T2 IRC Log: 2004-11-25
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.
--- Log opened Thu Nov 25 00:00:27 2004
00:16 < rxr> enjoy:
00:16 < rxr> http://www.exactcode.de/t2/packages/
00:16 < rxr> of coursed linked in the frontpage menu ...
00:16 < rxr> and some remaining stuff needs to be done - I'm working on it
00:28 < rxr> ok - basica fixed, example of newly fixed file:
00:28 < rxr> http://www.exactcode.de/t2/packages/aboot.html
00:33 -!- kensai [~kensai@64.237.129.108] has joined #t2
00:46 -!- kensai [~kensai@64.237.129.108] has quit ["Leaving"]
00:49 -!- kensai [~kensai@64.237.129.108] has joined #t2
00:56 -!- kensai [~kensai@64.237.129.108] has quit ["Debian Rulez"]
00:57 < rxr> I just browse the exactcode statistic - and I'm impressed how much traffic the box pushes into the net ...
01:17 < CIA-9> rene * r4837 /trunk/package/x11/rxvt-unicode/rxvt-unicode.desc: * updated rxvt-unicode (4.0 -> 4.1) - orignial file vanished
01:25 < CIA-9> rene * r4838 /trunk/package/base/hddtemp/hddtemp.desc:
01:25 < CIA-9> * updated hddtemp (0.3-beta11 -> 0.3-beta12) and updated the URL tag
01:25 < CIA-9> - previous site and file vanished
01:26 < CIA-9> rene * r4839 /trunk/package/x11/fox/fox.desc: * updated fox (1.3.12 -> 1.3.19) - original file vanished
01:39 < CIA-9> rene * r4840 /trunk/package/scientific/alliance/alliance.desc:
01:39 < CIA-9> * updated alliance (5.0-20040909 -> 5.0-20040928) - orignial file
01:39 < CIA-9> vanished
01:41 < rxr> 130(!) patches in submatser - they will have fun....
01:44 < rxr> no pcmcia on sparc... afaics
01:44 < jsaw> re
01:44 < rxr> ??? - of course sparc book had PCMCIA ...
01:44 < jsaw> hi rxr
01:44 < rxr> the rock submaster is a big trash deposit ...
01:44 < rxr> hi jsaw
01:45 < rxr> we have a preleminary package matrix ;-)
01:45 < jsaw> I've seen your comment but haven't had a look a it... let's see
01:46 < CIA-9> rene * r4841 /trunk/package/gnome2/gal/gal.desc:
01:46 < CIA-9> * fixed the gal URL to use the usual GNOME mirror - not a non-existing
01:46 < CIA-9> one so our master mirror can catch it ;-)
01:47 < jsaw> oh!
01:47 < jsaw> cool!
01:48 < rxr> hm - damn - on the mirror the nautilus file download always stops after 219.220 bytes ...
01:48 < rxr> hm - no - randomlly ...
01:50 < rxr> jsaw: how are you?
01:51 < rxr> hm - the gnome server is strange
01:57 < rxr> well - some random gnome mirror works as all the other download for T2 do ...
01:59 < jsaw> rxr: well, not so good today. Had a bit problems with my heart, but obviously nothing wrong on the organ level. "Herzbeklemmung" - most probably stress related.
02:00 < jsaw> going to make a "Belastungs EKG" tomorrow, but I guess, nothing will be found -> have to get a few days of holidays
02:00 < rxr> ic - then better relax a bit ... !
02:00 < rxr> XFce 4.2 - the Future is Now!
02:01 < rxr> ^- update anyone but jsaw ;-)
02:01 < jsaw> I do. Programming is relaxing ... :)
02:01 < jsaw> is xfce 4.2 ready?
02:01 < jsaw> or a rc?
02:02 < rxr> seems - ready - but just the osnews release I posted ...
02:02 < rxr> I saw a -rc before ... - so this must be final ...
02:03 < jsaw> the last entry on xfce.org is
02:03 < jsaw> 2004/11/14 - Xfce 4.2 Release Candidate 1 is out
02:04 < rxr> ouhm - I know why I do not like news sites ... they keep claiming releases there there are none more often with every day
02:04 < rxr> gcc was also announced when it hit the first master server and the gcc people intended to release it two days later ...
02:05 < rxr> and with mozilla they do the same ... - freak out when the first binary hit the server - but the source is still missing :-(
02:05 < jsaw> ...
02:08 < jsaw> btw xfce, there's a graphical source installer, maybe something we could have a look at:
02:08 < jsaw> http://www.os-cillation.com/article.php?sid=42
02:18 < rxr> oh
02:31 < CIA-9> rene * r4842 /trunk/package/scientific/alliance/ (alliance.conf gcc34.patch):
02:31 < CIA-9> * initial alliance fixup (gcc34.patch and build style) for recent
02:31 < CIA-9> alliance version bump
02:35 < rxr> ok - I'm in bed - n8 all
02:37 -!- _martin [~martin@brln-d9ba26a1.pool.mediaWays.net] has quit [Read error: 110 (Connection timed out)]
02:40 < jsaw> gn8 rxr
02:42 < CIA-9> jsaw * r4843 /trunk/package/office/ooo/ (ooo.conf ooo.desc setup-use-bz2.patch):
02:42 < CIA-9> * another small ooo step: at least start compiling...
02:42 < CIA-9> (breaks in STLport... again...)
02:49 < jsaw> <- gone
03:04 < mnemoc> rehi
03:05 < CIA-9> jsaw * r4844 /trunk/package/base/file/file.desc:
03:05 < CIA-9> * update file (4.11 -> 4.12)
03:05 < CIA-9> * v4.11 is gone on ftp server...
03:06 < mnemoc> rxr: your matrix only show 14 pacakges.... and http:// using * instead of repo is weird :p
03:25 < CIA-9> mnemoc * r4845 /trunk/package/network/ucspi-unix/head_n.patch: * fix deprecated usage of head on ucspi-unix
03:34 < CIA-9> mnemoc * r4846 /trunk/package/network/twoftpd/head_n.patch: * fix head -n patch of twoftpd, one entry was missing.
03:37 < CIA-9> mnemoc * r4847 /trunk/package/develop/mono-develop/mono-develop.desc: * priority of mono-develop increased to be built after gtk-sharp
03:44 < CIA-9> mnemoc * r4848 /trunk/package/security/keepalived/locations.patch: * keepalived's fix of make install adapted to new version
04:00 -!- MubeX [MuBeX@65-23-199-247.prtc.net] has joined #t2
04:00 -!- MubeX is now known as sparc-kly
04:17 -!- _martin [~martin@brln-d9ba2604.pool.mediaWays.net] has joined #t2
04:45 -!- Netsplit over, joins: jsaw
06:32 -!- Netsplit saberhagen.freenode.net <-> irc.freenode.net quits: valentin, jsaw, daja77, praenti, _Ragnar_, nzg, _martin, sparc-kly
06:40 -!- Netsplit over, joins: praenti, valentin
06:40 -!- Netsplit over, joins: jsaw, _martin, sparc-kly, nzg, daja77, _Ragnar_
07:41 -!- mnemoc_ [~amery@200.75.27.23] has joined #t2
07:54 -!- mnemoc [~amery@200.75.27.43] has quit [Read error: 110 (Connection timed out)]
07:54 -!- mnemoc_ is now known as mnemoc
08:37 -!- mtr [~michael@H804c.h.pppool.de] has joined #t2
08:47 < mtr> moin
10:50 -!- rxr_ [~rene@p213.54.211.44.tisdip.tiscali.de] has joined #t2
10:50 -!- Topic for #t2: T2 | the system development environment | http://www.exactcode.de/t2
10:50 -!- Topic set by valentin [] [Sun Nov 7 16:41:27 2004]
10:50 [Users #t2]
10:50 [ _martin ] [ daja77] [ mtr ] [ rxr ] [ valentin]
10:50 [ _Ragnar_] [ jsaw ] [ nzg ] [ rxr_ ]
10:50 [ CIA-9 ] [ mnemoc] [ praenti] [ sparc-kly]
10:50 -!- Irssi: #t2: Total of 13 nicks [0 ops, 0 halfops, 0 voices, 13 normal]
10:50 -!- Channel #t2 created Sun Aug 8 21:15:33 2004
10:50 -!- Irssi: Join to #t2 was synced in 11 secs
11:06 -!- rxr [~rene@p213.54.249.53.tisdip.tiscali.de] has quit [Read error: 110 (Connection timed out)]
12:01 -!- You're now known as rxr
12:02 < rxr> moin
12:28 < _martin> moin
12:28 -!- tschmidt_ [~tschmidt@pD95F868E.dip.t-dialin.net] has joined #t2
12:37 -!- nzg [~tschmidt@pD95F868E.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)]
13:33 -!- sparc-kly [MuBeX@65-23-199-247.prtc.net] has quit [Read error: 104 (Connection reset by peer)]
14:28 < rxr> moin _martin
14:48 < rxr> http://omega.astro.utoronto.ca/kst/kstplot.png
14:49 < rxr> http://omega.astro.utoronto.ca/kst/
16:12 -!- kensai [~kensai@64.237.129.108] has joined #t2
16:57 -!- kensai [~kensai@64.237.129.108] has quit ["Leaving"]
17:40 -!- kensai [~kensai@64.237.129.108] has joined #t2
17:45 -!- kensai [~kensai@64.237.129.108] has quit ["Leaving"]
18:25 -!- kensai [~kensai@64.237.129.108] has joined #t2
18:30 -!- kensai [~kensai@64.237.129.108] has quit ["Leaving"]
18:50 < mnemoc> rxr, valentin: how can i get the keyname of an iterator?
