Camelia, the Perl 6 bug

IRC log for #cdk, 2008-02-15

| Channels | #cdk index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
06:50 jbrefort joined #cdk
08:35 jonalv joined #cdk
08:46 dleidert joined #cdk
09:01 rojas1 joined #cdk
09:02 man-di dleidert: Hallo
09:02 dleidert hi
09:03 man-di dleidert: I havent had time to look into our little problem yet again
09:03 dleidert you could send me the build log
09:03 man-di can do later, currently at work
09:04 man-di dleidert: I cant plug off the network cable of my computer at home from here
09:04 dleidert I will go to get some sleep soon - was working the whole night :) so no need to hurry
09:04 man-di oh
09:04 man-di get sleep
09:04 man-di dleidert: you have weird working times
09:05 dleidert sometimes
09:06 dleidert I was checking the whole html-xml-utils man-pages
09:06 dleidert oh and the ones in debichem too
09:08 man-di dleidert: before you leave: can you tell me where I can get upstream sources for org.xmcml.*?
09:08 dleidert which package is that?
09:08 man-di jumbo52.jar
09:09 man-di its a requirement for cdk
09:09 man-di its an optional part which is not optional :-/
09:10 dleidert no idea, maybe Egon knows
09:10 man-di wil ask him when is back again
09:13 dleidert oh, you meant org.xmlcml
09:13 man-di dleidert: if you ever look for me, I
09:13 man-di dleidert: if you ever look for me,
09:13 man-di dleidert: if you ever look for me, I'm always on in #debian-java/irc.debian.org
09:13 man-di yes, you meant org.xmlcml
09:16 dleidert http://cml.cvs.sourceforge.ne​t/cml/jumbo53/src/org/xmlcml/
09:18 man-di will jumbo 5.4.1 from http://sourceforge.net/project​/showfiles.php?group_id=51361 work too?
09:18 man-di cdk seems to contain jumbo 5.2
09:19 dleidert Egon is IMO the much better person to ask these things :)
09:19 man-di oh well, jumbo 5.2 never existed upstream....
09:19 man-di that starts to become fun
09:44 steinbeck joined #cdk
09:59 thomas_ku joined #cdk
09:59 thomas_ku moin
09:59 man-di hello
10:01 masak joined #cdk
10:04 CIA-31 joined #cdk
10:04 carsten joined #cdk
10:07 steinbeck joined #cdk
10:07 jpansanel joined #cdk
10:53 de_maas joined #cdk
11:04 jpansanel moin
11:26 carsten Moin moin
11:50 Annzi joined #cdk
11:58 egonw joined #cdk
12:00 egonw moin
12:00 man-di egonw: moin
12:04 rojas1 hi egonw
12:05 egonw hi rojas1
12:05 rojas1 one question. may be you know. I made yestarday a update and upgrade with kubuntu and now it doesn't start the graphical login
12:06 egonw make sure the package kdm is installed
12:06 egonw (but not kdm-kde4)
12:06 egonw well... actually, that one too, if you use KDE4...
12:06 egonw but don't use kdm-kde4 as login manager...
12:08 rojas1 ok, thanks
14:24 jonalv Bioclipse 2 are lacking org.openscience.cdk.tools.HydrogenAdder. What's up with that?
14:25 rojas1 I don't know exactly but I think it was just moved to CDKHydrogenAdder
14:27 jonalv hm I ave compilation error here so I thought I should have a look at it
14:28 jonalv is it just a renaming then?
14:28 jonalv egonw, can you comment?
14:32 egonw jonalv: use CDKHydrogenAdder
14:34 jonalv is it just renamed?
14:37 egonw no, new implementation
14:38 egonw better implementation
14:38 egonw IMHO
14:38 egonw :)
14:41 * egonw is going to reboot into has patched kernel, and will brb
14:46 egonw joined #cdk
14:47 jonalv I seem to need addImplicitHydrogensToSatisfyValency but only addImplicitHydrogens is present on the new
14:47 egonw use the altter
14:47 egonw latter
14:47 egonw that's OK...
