnemoc> are they rebuilt into $file or $file.$$ ?
22:10 < rxr> of course we want to get rid of the old cluster build, but when you run to Build-Pkg in parallel locally you still get to toosl builds that wipe it's files each other ...
22:10 < rxr> might even be $file
22:10 -!- junix_ [~junix@pD95D0BB8.dip.t-dialin.net] has quit [Read error: 104 (Connection reset by peer)]
22:10 < rxr> but it should be locked and the 2nd one just wait until they are built ...
22:11 < rxr> I thnk
22:11 < rxr> think
22:11 < rxr> damn - I need to water my keyboard again - or press a bit harder again
22:11 < mnemoc> or use src/$config
22:12 < mnemoc> src/$config/
22:13 < rxr> == 12/31/04 21:59:13 =[1]=> Finished building package glibc.
22:13 < mnemoc> :D
22:13 < mnemoc> fresh build'
22:13 < mnemoc> ?
22:14 < rxr> GNU C Library stable release version 2.3.4, by Roland McGrath et al.
22:14 < rxr> Compiled by GNU CC version 3.4.3.
22:14 < rxr> Compiled on a Linux 2.6.8.1-dist system on 2004-12-31.
22:14 < rxr> yes
22:14 < rxr> fresh install build
22:14 < rxr> and I build on my running ibook
22:14 < rxr> if it is to defectify I'll be offline soon ...
22:14 < rxr> too even
22:14 < mnemoc> :)
22:15 < mnemoc> what will we do with linux headers?
22:16 < mnemoc> 2.6.8.1 until 2.6.11?
22:16 < mnemoc> patch files to build on 2.6.10?
22:16 < mnemoc> patch 2.6.10 to let packages build?
22:17 < rxr> we should never patch massive ammounts of user-space packages
22:17 < rxr> if the kernel headers are defect or opose pure junk they needs to be patched ...
22:17 < mnemoc> so (a) or (c) ?
22:17 < rxr> we could go .10 and patch them up ...
22:18 < mnemoc> for beta3 or beta4?
22:18 < rxr> but if it is quite some work and nonone is willing to do this - we could stay a bit longer on 8.1 ...
22:18 < mnemoc> a branch?
22:18 < rxr> branch for .10 fixup?
22:19 < mnemoc> a branch of package/base/linux-header to fixup
22:19 < rxr> well - if people are motivated to fix the headers up ..
22:19 < mnemoc> considering we choose (c)
22:19 < rxr> well - so far I o
22:19 < rxr> mostly see you beeing motivated for this path
22:19 < rxr> maybe join discussion on linuxe-kernel
22:19 < mnemoc> i have the motivation but not the knowledge
22:20 < mnemoc> lkml answer is well known
22:20 < rxr> well - just see what does not compile and tweak the headers accordingly - this is normally not that hard
22:20 < rxr> it was recently rediscussed
22:20 < mnemoc> yes? and the answer changed?
22:21 < rxr> and Linus - as usually - sayed (sometimes even worldy) "I don't care - it is the kernel - and user-space that does include the headers is broken"
22:21 < rxr> but this is plain stupid
22:21 < mnemoc> ACK
22:22 < rxr> somewhere the ABI (Application Binary Interface) needs to be defined
22:22 < mnemoc> yes
22:22 < rxr> and I - oposing the great guru - think if there is some kernel it needs to define the ABI to the outside
22:22 < rxr> whether using unified headers using #ifdef __KERNEL__ or two sets
22:23 < rxr> linux-$ver/include linux-$ver/include-usr (or so - just the example) is not to me to defice
22:23 < mnemoc> what fedora did?
22:23 < rxr> decide even
22:23 < mnemoc> i prefer #ifdef __KERNEL__
22:23 < rxr> but maintaining a different set is just error prone and loosing sync
22:23 < mnemoc> or even different headers on the same dif
22:23 < mnemoc> dir*
22:24 < rxr> redhat^Wfedora historically has thsi sanitized-kernel-header thing ...
22:24 < mnemoc> any clean patch we can stole^Wtake?
22:24 < rxr> some big tarball I think ...
22:24 < mnemoc> puaj
22:24 < rxr> google for sanitized kernel headers
22:24 < mnemoc> debian?
22:24 < rxr> you will find a lot
22:25 < rxr> I do not monitor debian that closely the overhead to get to the right source place is too high
22:25 < rxr> with the new fedory cvs this is quite nice to review their codebase - this is why I posted the link
22:26 < mnemoc> yes
22:28 * mnemoc greping his logs
22:29 < mnemoc> http://cvs.fedora.redhat.com/
22:29 < mnemoc> :)
22:29 < mnemoc> i shouldn't have compressed them :\
22:30 * mnemoc writing a bzgrep
22:31 < mnemoc> why do they use cvs??
22:33 < rxr> historic cruft
22:33 < rxr> we do most use cvs ...
22:33 < rxr> :-(
22:34 < mnemoc> yes, but you drop it two-three years ago?