18:51 < rxr> oh
18:51 < rxr> you are on a std::map?
18:52 < mnemoc> yes
18:52 < mnemoc> can i?
18:53 < mnemoc> i wanted to use printf instead of cout because cout << j << ": " << i->first << endl; looks expensive
18:55 < valentin> what map type ?
18:56 < valentin> key type
18:56 < valentin> std::string ?
18:56 < mnemoc> yes
18:56 < valentin> string has a method named c_str () returning a const char*
18:56 < valentin> what do you mean with expensive ?
18:57 < mnemoc> execcution time
18:57 < valentin> hm ? how many couts do you want to do ?
18:58 < valentin> you have a little time loss because of iostream buffering vs stdout file buffering
18:58 < mnemoc> i have 30 seconds to do minimax trees and log actions
18:58 < valentin> but how many couts do you do ?
18:58 < mnemoc> 0.5 seconds
18:59 < mnemoc> one per action
18:59 < valentin> you can try printf and check if that is faster
18:59 < mnemoc> uhm
18:59 < valentin> but in case you want to be realy fast, printf is still to slow - then you have to handcode formated output
19:00 < valentin> the problem with printf is, that the format expression is evaluatet in runtime, and that costs, too
19:00 < mnemoc> ic
19:00 < mnemoc> streams is faster then?
19:01 < valentin> directly write character buffers to STDOUT file should be the fastest, but that is much work when you output numbers and stuff
19:01 < rxr> the buffering issues are fix with gcc-3.4 ...
19:01 < valentin> anyone here wants to do a cout vs. printf benchmark ?
19:02 < rxr> no
19:02 < rxr> but normal cout << usage is a bit costly because of all the single string output - e.g.:
19:02 < valentin> mnemoc: there are some competetive algorithms that produce much output - like world class prime number generators
19:03 < valentin> they use handformated output because the main execution time is lost for the output
19:03 < rxr> cout << "index: " << i << " other debug data" << data << endl;
19:03 < rxr> you have 5 method calls here or more ...
19:04 < mnemoc> vs a printf of the same?
19:04 < mnemoc> vs=versus
19:04 < mnemoc> valentin: ic
19:05 < rxr> well - printf does also doe va_list iteration internally to find out the types ...
19:05 < valentin> yes - and it does in _runtime_
19:05 < rxr> yep
19:06 < rxr> I just wanted to point out a C++ output cost some people tend to overlook ...
19:06 < rxr> I'm not sure which one performs better - maybe they do not differ much anyway - and the C++ one is typesafe, elegant and extendable ...
19:06 < valentin> rxr: which one - the calls ?
19:06 < rxr> yep - the calls ...
19:06 < rxr> if they are not optimized away
19:08 < mnemoc> i need _speed_ too :)
19:08 < mnemoc> and low memory usage, to be able to build big trees
19:09 < mnemoc> i'll use cout
19:09 < rxr> maybe someone wanna benchmark a bit ? ;-)
19:09 < mnemoc> not today :p
19:11 * rxr could kick the libusb developers
19:12 < valentin> i benchmark it, but just a short test
19:12 < valentin> no time for long experiments
19:14 < rxr> cool
19:14 < daja77> rxr: for the missing documentation
19:14 < daja77> ?
19:14 < valentin> hi daja77
19:15 < daja77> hi valentin
19:15 < rxr> daja77: no for messing up the library names in the .8 version
19:15 < rxr> installing libs without .so ...
19:16 < valentin> mnemoc: write the log data into a file instead to std::out
19:16 < valentin> that should be faster than redirecting on the bash
19:16 < mnemoc> valentin: great :D
19:17 < rxr> why should that be faster?
19:17 < daja77> rxr: hm had that with another package too, could be libtool related perhaps?
19:17 < valentin> most of the time in my benchmark is spend in system not user
19:17 < valentin> not in process system
19:18 < valentin> real 0m58.841s
19:18 < valentin> user 0m26.843s
19:18 < valentin> sys 0m2.330s
19:18 < valentin> so half of the real time could have been the > /dev/null
19:18 < valentin> i'll veryfy that
19:18 < mnemoc> i don't need to log in realtime, as far as it produce a log
19:19 < valentin> the thing is, with > you have three buffers
19:20 < rxr> no
19:20 < rxr> a file descriptor is used directly ...
19:21 < valentin> real 0m15.865s
19:21 < valentin> user 0m7.899s
19:21 < valentin> sys 0m0.013s
19:21 < rxr> when you say my-prog < big-movie you get the the fd of that file as stdin - and you can even mmap it
19:21 < valentin> that is more than three times faster !!
19:21 < valentin> maybe smarter buffering, i.e no flush when std::cout is used ??
19:22 < valentin> so mnemoc: use a std::ofstream instead of cout
19:22 < valentin> and i did even write to disk
19:23 < rxr> what did you bench exactly?
19:23 < rxr> did you tried printf?
19:24 < valentin> i tried a version with cout << and a version with output <<
19:24 < valentin> and a version with printf (-> stdout)
19:24 < valentin> the printf is way faster than cout but still slower than output <<
19:25 < valentin> now i have to try fprintf
19:27 < valentin> hm, moment - the formating is the fastest thing - if i do no output at all ie write to an non-openend ostream
19:27 < valentin> most of the time seems to be wasted in that buffering stuff
19:29 < valentin> the times are not very clean because i compile ardour in parallel
19:29 < valentin> in general i would say print
19:29 < valentin> printf is d
19:29 < valentin> damn
19:29 < valentin> printf is almost twice as fast as cout
19:30 < valentin> and fprintf is faster than write to ofstream
19:30 -!- tschmidt_ [~tschmidt@pD95F868E.dip.t-dialin.net] has quit ["Verlassend"]
19:30 < daja77> printf is a fprintf wrapper ...
19:33 < valentin> but it is faster than redirection on the shell for some reason
19:33 < valentin> what is strange in case '>' just hands another descriptor
19:34 < valentin> can anyone veryfy that
19:34 < valentin> ?
19:34 * rxr would love too - but busy
19:35 -!- sparc-kly [~sparc@65-23-199-247.prtc.net] has joined #t2
19:39 < CIA-9> rene * r4849 /trunk/package/extra/libusb/libusb.conf: (log message trimmed)
19:39 < CIA-9> * fixed libusb regression introduced in 1.0.8 to install shared
19:39 < CIA-9> libraries named *.so ;-)
19:39 < CIA-9> it installed:
19:39 < CIA-9> libusb -> libusb-0.1.4.4.0
19:39 < CIA-9> libusb-0.1.4 -> libusb-0.1.4.4.0
19:39 < CIA-9> libusb-0.1.4.4.0
19:41 * mnemoc pukes
19:41 < rxr> mnemoc: ?
19:42 < mnemoc> too many numbers :)
19:43 < rxr> it is not such a big fix so ... - but maybe good I did it for you when there are too many numbers for you ...
19:43 -!- Minto [~chatzilla@82-217-66-7.cable.quicknet.nl] has joined #t2
19:43 < rxr> but the log got truncated by CIA - your mail is longer and includes more numbers ;-)
19:43 < rxr> hi sparc-kly
19:43 < rxr> hi Minto
19:43 < Minto> Hi rene
19:43 < sparc-kly> hi rxr :P
19:44 < rxr> dman - I think I'm overworked ...
19:44 * rxr feels so puffed out - although got out of the bed at 12:00 ... :-()
19:45 < Minto> Could I take some of your shoulder (some easy tasks)?
19:46 * mnemoc needs to understand c++ :\
19:46 < rxr> non - ease on my big todo ;-)
19:46 < rxr> but the big todo contains payed contract work ...
19:46 < Minto> It's a pity
19:46 < rxr> (like macosx device drivers ...)
19:47 < Minto> Ohhh, those are easy ;-)
19:47 < rxr> Minto: in which area of t2 would you like to work on?
19:48 < Minto> Any to get a better understanding of the internals
19:48 < Minto> But maybe I should start out easy
19:48 < Minto> Mayba adding an additional package
19:48 < rxr> Minto: do you already have a runnin t2 system?
19:48 < Minto> for RockLinux I once tried to add MailMan
19:49 < Minto> Not yet, I am still thinking about the proper hardware
19:49 < mnemoc> doens't ios::app came from #include ?
19:49 < rxr> oehm
19:49 < rxr> you have questions ...
19:50 < Minto> not really, recently I discovered the Tyan GX28 bare bone which seems to be a perfect candidate
19:52 < rxr> a bare bone as build server or workstation?