14:47 jonalv oki
14:47 egonw can you work it out?
14:47 egonw or should I boot my eclipse?
14:47 jonalv I am trying
14:47 jonalv honestly I don't know what it does
14:47 egonw I'll be around...
14:48 jonalv I just saw code that didn't compile so I thought: "Hey great oppurtunety to learn some CDK" :)
14:48 egonw :)
14:50 jonalv but now I seem to need a molecule builder instead of a String to create instances.  Is it cached for each builder?
14:51 egonw builders are cached
14:51 egonw that is...
14:51 egonw there a three builders...
14:51 egonw and only three...
14:51 egonw they are singleton classes, so don't worry about memory use...
14:51 egonw if you already have a molecule...
14:51 jonalv okey
14:51 jonalv well there are loops
14:52 jonalv and I need to move the code into the loop
14:52 jonalv Because I loop over molecules...
14:52 jonalv Guess that's not a problem though..
14:55 jonalv But my question was rather: Is the CDKHydrogenAdder instances cached within the class by the factory mthod?
14:55 jonalv for the different builder types that is...
14:56 jonalv oh I have the src I can just check it :)
14:57 egonw yes
14:57 egonw right
14:57 egonw we user svn:externals :)
14:57 egonw -r
14:57 jonalv nice
14:58 egonw considering the loop will iterature over IMoleculeSet.molecules() or so...
14:58 egonw you can instantiate the adder outside the loop, so, just once
14:59 jonalv but can't the diffferent molecules have different buiolder?
14:59 egonw theoretically: yes
14:59 jonalv *builders
14:59 jonalv son then that would be a nono
14:59 jonalv *so
14:59 egonw CDKHAdder uses .getInstance() right
15:00 egonw ?
15:00 jonalv yes
15:00 egonw that's fairly cheap call...
15:00 egonw so, if you want to make sure...
15:00 jonalv yes I think I do
15:00 egonw putting it inside the loop, would be fairly OK...
15:00 jonalv That's the way I am doing it now anyway
15:03 jonalv hm the old version seems to have returned a Map. The new one doesn't. The code seems to use the map for something... :/
15:04 egonw wtf?
15:04 egonw ok, I'm booting eclipse
15:04 jonalv :)
15:04 steinbeck hehe
15:04 egonw he steinbeck :)
15:04 steinbeck cheers
15:04 jonalv hi steinbeck, it's your old code btw :)
15:04 steinbeck which one?
15:04 steinbeck got a lot of it
15:04 steinbeck :-)
15:05 jonalv AddHydrogenAction
15:05 steinbeck uuuhhh
15:05 steinbeck I don't mind seeing this vanish
15:05 egonw yeah, may be best thing to do...
15:05 egonw it's not adding much functionality anyway
15:06 jonalv is it that broken?
15:06 * egonw will still boot his eclipse into BC2 mode
15:06 egonw well... it does not compile for you :)
15:06 jonalv but that is probably just because the AddHydrogen class is gone and replaced
15:11 egonw oi... have to resetup my BC2 env
15:11 jonalv oki, that will go pretty fast right? :)
15:12 egonw no idea
15:12 egonw at least I found the wiki page now :)
15:12 jonalv right, call if you ahve troubles
15:12 jonalv I _should_  be able to help... :)
15:15 jonalv it uses the Map for undo/redo purposes I think
15:16 egonw possibly
15:16 egonw if you think so... delete it...
15:16 jonalv uh?
15:16 egonw that concept will be dealt with differently in trunk/
15:16 jonalv trunk/ of CDK?
15:16 egonw yes
15:16 egonw which BC2 is using
15:16 egonw OK...
15:16 egonw help
15:16 jonalv yes
15:17 egonw I got circular deps in BC2
15:17 jonalv ?