22:34 -!- N0V4K [~N0V4K@246.Red-80-33-149.pooles.rima-tde.net] has joined #t2
22:34 < mnemoc> hi N0V4K
22:35 < N0V4K> mnemoc, buenas tio
22:36 < mnemoc> rxr: how is your english today?
22:36 < mnemoc> err
22:36 < mnemoc> N0V4K: how is your english today?
22:36 < mnemoc> rxr: sorry
22:36 < N0V4K> in my bead
22:37 < mnemoc> o_O
22:38 < mnemoc> rxr: http://cvs.fedora.redhat.com/viewcvs/devel/kernel/ <--- only three tiny patches for 2.6.10
22:38 < rxr> maybe ..
22:39 < rxr> no
22:40 < rxr> it seems they apply all the patches there to .10
22:40 < rxr> the versions seems to indicate when they intoduced them
22:40 < rxr> linux-2.6.0-must_check.patch
22:40 < rxr> 1.18
22:40 < rxr> 6 days
22:40 < rxr> davej
22:40 < rxr> rebase to 2.6.10
22:40 < mnemoc> oh
22:41 < rxr> linux-2.6.3-crash-driver.patch
22:41 < rxr> 1.3
22:41 < rxr> 2 months
22:41 < rxr> davej
22:41 < rxr> net/disk dump patch-o-rama.
22:41 < rxr> ah - ok:
22:41 < rxr> +config CRASH
22:41 < rxr> + tristate "Crash Utility memory driver"
22:42 < rxr> aiiiiiiiiiiiiiiii
22:42 < rxr> -> Building. Writing output to $root/var/adm/logs/9-glibc.out
22:42 < rxr> cat: write error: No space left on device
22:42 < mnemoc> :D
22:42 < rxr> good that I have ccache enab led ...
22:42 < mnemoc> ccache :)
22:42 < rxr> just that the damn glibc Makefiles are that slow ...
22:42 < mnemoc> ccache filled your disk
22:43 < rxr> not really
22:43 < rxr> I /only/ have 30Gb in the iBook - and for the latest 10.3 installation I increated the osx space to 10GB ...
22:44 < mnemoc> rxr: branches/$foo/{linux-header,glibc} ?
22:45 < rxr> well - let's discuss this on the list of there are enough people for this
22:46 < mnemoc> ok
22:47 < rxr> (e.g. I'm not that motivated right now ... - and have a quite full TODO for january)
22:49 < mnemoc> me too :)
22:50 < rxr> this are the fedory kernel headers.
22:50 < rxr> 4a1eb7e6e342fabb29a59e9425ff0d73 glibc-kernheaders.tar.bz2
22:50 < rxr> quite cool versioned, hm?
22:50 < rxr> and such commit messages:
22:50 < mnemoc> :)
22:50 < rxr> latest and greatest from upstream repository
22:50 < rxr> wow
22:50 < mnemoc> very explicative
22:50 < rxr> how many information it carries
22:54 < rxr> until glibc broke on my iBook due to out of space:
22:54 < rxr> -> Compiler Cache Hits while pkg build: 1.14% (29 hits, 2518 misses)
22:55 < mnemoc> 1% :\
22:55 < rxr> yeah - first build
22:56 < rxr> the 29 hits have most probably been configure test compiles ...
22:56 < rxr> at least me ./scripts/Cleanup rebuild will have 2518 hits ... more
22:57 < mnemoc> mail sent
22:57 < mnemoc> sure?
22:58 < rxr> yeah - after all the ccache work I'm sure
23:08 < rxr> ok:
23:08 < rxr> root@idefix:/home/rene# du -csh core*
23:08 < rxr> 2.4M core
23:08 < rxr> 47M core.20870
23:08 < rxr> 22M core.28893
23:08 < rxr> 39M core.29333
23:08 < rxr> 111M core.3601
23:08 < rxr> 625K core.4391
23:08 < rxr> 625K core.4396
23:08 < rxr> 221M total
23:09 < mnemoc> uhm
23:15 -!- mnemoc_ [~amery@200.75.27.11] has joined #t2
23:16 < rxr> 880M gsmp-trunk
23:16 < rxr> uhm2
23:17 < mnemoc_> after lot of cleaning yestarday i realiced i have no machine to do a full-generic build :\
23:17 < rxr> ouch
23:18 < mnemoc_> i was on 1197 of 13xx
23:18 < rxr> I also can not build powerpc right now, because my rs6k disks are full with enterprise video dumps ..
23:18 < mnemoc_> but i had no space to continue
23:18 < mnemoc_> :)
23:18 < rxr> but I have this new hitachi replaced disk to relocate my whole video/audio storage ...
23:19 -!- mnemoc [~amery@200.75.27.48] has quit [Nick collision from services.]
23:19 -!- mnemoc_ is now known as mnemoc
23:19 < rxr> I think I can start a 2.1.0-beat3 PowerPC build on 2 or 3rd january
23:19 < mnemoc> )
23:19 < mnemoc> =)
23:22 < rxr> 2.1G t2-trunk
23:22 < rxr> 1.1G xorg
23:23 < mnemoc> xorg?