19:52 < Minto> as full throttle build server
19:53 < Minto> currently I have an old system on which I installed rock linux (2.4.22-rock) a few months ago
19:54 < Minto> rxr: remember the chroot problem I had yesterday, I still have after an svn update
19:54 < rxr> hm
19:55 < rxr> I can not imagine off hand what it could be - maybe you could investigate?
19:55 < Minto> I don't know where to start
19:56 < mnemoc> aaaaaaaargh!!!!
19:56 < Minto> Hi mnemoc, are you okay????
19:57 < mnemoc> finally, built
19:57 < mnemoc> Minto: i'm loosing my mind :)
19:58 < Minto> Don't, you only have one (if you're not schizofrenic)
19:58 < rxr> mnemoc: what's up?
20:00 < rxr> valentin: no chance to get you to some beer in some jazz bar whatever today?
20:03 < valentin> rxr: no
20:03 < rxr> I thougth that ...
20:05 < valentin> rxr: maybe tomorow ?
20:06 < valentin> ardour ready ...
20:07 < rxr> maybe - I feel hevily overworked today no idea how it is tomorrow ...
20:07 < valentin> i do not feel like wanting to get smoked today
20:08 < rxr> haha - /me does not want to get smoked on any day ...
20:08 < valentin> we need some law like they have in New York - no smoking in public areas
20:08 < rxr> well - I SMSed someone else - but there the chances are minimal, too ...
20:09 < rxr> well - when we start this way some people wanna have more us laws - and most of the other I would like to avoid
20:09 < valentin> maybe just increasing tobacco taxes for another 500% ?
20:10 < rxr> yep
20:15 < Minto> rxr: On that generic build error I got I have a src.binutils.xxxx leftover with only..
20:15 < Minto> the chroot.sh and debug.sh in it.
20:22 < Minto> after running the debug script I get: /bin/bash not found.
20:23 < Minto> Upon checking the in build/generic-2.1.0xxxxx/bin it indeed is not there.
20:23 < Minto> How should I proceed?
20:37 < rxr> chroot build/generic-*/
20:37 < rxr> does work?
20:37 < rxr> or does it yield the same bin/bash error?
20:45 < Minto> chroot build/generic-*/ gives the same error
20:45 < rxr> was bash built?
20:45 < rxr> e.g. build/generic-*/var/adm/logs/*.bash* present?
20:46 < Minto> Nope, no bash log present
20:48 < rxr> hm
20:49 < Minto> I used the minimal package selection template
20:49 < rxr> ahhhhhhhhhhhh!
20:49 < rxr> no wonder it did not work ...
20:49 < rxr> *g* - you should mention that when you say "generic target" :-)
20:49 < rxr> let me commit a fix
20:49 < Minto> sorry?
20:49 -!- sparc-kly [~sparc@65-23-199-247.prtc.net] has quit ["Leaving"]
20:50 < Minto> What needs to be modified? Maybe I could do it myself.
20:51 < Minto> I guess the bash package needs ot be added, but how.
20:52 < rxr> see the commit just done
20:52 < rxr> svn up
20:52 < CIA-9> rene * r4850 /trunk/misc/pkgsel/minimal.in: * hotfixed minimal package selection template
20:52 < rxr> you can review the changeset:
20:52 < rxr> svn diff -r4849:4850
20:52 < rxr> the templates are in misc/pkg-sel
20:53 < rxr> the templates are in misc/pkgsel even
20:53 < Minto> I think I did the change myself just now. Let's see mine works. Then I'll compare it to yours
20:55 < Minto> Comparing first is quicker. I did the same, however I only added the bash package. I guess the others are needed too.
21:00 < Minto> Uhm, I get lots of red lines...
21:01 < Minto> a lot of logfile shoudl not be present and a lot of missing binary files for *.gem not present.
21:01 < Minto> example lines are:
21:01 < Minto> build/generic-2.1.0-beta-x86-generic-expert/var/adm/logs/1-wget.log should not be present (now in src/invalid-files.lst)!
21:01 < Minto> Binary file for wget-1.9.1.gem not present. Skipped in package database.
21:03 < rxr> this is the post-build CD pre-creation stuff
21:03 < rxr> take a look where the build aborted before
21:03 < rxr> and check if you have a valid package selection template - e.g. no mix of your and my modifitations ...
21:03 < rxr> what does svn status misc/pkgsel output?
21:04 < CIA-9> rene * r4851 /trunk/misc/pkgsel/ (benchmark.in minimal+xfree.in minimal-desktop.in): * more trivial package preselection template fixes
21:08 < Minto> the svn instruction gave nothing, however my change and yours where indeed mixed.
21:08 < Minto> I changed it and things look better now.
21:11 < rxr> .oO
21:11 < rxr> Systemvoraussetzungen
21:11 < rxr> Zur Nutzung von Musicload benötigen Sie mindestens Version 8 des Windows Media Player.
21:11 < rxr>
21:11 < rxr> Die aktuellste Version können Sie jetzt kostenlos downloaden:
21:11 < mnemoc> ?
21:12 < rxr> it is the german error message of http://www.musicload.de that tells me that I need the Windows Media Player in order to utilize this musc download service ...
21:14 < rxr> wow - d
21:14 < rxr> wow - the music is exported via Windows Media Audio-Format (WMA) files ...
21:15 < rxr> how useful ...
21:15 < mnemoc> :\
21:15 < rxr> yeah :-/
21:21 < Minto> rxr: additional package where build okay, however binutils in [2] now gives:
21:21 < Minto> ==[apply_patchfiles:541 (last $?=0)> '[' '!' -e /ROCK/pa ..
21:22 < rxr> ah - patch missing
21:22 < rxr> let me add them to svn
21:23 < rxr> done
21:23 < CIA-9> rene * r4852 /trunk/misc/pkgsel/minimal.in: * added patch to the minimal template
21:24 < Minto> thanks
21:24 < rxr> you could add other missing packages yourself and send the changes obtained by svn diff misc/pkgsel/minimal.in ...
21:24 < Minto> I would if i knew what was wrong.
21:25 < Minto> But now I know this kind of error probably indicate a missing package
21:28 < rxr> yeah ;-)
21:28 < rxr> we moved a lot packges around to get rid of those personally named repositories from ROCK Linux ...
21:29 < rxr> so those hardcoded package preselection templates that relied on the old package distributiion accross the repositories do not work anymore ...
21:29 < rxr> but it it good that they get fixed now that you use them ;-)
21:31 < Minto> I see what I run into next and maybe be able to fix it myself. Hopefully I sink ;-) deeper and deeper into T2.
21:32 < rxr> ;-)
21:33 < Minto> By the way, where you the one to set up MailMan for RockLinux?
21:34 < Minto> I always wonder why RockLinux uses MailMan but has no package for it.
21:34 < rxr> hehe - no clifford set it up himself
21:34 < rxr> after we bugged him for over a year that we want more then one mailing list ...
21:35 < rxr> I do not like mailman much - so far I think of it as a unreviewable moster - that is why we use a custom (read hacked) version of minimalist ...
21:35 < Minto> Why customizing?
21:36 < rxr> added features and standard conformace ...
21:37 < Minto> Hmm, GForge uses MailMan, another reason for you not to use GForge.
21:47 < mnemoc> rxr: do you have a minute?
21:52 -!- kensai [~kensai@64.237.129.108] has joined #t2
21:53 < rxr> mnemoc: yep
21:53 < mnemoc> good :)
21:54 < mnemoc> i have a working: inline void SetLogger( std::ofstream &logger ) { logger << "foo\n;" }
21:54 < mnemoc> how can i make a public log to 'remember' logger?
21:54 < rxr> oh my ...
21:55 < mnemoc> private:
21:55 < mnemoc> 18 std::ofstream *log;
21:55 < mnemoc> log = logger; fails
21:55 < mnemoc> and every convination of & and * :\
21:55 < mnemoc> i'm damn confused
21:56 < mnemoc> everything i try gives ugly compilation errors
21:56 < Minto> have you tried log = &logger ?
21:56 < mnemoc> o_O
21:56 < mnemoc> thanks
21:56 < Minto> your welcome
21:57 < mnemoc> my brain in not working these days
21:57 < Minto> It seems like the visit of Bush has drained a lot of energy over there.
21:58 < mnemoc> absolutly
21:59 < mnemoc> errr...
21:59 < Minto> Uhmm, does my suggestion nog work?
21:59 < mnemoc> how should i do it to use: log << "foo\n"; ? std::ofstream &foo; ?
22:00 < Minto> try (*log) << "foo\n"
22:02 -!- kensai [~kensai@64.237.129.108] has quit ["Leaving"]
22:02 < mnemoc> aaarg... now gcc think it's a double :\
22:03 < mnemoc> Minto: how should i implement on abstract class something will be private?
22:03 < mnemoc> friend: makes it 'double' :\
22:04 < mnemoc> protected :p
22:05 < Minto> ?? I am sure I understand what you want to accomplish
22:05 < mnemoc> yes you did
22:05 < mnemoc> can i do that _and_ use it as log << inside the methods?