15:17 jonalv aouch
15:17 jonalv what
15:17 jonalv ?
15:17 egonw a couple of them
15:17 jonalv a couple of them? :)
15:17 egonw e.g. net.bioclipse.r
15:17 egonw .ui
15:18 egonw ,loggin
15:18 egonw .core
15:18 jonalv have you checked out everyhting from the beginning?
15:18 egonw .target.platform + all plugins from core/plugins
15:18 jonalv .core shouldn't depend on r any longer
15:19 egonw ok. I see this:
15:19 egonw .logging has no plugin.xml
15:19 egonw and indicates a dep on .core
15:19 egonw and .core deps on .logging
15:19 egonw that's one circular
15:20 jonalv my logging doesn't depend on core
15:20 egonw should .logging not have plugin.xml?
15:20 jonalv Well I guess we haven't needed anyone...
15:20 jonalv ...yet
15:21 jonalv egonw, I can't quite shake the feeling that you have old plugins together with new plugins in your workspace
15:23 egonw mmm...
15:27 jonalv egonw? any progress?
15:43 egonw no
15:44 egonw ran into someone who showed me a nice PhD thesis...
15:44 jonalv ah
15:45 * egonw is making a fresh checkout
15:45 jonalv good
15:49 egonw still checking out...
15:51 jonalv hm not sure it should take long time. But it is sort of big :)
15:53 masak jonalv: not that big, is it?
15:53 * masak is more ready to blame subclipse
15:53 egonw ah, doing it from the command line
15:53 jonalv not subclipse then... :)
15:54 egonw ok, imported things...
15:54 egonw it's building
15:54 masak egonw++ # command line
15:54 egonw yeah, there they are again
15:54 egonw cyclic deps
15:54 egonw sorry jonalv
15:54 jonalv my bioclipse 2 workspace is almost 100 MB
15:54 jonalv but
15:55 jonalv but
15:55 jonalv I don't have them...
15:55 egonw that's because you use subclipse
15:55 jonalv explain
15:55 egonw btw... let's move this to #bioclipse
15:55 jonalv oki
15:57 steinbeck left #cdk
16:04 CIA-31 bioclipse: carl_masak * r4888 /trunk/playground/plugins/net.biocl​ipse.base2/src/net/bioclipse/base2/ (db/Base2Database.java model/Database.java):
16:04 CIA-31 bioclipse: [Base2Database.java]
16:04 CIA-31 bioclipse: [Database.java]
16:04 CIA-31 bioclipse: * better error reporting when server is unavailable
16:23 CIA-31 bioclipse: egonw * r4889 /branches/bioclipse2/core/plugins/ (4 files in 4 dirs): Removed log4j from net.bioclipse.core to solve cyclic dependencies: it now compiles again
17:14 dleidert joined #cdk
17:22 CIA-31 cdk: miguelrojasch * r10115 /branches/miguelrojasch/react​ion/src/org/openscience/cdk/ (11 files in 7 dirs): updating changed name for class
17:23 rojas1 left #cdk
17:27 man-di egonw: what is the chance to get a the parts in cdk that are marked optional to be optional actually?
17:27 egonw very good
17:27 egonw theoretically...
17:28 egonw I guess the problem that it current fails to compile on those bits?
17:28 man-di right
17:28 man-di and the logic that checks if the optional jar exist is wrong/doesn't work
17:29 egonw ok, will check
17:29 egonw where's the debian/ source ?
17:29 egonw will try to find time to set up a debian testing vmware/kvm thingy
17:29 man-di egonw: the old one in pkg-java svn and the new one on my hard disk
17:29 egonw or chroot otherwis
17:30 egonw ah, ack
17:30 egonw might you send me a copy of your new one?
17:30 egonw what you have now?