23:23 < rxr> yeap
23:23 < rxr> ok - my iBook glibc is at iconvdata
23:23 < mnemoc> svn?
23:24 < rxr> nope - cvs of course
23:24 < mnemoc> .oO( we could use ccache data to add a progress bar to the builds )o
23:24 < rxr> we do not use ccache for this ... ;-)
23:25 < mnemoc> yes yes, i meaned checkouts
23:25 < mnemoc> but we could :)
23:25 < rxr> scrollback - we mentioned that this afternoon
23:25 < mnemoc> yes?
23:25 < rxr> yep
23:25 < rxr> maybe around 14.oo
23:25 < rxr> so 10.oo on your side or so ...
23:28 < mnemoc> :D
23:28 < rxr> my glibc is running post-install adaptions
23:29 < mnemoc> nice
23:29 < mnemoc> i guess you should do 2.6.10 there too, to have something to ./debug.sh :)
23:31 < rxr> ah - no /me blinded
23:31 < rxr> if - that is the big question
23:31 < rxr> if debug.sh is fixed then ...
23:32 < rxr> so - some cleanup:
23:32 < rxr> 14622012 13169836 1452176 91% /
23:32 < rxr> 9% more ...
23:33 < rxr> hm - there is a ATI Mobility _T2_ ;-)
23:34 < mnemoc> what do you think about t2sde.org?
23:35 < rxr> yeah - the questoin is with - ?
23:36 < rxr> t2-sde.org?
23:36 < rxr> or t2-project.org ?
23:37 < mnemoc> t2-project remembers me the first project i destroyed harbour-project.com
23:38 < mnemoc> my first approach to opensource... on 1999, i sent a mail due to some problems building it on linux (it was designed for dos)
23:38 < rxr> do nto know harbour ...
23:38 < mnemoc> a flamewar started and xharbour forked two weeks after that
23:39 < mnemoc> a clipper compiler
23:39 < mnemoc> both communities are agonic
23:39 < mnemoc> and dying day by day
23:40 < rxr> or - so better t2-sde or t2sde ;-)
23:40 < mnemoc> but t2-project.org sounds much better that t2sde.org
23:40 < mnemoc> as general rule i dislike '-'s, but looks better with
23:42 < mnemoc> that's another mail to set to /dev/null^Wt2ml
23:42 < mnemoc> maybe as selfreply to copyright one
23:44 < rxr> huh? t2 is not /dev/null!
23:45 < mnemoc> you don't answer my mails :)
23:46 < rxr> I don't ?
23:46 < rxr> I thought I do n
23:46 < rxr> mostly
23:46 < rxr> at least I just replied to the branch mail ... ;-)
23:46 < mnemoc> :D
23:46 < mnemoc> just got here
23:49 < rxr> cache size 602.8 Mbytes
23:49 < rxr> max cache size 976.6 Mbytes
23:49 < rxr> and this is just my local iBook rebulids
23:49 < rxr> not a generic build
23:49 < rxr> I think we should increate the ccache default of 1GB for t2 builds ...
23:50 < mnemoc> yes
23:50 < mnemoc> 1G is very low
23:50 < mnemoc> even an option on Config
23:50 < mnemoc> expert.
23:50 < rxr> # ./scripts/Create-ErrList -cfg install
23:50 < rxr> Error logs from install-2.1.0-beta3-x86-bootdisk:
23:50 < rxr> 161 builds total, 72 completed fine, 0 with errors.
23:51 < mnemoc> make something fail to try ./debug.sh :)
23:51 < rxr> yeah - will soon
23:51 < rxr> 23:51 here ...
23:51 -!- N0V4K [~N0V4K@246.Red-80-33-149.pooles.rima-tde.net] has quit [Remote closed the connection]
23:51 < mnemoc> cu next year rxr
23:52 < rxr> so first off to the balon (or the en equivalent) and a bit new year celebrating
23:52 < mnemoc> your family is waiting
23:52 < rxr> yeah - cu in 2005 when it started over here ...
23:53 < mnemoc> this Download -all on 'my' mail server is delaying arrivals importantly
23:54 < mnemoc> if your server is ntp-ed too, it took 17 minutes to get into mine
23:55 < rxr> my boxes "should" be ntp synced ...
23:55 < mnemoc> so mine is damn delayed on it's deliveries :)
23:56 < rxr> I double check the ntp conficuration in 2005 ;-)
23:56 < rxr> so - now my iBook glibc is sintalling
23:56 < rxr> maybe it finished this year
23:56 < mnemoc> 4 minutes left... GO!
23:56 < mnemoc> :D
23:57 < mnemoc> progress bar ;)
23:57 < rxr> I'm already "here"
23:57 < rxr> light book with w-lan .. all it needs ...
23:57 < mnemoc> :)
--- Log closed Sat Jan 01 00:00:46 2005