22:06 < Minto> I am not sure if ofstream has operator =, but if it has
22:07 < Minto> use std::ofstreamd log; (no pointer)
22:07 < Minto> then you can use log = logger in the setLogger method.
22:08 < Minto> using the log would then be log <<
22:08 < mnemoc> /usr/lib/gcc-lib/i686-unknown-linux-gnu/3.3.3/include/c++/bits/ios_base.h: In
22:08 < mnemoc> member function `std::basic_ios >&
22:08 < mnemoc> std::basic_ios >::operator=(const
22:08 < mnemoc> std::basic_ios >&)'
22:08 < mnemoc> sucks :\
22:08 < mnemoc> thanks you anyway
22:08 < mnemoc> i'll have to use *log
22:09 < Minto> why does it suck??
22:10 < rxr> is there some unix way to send celular phone SMSs via some convienient and cheap interface?
22:11 < rxr> so I do not need to use the 9 button phone interface that is that slow that often ... ?
22:13 < rxr> hm - any idea what those should be in the usb20.zip containing the vanilla 2.0 USB standard?
22:13 < rxr> 0 02-12-04 01:08 __MACOSX/
22:13 < rxr> 0 02-12-04 01:08 __MACOSX/usb_20/
22:13 < rxr> 82 02-12-04 00:55 __MACOSX/usb_20/._OTG_Errata_Update_1.0.pdf
22:18 < rxr> hm - guess I shoudl review this in more depth:
22:18 < rxr> 9.1.1.5 Configured
22:18 < rxr> Before a USB device's function may be used, the device must be configured. From the device's
22:18 < rxr> perspective, configuration involves correctly processing a SetConfiguration() request with a non-zero
22:18 < rxr> configuration value. Configuring a device or changing an alternate setting causes all of the status and
22:18 < rxr> configuration values associated with endpoints in the affected interfaces to be set to their default values.
22:18 < rxr> This includes setting the data toggle of any endpoint using data toggles to the value DATA0.
22:18 < rxr> hm - also interesting to know:
22:18 < rxr> 9.1.1.6 Suspended
22:18 < rxr> In order to conserve power, USB devices automatically enter the Suspended state when the device has
22:18 < rxr> observed no bus traffic for a specified period (refer to Chapter 7). When suspended, the USB device
22:18 < rxr> maintains any internal status, including its address and configuration.
22:18 < rxr> All devices must suspend if bus activity has not been observed for the length of time specified in
22:18 < rxr> Chapter 7. Attached devices must be prepared to suspend at any time they are powered, whether they have
22:25 < CIA-9> rene * r4853 /trunk/package/scientific/alliance/gcc34.patch:
22:25 < CIA-9> * readded a few bits to the alliance/gcc34.patch - builds further but
22:25 < CIA-9> still running
22:29 < rxr> ok - I know what I'm hacking the next hours ...
22:45 < Minto> bedtime for me, gn8 all
22:45 -!- Minto [~chatzilla@82-217-66-7.cable.quicknet.nl] has quit ["ChatZilla 0.9.61 [Mozilla rv:1.7.3/20040910]"]
22:45 < mnemoc> Minto,rxr: how can i format the integers i dump to cout? (%3d of printf)
22:47 < rxr> what does %3 did ? field width of 3 chars?
22:47 < rxr> << std::setw (3)
22:47 < rxr> if it is the width ...
22:48 < rxr> and
22:48 < rxr> << std::left
22:48 < rxr> << std::right
22:48 < rxr> for the alignment
22:48 < rxr> but default should be ok for you ...
22:48 < mnemoc> i want 'air' between numbers :)
22:49 < rxr> like fixed width in table form?
22:49 < rxr> then the above is for you
22:49 < rxr> btw. is this some university project? or what is it? ;-)
22:50 < mnemoc> boards/basic_board.cc:38: error: `setw' undeclared in namespace `std'
22:50 < mnemoc> university
22:50 < mnemoc> for AI
22:50 < rxr> #include
22:50 < rxr> ic
22:51 < rxr> I love those:
22:52 < rxr> "Default state: Device behavior when this request is received while the device is in the Default state
22:52 < rxr> is not specified.
22:52 < rxr> "
23:17 < mnemoc> rxr: how do i flush cin ?
23:19 < mnemoc> rxr: ?
23:23 < rxr> you need a C++ book
23:24 < rxr> cin.flush ()
23:24 < rxr> ?
23:24 < rxr> tell me if it worked ;-)
23:24 < rxr> did the above setw help you?
23:26 < mnemoc> nope
23:26 < mnemoc> including iomanip killled my class :p
23:27 < mnemoc> players/human_player.cc:18: error: `flush' undeclared (first use this function)
23:27 < mnemoc> calling it as std::cin.flush();
23:28 < mnemoc> rxr: ping :p
23:30 < rxr> pong
23:31 < rxr> well std::cin.flush(); of course
23:31 < rxr> including iomanip should not kill anything ...
23:31 < rxr> both work here.
23:31 < rxr> maybe inspirate you with this
23:31 < mnemoc> std::cin.flush(); doesn't work :(
23:32 < rxr> http://svn.exactcode.de/gsmp/trunk/pcm/tests/Bench.cc
23:32 < rxr> what is the error?
23:32 < mnemoc> players/human_player.cc: In member function `Move HumanPlayer::AskMove()':
23:32 < mnemoc> players/human_player.cc:18: error: `flush' undeclared (first use this function)
23:32 < mnemoc> players/human_player.cc:18: error: (Each undeclared identifier is reported only
23:33 < rxr> show me the relevant lines ;-)
23:33 < mnemoc> Move m;
23:33 < mnemoc> 18 std::cin.flush();
23:33 < mnemoc> 19 std::cout << "Fila a jugar? [1-9] ";
23:33 < mnemoc> 20 std::cin >> m.X;
23:34 < mnemoc> C++ hates me
23:34 < mnemoc> i should have coded in plain C :\
23:34 < rxr> hm - seems only ostream has flufh
23:34 < rxr> did you choose C++ yourself?
23:34 < mnemoc> yes
23:35 < rxr> let me open a book
23:35 < mnemoc> you told "C++ is the most wonderfull language in the world" and i believed you :p
23:35 < rxr> well - although I did not say exactly that - C++ is great ...
23:41 < rxr> for some yet unknown reason there is only flush for ostream's
23:42 < mnemoc> :\
23:43 < mnemoc> i want to read ints, if i add another thing it gets crazy :\
23:43 < mnemoc> that's why i want to flush it
23:43 < mnemoc> as i would do on C
23:44 < rxr> getline reads one line ...
23:45 < mnemoc> atoi?
23:45 < mnemoc> C++ is not as magic as i would expected :\
23:46 < rxr> noone said C++ is magic ...
23:47 < rxr> ok - why do you got to atoi?
23:47 < rxr> although I do not know some flush method for istreams off-hand (seme
23:47 < rxr> seems I never needed that)
23:47 < rxr> you could still do
23:47 < rxr> int i;M
23:48 < rxr> hm - wait
23:48 < mnemoc> i is int :)
23:49 < mnemoc> how would you safely as for three _numbers_
23:49 < mnemoc> ?
23:49 < rxr> what do you mean=
23:50 < rxr> ? even
23:50 < rxr> "as for three" what do you mean?
23:50 < mnemoc> ask*
23:51 < mnemoc> make it simple
23:51 < mnemoc> ask for one number
23:51 < mnemoc> until it gets a number
23:52 < rxr> mom
23:52 < mnemoc> i=-1; while(i<0) { cin >> i; }
23:52 < mnemoc> if you press enter it goes crazy :)
23:52 < rxr> damn - I do too less C++ these days ...
23:53 < mnemoc> what do you use? visual basic? :p
23:54 < rxr> no - too much bash hacking for t2 - and too much open source C crap ...
23:54 < rxr> ok it is that easy:
23:54 < rxr> cout << "your msg" << endl;
23:54 -!- kensai [~kensai@64.237.129.108] has joined #t2
23:54 < rxr> while (! (cin >> i) )
23:54 < rxr> {
23:54 < rxr> cout << "better input a number" << endl;
23:54 < rxr> cin.getline();
23:54 < rxr> }
23:55 < rxr> I'm not sure about the getline#
23:55 < mnemoc> aha
23:55 < rxr> let me take a look
23:55 < mnemoc> i think getline does the trick
23:55 < rxr> I really do not know why they left out flush for istreams - but I have not missed it until tonight ...
23:56 < rxr> ah - no it does not have a getline like that ...
23:57 < rxr> if you want s.th. complex but correct:
23:57 < rxr> cin.ignore (numeric_limits::max(), '\n');
23:57 < rxr> ;-)
23:57 < mnemoc> o_O
23:58 < mnemoc> ok =)
23:58 < mnemoc> should i run that before cin >> i ?
23:58 < mnemoc> once or everytime?
23:58 < rxr> you should run it in the loop, too - to remove incorrect input of course ...
23:58 < rxr> when you have a dummy object you could do:
23:59 < rxr> getline (cin, stdstring);
23:59 < mnemoc> your complex thing didn't complied :p
23:59 -!- kensai [~kensai@64.237.129.108] has quit ["Leaving"]
23:59 < rxr> you need to include some limits header ...