17:30 egonw will look at it this weekend
17:30 man-di egonw: development of debian packages is against sid/unstbale not testing. with testing you might have extra problems due to ongoing transitions
17:30 man-di I can do later, sure
17:31 egonw ack
17:31 man-di I will need to move the changes to build.xml to patch files under debian/patches
17:31 egonw explain... you need build.xml to apply patches?
17:32 man-di I did some changes to build.xml to work on debian for now
17:32 dleidert egonw: you can setup a pbuilder-CHROOT to build the packages on Lenny with a Sid environment
17:33 dleidert it's pretty easy
17:33 man-di I will need to move these changes to extra patch files to make the debian dir work for you as it does for me
17:33 egonw dleidert: right...
17:33 egonw never done a pbuilder chroot...
17:33 egonw but have had debian chroots on my ubuntu system...
17:33 egonw will set one up for unstable
17:34 man-di egonw: pbuilder is better for package building as it is always clean
17:34 dleidert that's true, unfortunately it doesn't handle the sun-packages well atm
17:34 man-di dleidert: there is a nice trick for that
17:35 dleidert you have to edit the config-file to set DEBIAN_FRONTEND
17:35 dleidert any other trick I don't know?
17:35 man-di dleidert: just preset the debconf database once with pbuilder login --save-after-login
17:35 dleidert that was also my idea - unfortunately I did not know, how to do that
17:35 dleidert any short-howto?
17:36 man-di and then 'echo "some value" | debconf-set-value' or similar, I have written that down somewhere.
17:36 dleidert I was looking through the man-pages, but did not find/understand howto preset values
17:36 dleidert ok
17:36 man-di for sun java 5: echo "sun-java5-jdk shared/accepted-sun-dlj-v1-1 boolean true" | debconf-set-selections
17:36 man-di similr for sun java 6
17:36 dleidert oh nice
17:37 * dleidert will update his pbuilder CHROOTs
17:37 man-di dleidert: my old workaround was en extra pbuilder base.tgz with sun java jdk preinstalled
17:38 man-di dleidert: I still need two pbuilders, one for main-only and one for main/contrib/non-free
17:38 man-di just to make sure packages in main build-depends on main only
17:38 dleidert yeah, this is also something I miss ... sometimes I would like to extend my Sid-CHROOT with experimental sources
17:38 dleidert but I need two tarballs two
17:38 egonw man-di: nifty
17:39 egonw ok, will check out how to set up pbuilder
17:39 man-di dleidert: experimental, another pbuilder base.tgz :-/
17:39 dleidert would be a nice feature to be able to extend the categories or sources for one tarball
17:39 dleidert just for building
17:39 man-di dleidert: but sid is nice, I have now my own sid mirror at home, helps alot
17:40 dleidert well, caching the packages already helps a lot
17:40 dleidert how large is the mirror?
17:42 man-di around 50 GB with sid/source, sid/i386, sid/amd64, sid/arm and etch/i386
17:42 man-di but I update it only once a week at work, its on a portable USB disk
17:42 man-di I will need to setup piuparts
17:43 man-di I havent used piuparts yet, but I should
17:49 dleidert yeah, it triggers some installation issues
17:49 jbrefort joined #cdk
17:54 CIA-31 bioclipse: egonw * r4890 /branches/bioclipse2/core/plugins/ (4 files in 4 dirs): Reverted patch 4889: it's not the definate solution
17:56 dleidert man-di: thanks for the tip, SUN java packages are now installed in noninteractive mode
17:57 egonw man-di: need to catch the train
17:57 egonw will be online later this weekend...
18:36 rojasm joined #cdk
18:53 rojasm joined #cdk
18:56 rojasm joined #cdk
19:00 Annzi left #cdk
19:06 rojasm joined #cdk
19:08 rojasm joined #cdk
19:17 rojasm joined #cdk
23:00 rojasm left #cdk

| Channels | #cdk index | Today | | Search | Google Search | Plain-Text | summary