--- Log closed Fri Nov 26 00:00:42 2004
00:16 < rxr> enjoy:
00:16 < rxr> http://www.exactcode.de/t2/packages/
00:16 < rxr> of coursed linked in the frontpage menu ...
00:16 < rxr> and some remaining stuff needs to be done - I'm working on it
00:28 < rxr> ok - basica fixed, example of newly fixed file:
00:28 < rxr> http://www.exactcode.de/t2/packages/aboot.html
00:33 -!- kensai [~kensai@64.237.129.108] has joined #t2
00:46 -!- kensai [~kensai@64.237.129.108] has quit ["Leaving"]
00:49 -!- kensai [~kensai@64.237.129.108] has joined #t2
00:56 -!- kensai [~kensai@64.237.129.108] has quit ["Debian Rulez"]
00:57 < rxr> I just browse the exactcode statistic - and I'm impressed how much traffic the box pushes into the net ...
01:17 < CIA-9> rene * r4837 /trunk/package/x11/rxvt-unicode/rxvt-unicode.desc: * updated rxvt-unicode (4.0 -> 4.1) - orignial file vanished
01:25 < CIA-9> rene * r4838 /trunk/package/base/hddtemp/hddtemp.desc:
01:25 < CIA-9> * updated hddtemp (0.3-beta11 -> 0.3-beta12) and updated the URL tag
01:25 < CIA-9> - previous site and file vanished
01:26 < CIA-9> rene * r4839 /trunk/package/x11/fox/fox.desc: * updated fox (1.3.12 -> 1.3.19) - original file vanished
01:39 < CIA-9> rene * r4840 /trunk/package/scientific/alliance/alliance.desc:
01:39 < CIA-9> * updated alliance (5.0-20040909 -> 5.0-20040928) - orignial file
01:39 < CIA-9> vanished
01:41 < rxr> 130(!) patches in submatser - they will have fun....
01:44 < rxr> no pcmcia on sparc... afaics
01:44 < jsaw> re
01:44 < rxr> ??? - of course sparc book had PCMCIA ...
01:44 < jsaw> hi rxr
01:44 < rxr> the rock submaster is a big trash deposit ...
01:44 < rxr> hi jsaw
01:45 < rxr> we have a preleminary package matrix ;-)
01:45 < jsaw> I've seen your comment but haven't had a look a it... let's see
01:46 < CIA-9> rene * r4841 /trunk/package/gnome2/gal/gal.desc:
01:46 < CIA-9> * fixed the gal URL to use the usual GNOME mirror - not a non-existing
01:46 < CIA-9> one so our master mirror can catch it ;-)
01:47 < jsaw> oh!
01:47 < jsaw> cool!
01:48 < rxr> hm - damn - on the mirror the nautilus file download always stops after 219.220 bytes ...
01:48 < rxr> hm - no - randomlly ...
01:50 < rxr> jsaw: how are you?
01:51 < rxr> hm - the gnome server is strange
01:57 < rxr> well - some random gnome mirror works as all the other download for T2 do ...
01:59 < jsaw> rxr: well, not so good today. Had a bit problems with my heart, but obviously nothing wrong on the organ level. "Herzbeklemmung" - most probably stress related.
02:00 < jsaw> going to make a "Belastungs EKG" tomorrow, but I guess, nothing will be found -> have to get a few days of holidays
02:00 < rxr> ic - then better relax a bit ... !
02:00 < rxr> XFce 4.2 - the Future is Now!
02:01 < rxr> ^- update anyone but jsaw ;-)
02:01 < jsaw> I do. Programming is relaxing ... :)
02:01 < jsaw> is xfce 4.2 ready?
02:01 < jsaw> or a rc?
02:02 < rxr> seems - ready - but just the osnews release I posted ...
02:02 < rxr> I saw a -rc before ... - so this must be final ...
02:03 < jsaw> the last entry on xfce.org is
02:03 < jsaw> 2004/11/14 - Xfce 4.2 Release Candidate 1 is out
02:04 < rxr> ouhm - I know why I do not like news sites ... they keep claiming releases there there are none more often with every day
02:04 < rxr> gcc was also announced when it hit the first master server and the gcc people intended to release it two days later ...
02:05 < rxr> and with mozilla they do the same ... - freak out when the first binary hit the server - but the source is still missing :-(
02:05 < jsaw> ...
02:08 < jsaw> btw xfce, there's a graphical source installer, maybe something we could have a look at:
02:08 < jsaw> http://www.os-cillation.com/article.php?sid=42
02:18 < rxr> oh
02:31 < CIA-9> rene * r4842 /trunk/package/scientific/alliance/ (alliance.conf gcc34.patch):
02:31 < CIA-9> * initial alliance fixup (gcc34.patch and build style) for recent
02:31 < CIA-9> alliance version bump
02:35 < rxr> ok - I'm in bed - n8 all
02:37 -!- _martin [~martin@brln-d9ba26a1.pool.mediaWays.net] has quit [Read error: 110 (Connection timed out)]
02:40 < jsaw> gn8 rxr
02:42 < CIA-9> jsaw * r4843 /trunk/package/office/ooo/ (ooo.conf ooo.desc setup-use-bz2.patch):
02:42 < CIA-9> * another small ooo step: at least start compiling...
02:42 < CIA-9> (breaks in STLport... again...)
02:49 < jsaw> <- gone
03:04 < mnemoc> rehi
03:05 < CIA-9> jsaw * r4844 /trunk/package/base/file/file.desc:
03:05 < CIA-9> * update file (4.11 -> 4.12)
03:05 < CIA-9> * v4.11 is gone on ftp server...
03:06 < mnemoc> rxr: your matrix only show 14 pacakges.... and http:// using * instead of repo is weird :p
03:25 < CIA-9> mnemoc * r4845 /trunk/package/network/ucspi-unix/head_n.patch: * fix deprecated usage of head on ucspi-unix
03:34 < CIA-9> mnemoc * r4846 /trunk/package/network/twoftpd/head_n.patch: * fix head -n patch of twoftpd, one entry was missing.
03:37 < CIA-9> mnemoc * r4847 /trunk/package/develop/mono-develop/mono-develop.desc: * priority of mono-develop increased to be built after gtk-sharp
03:44 < CIA-9> mnemoc * r4848 /trunk/package/security/keepalived/locations.patch: * keepalived's fix of make install adapted to new version
04:00 -!- MubeX [MuBeX@65-23-199-247.prtc.net] has joined #t2
04:00 -!- MubeX is now known as sparc-kly
04:17 -!- _martin [~martin@brln-d9ba2604.pool.mediaWays.net] has joined #t2
04:45 -!- Netsplit over, joins: jsaw
06:32 -!- Netsplit saberhagen.freenode.net <-> irc.freenode.net quits: valentin, jsaw, daja77, praenti, _Ragnar_, nzg, _martin, sparc-kly
06:40 -!- Netsplit over, joins: praenti, valentin
06:40 -!- Netsplit over, joins: jsaw, _martin, sparc-kly, nzg, daja77, _Ragnar_
07:41 -!- mnemoc_ [~amery@200.75.27.23] has joined #t2
07:54 -!- mnemoc [~amery@200.75.27.43] has quit [Read error: 110 (Connection timed out)]
07:54 -!- mnemoc_ is now known as mnemoc
08:37 -!- mtr [~michael@H804c.h.pppool.de] has joined #t2
08:47 < mtr> moin
10:50 -!- rxr_ [~rene@p213.54.211.44.tisdip.tiscali.de] has joined #t2
10:50 -!- Topic for #t2: T2 | the system development environment | http://www.exactcode.de/t2
10:50 -!- Topic set by valentin [] [Sun Nov 7 16:41:27 2004]
10:50 [Users #t2]
10:50 [ _martin ] [ daja77] [ mtr ] [ rxr ] [ valentin]
10:50 [ _Ragnar_] [ jsaw ] [ nzg ] [ rxr_ ]
10:50 [ CIA-9 ] [ mnemoc] [ praenti] [ sparc-kly]
10:50 -!- Irssi: #t2: Total of 13 nicks [0 ops, 0 halfops, 0 voices, 13 normal]
10:50 -!- Channel #t2 created Sun Aug 8 21:15:33 2004
10:50 -!- Irssi: Join to #t2 was synced in 11 secs
11:06 -!- rxr [~rene@p213.54.249.53.tisdip.tiscali.de] has quit [Read error: 110 (Connection timed out)]
12:01 -!- You're now known as rxr
12:02 < rxr> moin
12:28 < _martin> moin
12:28 -!- tschmidt_ [~tschmidt@pD95F868E.dip.t-dialin.net] has joined #t2
12:37 -!- nzg [~tschmidt@pD95F868E.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)]
13:33 -!- sparc-kly [MuBeX@65-23-199-247.prtc.net] has quit [Read error: 104 (Connection reset by peer)]
14:28 < rxr> moin _martin
14:48 < rxr> http://omega.astro.utoronto.ca/kst/kstplot.png
14:49 < rxr> http://omega.astro.utoronto.ca/kst/
16:12 -!- kensai [~kensai@64.237.129.108] has joined #t2
16:57 -!- kensai [~kensai@64.237.129.108] has quit ["Leaving"]
17:40 -!- kensai [~kensai@64.237.129.108] has joined #t2
17:45 -!- kensai [~kensai@64.237.129.108] has quit ["Leaving"]
18:25 -!- kensai [~kensai@64.237.129.108] has joined #t2
18:30 -!- kensai [~kensai@64.237.129.108] has quit ["Leaving"]
18:50 < mnemoc> rxr, valentin: how can i get the keyname of an iterator?
18:51 < rxr> oh
18:51 < rxr> you are on a std::map?
18:52 < mnemoc> yes
18:52 < mnemoc> can i?
18:53 < mnemoc> i wanted to use printf instead of cout because cout << j << ": " << i->first << endl; looks expensive
18:55 < valentin> what map type ?
18:56 < valentin> key type
18:56 < valentin> std::string ?
18:56 < mnemoc> yes
18:56 < valentin> string has a method named c_str () returning a const char*
18:56 < valentin> what do you mean with expensive ?
18:57 < mnemoc> execcution time
18:57 < valentin> hm ? how many couts do you want to do ?
18:58 < valentin> you have a little time loss because of iostream buffering vs stdout file buffering
18:58 < mnemoc> i have 30 seconds to do minimax trees and log actions
18:58 < valentin> but how many couts do you do ?
18:58 < mnemoc> 0.5 seconds
18:59 < mnemoc> one per action
18:59 < valentin> you can try printf and check if that is faster
18:59 < mnemoc> uhm
18:59 < valentin> but in case you want to be realy fast, printf is still to slow - then you have to handcode formated output
19:00 < valentin> the problem with printf is, that the format expression is evaluatet in runtime, and that costs, too
19:00 < mnemoc> ic
19:00 < mnemoc> streams is faster then?
19:01 < valentin> directly write character buffers to STDOUT file should be the fastest, but that is much work when you output numbers and stuff
19:01 < rxr> the buffering issues are fix with gcc-3.4 ...
19:01 < valentin> anyone here wants to do a cout vs. printf benchmark ?
19:02 < rxr> no
19:02 < rxr> but normal cout << usage is a bit costly because of all the single string output - e.g.:
19:02 < valentin> mnemoc: there are some competetive algorithms that produce much output - like world class prime number generators
19:03 < valentin> they use handformated output because the main execution time is lost for the output
19:03 < rxr> cout << "index: " << i << " other debug data" << data << endl;
19:03 < rxr> you have 5 method calls here or more ...
19:04 < mnemoc> vs a printf of the same?
19:04 < mnemoc> vs=versus
19:04 < mnemoc> valentin: ic
19:05 < rxr> well - printf does also doe va_list iteration internally to find out the types ...
19:05 < valentin> yes - and it does in _runtime_
19:05 < rxr> yep
19:06 < rxr> I just wanted to point out a C++ output cost some people tend to overlook ...
19:06 < rxr> I'm not sure which one performs better - maybe they do not differ much anyway - and the C++ one is typesafe, elegant and extendable ...
19:06 < valentin> rxr: which one - the calls ?
19:06 < rxr> yep - the calls ...
19:06 < rxr> if they are not optimized away
19:08 < mnemoc> i need _speed_ too :)
19:08 < mnemoc> and low memory usage, to be able to build big trees
19:09 < mnemoc> i'll use cout
19:09 < rxr> maybe someone wanna benchmark a bit ? ;-)
19:09 < mnemoc> not today :p
19:11 * rxr could kick the libusb developers
19:12 < valentin> i benchmark it, but just a short test
19:12 < valentin> no time for long experiments
19:14 < rxr> cool
19:14 < daja77> rxr: for the missing documentation
19:14 < daja77> ?
19:14 < valentin> hi daja77
19:15 < daja77> hi valentin
19:15 < rxr> daja77: no for messing up the library names in the .8 version
19:15 < rxr> installing libs without .so ...
19:16 < valentin> mnemoc: write the log data into a file instead to std::out
19:16 < valentin> that should be faster than redirecting on the bash
19:16 < mnemoc> valentin: great :D
19:17 < rxr> why should that be faster?
19:17 < daja77> rxr: hm had that with another package too, could be libtool related perhaps?
19:17 < valentin> most of the time in my benchmark is spend in system not user
19:17 < valentin> not in process system
19:18 < valentin> real 0m58.841s
19:18 < valentin> user 0m26.843s
19:18 < valentin> sys 0m2.330s
19:18 < valentin> so half of the real time could have been the > /dev/null
19:18 < valentin> i'll veryfy that
19:18 < mnemoc> i don't need to log in realtime, as far as it produce a log
19:19 < valentin> the thing is, with > you have three buffers
19:20 < rxr> no
19:20 < rxr> a file descriptor is used directly ...
19:21 < valentin> real 0m15.865s
19:21 < valentin> user 0m7.899s
19:21 < valentin> sys 0m0.013s
19:21 < rxr> when you say my-prog < big-movie you get the the fd of that file as stdin - and you can even mmap it
19:21 < valentin> that is more than three times faster !!
19:21 < valentin> maybe smarter buffering, i.e no flush when std::cout is used ??
19:22 < valentin> so mnemoc: use a std::ofstream instead of cout
19:22 < valentin> and i did even write to disk
19:23 < rxr> what did you bench exactly?
19:23 < rxr> did you tried printf?
19:24 < valentin> i tried a version with cout << and a version with output <<
19:24 < valentin> and a version with printf (-> stdout)
19:24 < valentin> the printf is way faster than cout but still slower than output <<
19:25 < valentin> now i have to try fprintf
19:27 < valentin> hm, moment - the formating is the fastest thing - if i do no output at all ie write to an non-openend ostream
19:27 < valentin> most of the time seems to be wasted in that buffering stuff
19:29 < valentin> the times are not very clean because i compile ardour in parallel
19:29 < valentin> in general i would say print
19:29 < valentin> printf is d
19:29 < valentin> damn
19:29 < valentin> printf is almost twice as fast as cout
19:30 < valentin> and fprintf is faster than write to ofstream
19:30 -!- tschmidt_ [~tschmidt@pD95F868E.dip.t-dialin.net] has quit ["Verlassend"]
19:30 < daja77> printf is a fprintf wrapper ...
19:33 < valentin> but it is faster than redirection on the shell for some reason
19:33 < valentin> what is strange in case '>' just hands another descriptor
19:34 < valentin> can anyone veryfy that
19:34 < valentin> ?
19:34 * rxr would love too - but busy
19:35 -!- sparc-kly [~sparc@65-23-199-247.prtc.net] has joined #t2
19:39 < CIA-9> rene * r4849 /trunk/package/extra/libusb/libusb.conf: (log message trimmed)
19:39 < CIA-9> * fixed libusb regression introduced in 1.0.8 to install shared
19:39 < CIA-9> libraries named *.so ;-)
19:39 < CIA-9> it installed:
19:39 < CIA-9> libusb -> libusb-0.1.4.4.0
19:39 < CIA-9> libusb-0.1.4 -> libusb-0.1.4.4.0
19:39 < CIA-9> libusb-0.1.4.4.0
19:41 * mnemoc pukes
19:41 < rxr> mnemoc: ?
19:42 < mnemoc> too many numbers :)
19:43 < rxr> it is not such a big fix so ... - but maybe good I did it for you when there are too many numbers for you ...
19:43 -!- Minto [~chatzilla@82-217-66-7.cable.quicknet.nl] has joined #t2
19:43 < rxr> but the log got truncated by CIA - your mail is longer and includes more numbers ;-)
19:43 < rxr> hi sparc-kly
19:43 < rxr> hi Minto
19:43 < Minto> Hi rene
19:43 < sparc-kly> hi rxr :P
19:44 < rxr> dman - I think I'm overworked ...
19:44 * rxr feels so puffed out - although got out of the bed at 12:00 ... :-()
19:45 < Minto> Could I take some of your shoulder (some easy tasks)?
19:46 * mnemoc needs to understand c++ :\
19:46 < rxr> non - ease on my big todo ;-)
19:46 < rxr> but the big todo contains payed contract work ...
19:46 < Minto> It's a pity
19:46 < rxr> (like macosx device drivers ...)
19:47 < Minto> Ohhh, those are easy ;-)
19:47 < rxr> Minto: in which area of t2 would you like to work on?
19:48 < Minto> Any to get a better understanding of the internals
19:48 < Minto> But maybe I should start out easy
19:48 < Minto> Mayba adding an additional package
19:48 < rxr> Minto: do you already have a runnin t2 system?
19:48 < Minto> for RockLinux I once tried to add MailMan
19:49 < Minto> Not yet, I am still thinking about the proper hardware
19:49 < mnemoc> doens't ios::app came from #include
19:49 < rxr> oehm
19:49 < rxr> you have questions ...
19:50 < Minto> not really, recently I discovered the Tyan GX28 bare bone which seems to be a perfect candidate
19:52 < rxr> a bare bone as build server or workstation?
19:52 < Minto> as full throttle build server
19:53 < Minto> currently I have an old system on which I installed rock linux (2.4.22-rock) a few months ago
19:54 < Minto> rxr: remember the chroot problem I had yesterday, I still have after an svn update
19:54 < rxr> hm
19:55 < rxr> I can not imagine off hand what it could be - maybe you could investigate?
19:55 < Minto> I don't know where to start
19:56 < mnemoc> aaaaaaaargh!!!!
19:56 < Minto> Hi mnemoc, are you okay????
19:57 < mnemoc> finally, built
19:57 < mnemoc> Minto: i'm loosing my mind :)
19:58 < Minto> Don't, you only have one (if you're not schizofrenic)
19:58 < rxr> mnemoc: what's up?
20:00 < rxr> valentin: no chance to get you to some beer in some jazz bar whatever today?
20:03 < valentin> rxr: no
20:03 < rxr> I thougth that ...
20:05 < valentin> rxr: maybe tomorow ?
20:06 < valentin> ardour ready ...
20:07 < rxr> maybe - I feel hevily overworked today no idea how it is tomorrow ...
20:07 < valentin> i do not feel like wanting to get smoked today
20:08 < rxr> haha - /me does not want to get smoked on any day ...
20:08 < valentin> we need some law like they have in New York - no smoking in public areas
20:08 < rxr> well - I SMSed someone else - but there the chances are minimal, too ...
20:09 < rxr> well - when we start this way some people wanna have more us laws - and most of the other I would like to avoid
20:09 < valentin> maybe just increasing tobacco taxes for another 500% ?
20:10 < rxr> yep
20:15 < Minto> rxr: On that generic build error I got I have a src.binutils.xxxx leftover with only..
20:15 < Minto> the chroot.sh and debug.sh in it.
20:22 < Minto> after running the debug script I get: /bin/bash not found.
20:23 < Minto> Upon checking the in build/generic-2.1.0xxxxx/bin it indeed is not there.
20:23 < Minto> How should I proceed?
20:37 < rxr> chroot build/generic-*/
20:37 < rxr> does work?
20:37 < rxr> or does it yield the same bin/bash error?
20:45 < Minto> chroot build/generic-*/ gives the same error
20:45 < rxr> was bash built?
20:45 < rxr> e.g. build/generic-*/var/adm/logs/*.bash* present?
20:46 < Minto> Nope, no bash log present
20:48 < rxr> hm
20:49 < Minto> I used the minimal package selection template
20:49 < rxr> ahhhhhhhhhhhh!
20:49 < rxr> no wonder it did not work ...
20:49 < rxr> *g* - you should mention that when you say "generic target" :-)
20:49 < rxr> let me commit a fix
20:49 < Minto> sorry?
20:49 -!- sparc-kly [~sparc@65-23-199-247.prtc.net] has quit ["Leaving"]
20:50 < Minto> What needs to be modified? Maybe I could do it myself.
20:51 < Minto> I guess the bash package needs ot be added, but how.
20:52 < rxr> see the commit just done
20:52 < rxr> svn up
20:52 < CIA-9> rene * r4850 /trunk/misc/pkgsel/minimal.in: * hotfixed minimal package selection template
20:52 < rxr> you can review the changeset:
20:52 < rxr> svn diff -r4849:4850
20:52 < rxr> the templates are in misc/pkg-sel
20:53 < rxr> the templates are in misc/pkgsel even
20:53 < Minto> I think I did the change myself just now. Let's see mine works. Then I'll compare it to yours
20:55 < Minto> Comparing first is quicker. I did the same, however I only added the bash package. I guess the others are needed too.
21:00 < Minto> Uhm, I get lots of red lines...
21:01 < Minto> a lot of logfile shoudl not be present and a lot of missing binary files for *.gem not present.
21:01 < Minto> example lines are:
21:01 < Minto> build/generic-2.1.0-beta-x86-generic-expert/var/adm/logs/1-wget.log should not be present (now in src/invalid-files.lst)!
21:01 < Minto> Binary file for wget-1.9.1.gem not present. Skipped in package database.
21:03 < rxr> this is the post-build CD pre-creation stuff
21:03 < rxr> take a look where the build aborted before
21:03 < rxr> and check if you have a valid package selection template - e.g. no mix of your and my modifitations ...
21:03 < rxr> what does svn status misc/pkgsel output?
21:04 < CIA-9> rene * r4851 /trunk/misc/pkgsel/ (benchmark.in minimal+xfree.in minimal-desktop.in): * more trivial package preselection template fixes
21:08 < Minto> the svn instruction gave nothing, however my change and yours where indeed mixed.
21:08 < Minto> I changed it and things look better now.
21:11 < rxr> .oO
21:11 < rxr> Systemvoraussetzungen
21:11 < rxr> Zur Nutzung von Musicload benötigen Sie mindestens Version 8 des Windows Media Player.
21:11 < rxr>
21:11 < rxr> Die aktuellste Version können Sie jetzt kostenlos downloaden:
21:11 < mnemoc> ?
21:12 < rxr> it is the german error message of http://www.musicload.de that tells me that I need the Windows Media Player in order to utilize this musc download service ...
21:14 < rxr> wow - d
21:14 < rxr> wow - the music is exported via Windows Media Audio-Format (WMA) files ...
21:15 < rxr> how useful ...
21:15 < mnemoc> :\
21:15 < rxr> yeah :-/
21:21 < Minto> rxr: additional package where build okay, however binutils in [2] now gives:
21:21 < Minto> ==[apply_patchfiles:541 (last $?=0)> '[' '!' -e /ROCK/pa ..
21:22 < rxr> ah - patch missing
21:22 < rxr> let me add them to svn
21:23 < rxr> done
21:23 < CIA-9> rene * r4852 /trunk/misc/pkgsel/minimal.in: * added patch to the minimal template
21:24 < Minto> thanks
21:24 < rxr> you could add other missing packages yourself and send the changes obtained by svn diff misc/pkgsel/minimal.in ...
21:24 < Minto> I would if i knew what was wrong.
21:25 < Minto> But now I know this kind of error probably indicate a missing package
21:28 < rxr> yeah ;-)
21:28 < rxr> we moved a lot packges around to get rid of those personally named repositories from ROCK Linux ...
21:29 < rxr> so those hardcoded package preselection templates that relied on the old package distributiion accross the repositories do not work anymore ...
21:29 < rxr> but it it good that they get fixed now that you use them ;-)
21:31 < Minto> I see what I run into next and maybe be able to fix it myself. Hopefully I sink ;-) deeper and deeper into T2.
21:32 < rxr> ;-)
21:33 < Minto> By the way, where you the one to set up MailMan for RockLinux?
21:34 < Minto> I always wonder why RockLinux uses MailMan but has no package for it.
21:34 < rxr> hehe - no clifford set it up himself
21:34 < rxr> after we bugged him for over a year that we want more then one mailing list ...
21:35 < rxr> I do not like mailman much - so far I think of it as a unreviewable moster - that is why we use a custom (read hacked) version of minimalist ...
21:35 < Minto> Why customizing?
21:36 < rxr> added features and standard conformace ...
21:37 < Minto> Hmm, GForge uses MailMan, another reason for you not to use GForge.
21:47 < mnemoc> rxr: do you have a minute?
21:52 -!- kensai [~kensai@64.237.129.108] has joined #t2
21:53 < rxr> mnemoc: yep
21:53 < mnemoc> good :)
21:54 < mnemoc> i have a working: inline void SetLogger( std::ofstream &logger ) { logger << "foo\n;" }
21:54 < mnemoc> how can i make a public log to 'remember' logger?
21:54 < rxr> oh my ...
21:55 < mnemoc> private:
21:55 < mnemoc> 18 std::ofstream *log;
21:55 < mnemoc> log = logger; fails
21:55 < mnemoc> and every convination of & and * :\
21:55 < mnemoc> i'm damn confused
21:56 < mnemoc> everything i try gives ugly compilation errors
21:56 < Minto> have you tried log = &logger ?
21:56 < mnemoc> o_O
21:56 < mnemoc> thanks
21:56 < Minto> your welcome
21:57 < mnemoc> my brain in not working these days
21:57 < Minto> It seems like the visit of Bush has drained a lot of energy over there.
21:58 < mnemoc> absolutly
21:59 < mnemoc> errr...
21:59 < Minto> Uhmm, does my suggestion nog work?
21:59 < mnemoc> how should i do it to use: log << "foo\n"; ? std::ofstream &foo; ?
22:00 < Minto> try (*log) << "foo\n"
22:02 -!- kensai [~kensai@64.237.129.108] has quit ["Leaving"]
22:02 < mnemoc> aaarg... now gcc think it's a double :\
22:03 < mnemoc> Minto: how should i implement on abstract class something will be private?
22:03 < mnemoc> friend: makes it 'double' :\
22:04 < mnemoc> protected :p
22:05 < Minto> ?? I am sure I understand what you want to accomplish
22:05 < mnemoc> yes you did
22:05 < mnemoc> can i do that _and_ use it as log << inside the methods?
22:06 < Minto> I am not sure if ofstream has operator =, but if it has
22:07 < Minto> use std::ofstreamd log; (no pointer)
22:07 < Minto> then you can use log = logger in the setLogger method.
22:08 < Minto> using the log would then be log <<
22:08 < mnemoc> /usr/lib/gcc-lib/i686-unknown-linux-gnu/3.3.3/include/c++/bits/ios_base.h: In
22:08 < mnemoc> member function `std::basic_ios
22:08 < mnemoc> std::basic_ios
22:08 < mnemoc> std::basic_ios
22:08 < mnemoc> sucks :\
22:08 < mnemoc> thanks you anyway
22:08 < mnemoc> i'll have to use *log
22:09 < Minto> why does it suck??
22:10 < rxr> is there some unix way to send celular phone SMSs via some convienient and cheap interface?
22:11 < rxr> so I do not need to use the 9 button phone interface that is that slow that often ... ?
22:13 < rxr> hm - any idea what those should be in the usb20.zip containing the vanilla 2.0 USB standard?
22:13 < rxr> 0 02-12-04 01:08 __MACOSX/
22:13 < rxr> 0 02-12-04 01:08 __MACOSX/usb_20/
22:13 < rxr> 82 02-12-04 00:55 __MACOSX/usb_20/._OTG_Errata_Update_1.0.pdf
22:18 < rxr> hm - guess I shoudl review this in more depth:
22:18 < rxr> 9.1.1.5 Configured
22:18 < rxr> Before a USB device's function may be used, the device must be configured. From the device's
22:18 < rxr> perspective, configuration involves correctly processing a SetConfiguration() request with a non-zero
22:18 < rxr> configuration value. Configuring a device or changing an alternate setting causes all of the status and
22:18 < rxr> configuration values associated with endpoints in the affected interfaces to be set to their default values.
22:18 < rxr> This includes setting the data toggle of any endpoint using data toggles to the value DATA0.
22:18 < rxr> hm - also interesting to know:
22:18 < rxr> 9.1.1.6 Suspended
22:18 < rxr> In order to conserve power, USB devices automatically enter the Suspended state when the device has
22:18 < rxr> observed no bus traffic for a specified period (refer to Chapter 7). When suspended, the USB device
22:18 < rxr> maintains any internal status, including its address and configuration.
22:18 < rxr> All devices must suspend if bus activity has not been observed for the length of time specified in
22:18 < rxr> Chapter 7. Attached devices must be prepared to suspend at any time they are powered, whether they have
22:25 < CIA-9> rene * r4853 /trunk/package/scientific/alliance/gcc34.patch:
22:25 < CIA-9> * readded a few bits to the alliance/gcc34.patch - builds further but
22:25 < CIA-9> still running
22:29 < rxr> ok - I know what I'm hacking the next hours ...
22:45 < Minto> bedtime for me, gn8 all
22:45 -!- Minto [~chatzilla@82-217-66-7.cable.quicknet.nl] has quit ["ChatZilla 0.9.61 [Mozilla rv:1.7.3/20040910]"]
22:45 < mnemoc> Minto,rxr: how can i format the integers i dump to cout? (%3d of printf)
22:47 < rxr> what does %3 did ? field width of 3 chars?
22:47 < rxr> << std::setw (3)
22:47 < rxr> if it is the width ...
22:48 < rxr> and
22:48 < rxr> << std::left
22:48 < rxr> << std::right
22:48 < rxr> for the alignment
22:48 < rxr> but default should be ok for you ...
22:48 < mnemoc> i want 'air' between numbers :)
22:49 < rxr> like fixed width in table form?
22:49 < rxr> then the above is for you
22:49 < rxr> btw. is this some university project? or what is it? ;-)
22:50 < mnemoc> boards/basic_board.cc:38: error: `setw' undeclared in namespace `std'
22:50 < mnemoc> university
22:50 < mnemoc> for AI
22:50 < rxr> #include
22:50 < rxr> ic
22:51 < rxr> I love those:
22:52 < rxr> "Default state: Device behavior when this request is received while the device is in the Default state
22:52 < rxr> is not specified.
22:52 < rxr> "
23:17 < mnemoc> rxr: how do i flush cin ?
23:19 < mnemoc> rxr: ?
23:23 < rxr> you need a C++ book
23:24 < rxr> cin.flush ()
23:24 < rxr> ?
23:24 < rxr> tell me if it worked ;-)
23:24 < rxr> did the above setw help you?
23:26 < mnemoc> nope
23:26 < mnemoc> including iomanip killled my class :p
23:27 < mnemoc> players/human_player.cc:18: error: `flush' undeclared (first use this function)
23:27 < mnemoc> calling it as std::cin.flush();
23:28 < mnemoc> rxr: ping :p
23:30 < rxr> pong
23:31 < rxr> well std::cin.flush(); of course
23:31 < rxr> including iomanip should not kill anything ...
23:31 < rxr> both work here.
23:31 < rxr> maybe inspirate you with this
23:31 < mnemoc> std::cin.flush(); doesn't work :(
23:32 < rxr> http://svn.exactcode.de/gsmp/trunk/pcm/tests/Bench.cc
23:32 < rxr> what is the error?
23:32 < mnemoc> players/human_player.cc: In member function `Move HumanPlayer::AskMove()':
23:32 < mnemoc> players/human_player.cc:18: error: `flush' undeclared (first use this function)
23:32 < mnemoc> players/human_player.cc:18: error: (Each undeclared identifier is reported only
23:33 < rxr> show me the relevant lines ;-)
23:33 < mnemoc> Move m;
23:33 < mnemoc> 18 std::cin.flush();
23:33 < mnemoc> 19 std::cout << "Fila a jugar? [1-9] ";
23:33 < mnemoc> 20 std::cin >> m.X;
23:34 < mnemoc> C++ hates me
23:34 < mnemoc> i should have coded in plain C :\
23:34 < rxr> hm - seems only ostream has flufh
23:34 < rxr> did you choose C++ yourself?
23:34 < mnemoc> yes
23:35 < rxr> let me open a book
23:35 < mnemoc> you told "C++ is the most wonderfull language in the world" and i believed you :p
23:35 < rxr> well - although I did not say exactly that - C++ is great ...
23:41 < rxr> for some yet unknown reason there is only flush for ostream's
23:42 < mnemoc> :\
23:43 < mnemoc> i want to read ints, if i add another thing it gets crazy :\
23:43 < mnemoc> that's why i want to flush it
23:43 < mnemoc> as i would do on C
23:44 < rxr> getline reads one line ...
23:45 < mnemoc> atoi?
23:45 < mnemoc> C++ is not as magic as i would expected :\
23:46 < rxr> noone said C++ is magic ...
23:47 < rxr> ok - why do you got to atoi?
23:47 < rxr> although I do not know some flush method for istreams off-hand (seme
23:47 < rxr> seems I never needed that)
23:47 < rxr> you could still do
23:47 < rxr> int i;M
23:48 < rxr> hm - wait
23:48 < mnemoc> i is int :)
23:49 < mnemoc> how would you safely as for three _numbers_
23:49 < mnemoc> ?
23:49 < rxr> what do you mean=
23:50 < rxr> ? even
23:50 < rxr> "as for three" what do you mean?
23:50 < mnemoc> ask*
23:51 < mnemoc> make it simple
23:51 < mnemoc> ask for one number
23:51 < mnemoc> until it gets a number
23:52 < rxr> mom
23:52 < mnemoc> i=-1; while(i<0) { cin >> i; }
23:52 < mnemoc> if you press enter it goes crazy :)
23:52 < rxr> damn - I do too less C++ these days ...
23:53 < mnemoc> what do you use? visual basic? :p
23:54 < rxr> no - too much bash hacking for t2 - and too much open source C crap ...
23:54 < rxr> ok it is that easy:
23:54 < rxr> cout << "your msg" << endl;
23:54 -!- kensai [~kensai@64.237.129.108] has joined #t2
23:54 < rxr> while (! (cin >> i) )
23:54 < rxr> {
23:54 < rxr> cout << "better input a number" << endl;
23:54 < rxr> cin.getline();
23:54 < rxr> }
23:55 < rxr> I'm not sure about the getline#
23:55 < mnemoc> aha
23:55 < rxr> let me take a look
23:55 < mnemoc> i think getline does the trick
23:55 < rxr> I really do not know why they left out flush for istreams - but I have not missed it until tonight ...
23:56 < rxr> ah - no it does not have a getline like that ...
23:57 < rxr> if you want s.th. complex but correct:
23:57 < rxr> cin.ignore (numeric_limits
23:57 < rxr> ;-)
23:57 < mnemoc> o_O
23:58 < mnemoc> ok =)
23:58 < mnemoc> should i run that before cin >> i ?
23:58 < mnemoc> once or everytime?
23:58 < rxr> you should run it in the loop, too - to remove incorrect input of course ...
23:58 < rxr> when you have a dummy object you could do:
23:59 < rxr> getline (cin, stdstring);
23:59 < mnemoc> your complex thing didn't complied :p
23:59 -!- kensai [~kensai@64.237.129.108] has quit ["Leaving"]
23:59 < rxr> you need to include some limits header ...
--- Log closed Fri Nov 26 00:00:42 2004