Camelia, the Perl 6 bug

IRC log for #cdk, 2009-10-13

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

All times shown according to UTC.

Time Nick Message
00:38 slyrus_ joined #cdk
05:12 jbrefort joined #cdk
05:51 bag joined #cdk
06:38 Gpox joined #cdk
07:26 s_wolf joined #cdk
07:33 miguelrojasch joined #cdk
07:49 sneumann joined #cdk
07:53 egonw joined #cdk
07:57 jonalv joined #cdk
08:02 egonw jonalv: fingerprints, not?
08:02 egonw test performance
08:02 egonw that's 1
08:03 jonalv egonw: yea
08:03 egonw 2. test performance
08:03 egonw 1 = speed
08:03 egonw 2 = that ratio
08:03 jonalv timne efficiency
08:08 jonalv so where do we start? I don't even know thr classes for the different fingerprinters in CDK...
08:10 egonw open IFingerPrinter
08:10 egonw and look for classes implementing that interface
08:10 jonalv oki
08:11 jonalv -me is just gonna get a fresh CDK checkout
08:12 egonw make sure to pick *master*
08:13 jonalv egonw: I ahve your master up to date now
08:14 jonalv egonw: is that the 'official; *master* btw?
08:14 egonw github.com/egonw/cdk
08:15 egonw yes, that's one of the two official repositories
08:15 jonalv oki
08:15 egonw mirroring each other
08:17 jonalv wow, we have: EStateFingerprinter, ExtendedFingerprinter, Fingerprinter, MACCSFingerprinter, PubchemFingerprinter and SubstructureFingerprinter
08:21 egonw yes, not bad :)
08:21 egonw and I will hack up another one today :)
08:21 miguelrojasch1 joined #cdk
08:21 miguelrojasch1 egonw: could you find some email about the patch?
08:21 egonw miguelrojasch1: just awake 30 minutes now...
08:21 egonw my brain is still booting...
08:22 egonw miguelrojasch1: something else... are you in Leiden in the week of 20 Nov
08:22 miguelrojasch1 :)
08:22 miguelrojasch1 I check
08:22 egonw I will be in Amsterdam on 20 Nov for the SWAT4LS conference
08:23 miguelrojasch1 Yes I will be
08:23 miguelrojasch1 in Leiden
08:23 jonalv egonw: what test data file were we planning to use again?
08:23 egonw cheminfbench
08:23 egonw mom, let me look up the repos
08:24 egonw http://github.com/egonw/cheminfbenchmark
08:24 zarah egonw's link is also http://tinyurl.com/ygl9ywo
08:27 CIA-29 cdk: shk3 * r14970 /jchempaint/trunk/.classpath: removed stuff from classpath which should not be there
08:29 jonalv egonw: what say we make our cheminfbenchmark a plugin so it can use CDK?
08:29 egonw what do you mean?
08:30 jonalv egonw: right click and convert it into an Eclipse plugin so it can have deps
08:30 jonalv and we can handle it like we are used to hadnle stuff...
08:31 jonalv otherwise we need to rebuild a cdk jar each time we want to test run a new version of stuff. tight?
08:31 egonw ah... no, don't think so
08:31 jonalv egonw: why not?
08:31 egonw it uses a ant build.xml to run the testing
08:31 jonalv doesn't ant support that?
08:31 egonw check the fingerprint.xml config file
08:32 egonw the things is more like this:
08:32 egonw the testing tools it uses expects things to be in a given jar
08:32 egonw because it will test *multiple* cdk versions
08:33 jonalv hm well mine is complaining about not finging cdk-lean35.jar...
08:34 jonalv egonw: you mean fingerprint-config.xml btw?
08:34 egonw ah... that should not be part of master
08:35 egonw that should be part of my *lean* branch
08:35 jonalv egonw:  I don't even see a 'lean' branch
08:35 egonw yes, weird
08:36 egonw just checking the fip-config.xml and don't see lean mentioned...
08:36 egonw where is it complaining?
08:36 jonalv egonw: in the error view in Eclipse
08:37 egonw ok, so .classpath or so
08:37 egonw ah, indeed
08:37 egonw ok, hang on
08:39 jonalv @ping
08:39 zarah jonalv: pong
08:41 CIA-29 cheminfbenchmark: Egon Willighagen master * r04a304b / .classpath : Removed cdk-lean.jar which does not exist in the master branch - http://bit.ly/49x3GT
08:42 jonalv egonw: okey, so how do I testrun this cheminfbenchmark stuff now?
08:43 egonw see the README
08:43 egonw down the bottom it gives:
08:43 egonw ant -Dconfig=test-config.xml run
08:43 egonw where test-config.xml would be fingerprint-config.xml
08:44 jonalv hm I must learn to run ant stuff from within Eclipse
08:45 egonw oh simple
08:45 egonw just open the Ant view
08:45 egonw right click the target
08:45 egonw but
08:45 jonalv egonw: the ant view is empty
08:45 egonw not sure how to set -Dconfig values
08:45 egonw ah, good :)
08:46 jonalv good?
08:47 egonw that you knew about the vie
08:47 egonw w
08:48 * jonalv tried to run that from commandline and got a stacktrace
08:48 egonw I'll add the latest cdk 1.3.x
08:48 jonalv is this even supposed to work?
08:48 egonw yes, it is
08:48 egonw it did, at least
08:48 egonw this June I used it extensively
08:48 egonw for testing my 'lean' implementation of the CDK interfaces
08:49 * jonalv tgries the fingerptint config instead then..
08:50 egonw pastebin/ for the stacktrace (ST) ?
08:50 jonalv egonw: nah I used the test config but that one seems to have been removed sometime...
08:51 * jonalv is testing the fingerrint config now. How long is that supposed to take?
08:51 egonw oh, what were you trying before then?
08:51 egonw a few minutes, I guess
08:51 jonalv done
08:52 jonalv I was trying the test-config.xml, the one suggested in the README file
08:53 egonw hahaha
08:53 egonw there is no test-config.xml
08:53 egonw no, that's indeed not supposed to work :)
08:53 jonalv egonw: no there isn't
08:53 jonalv that's a bit borken if you ask me...
08:53 egonw it talks swedish?
08:53 jonalv anyways, I am looking at the report.html now
08:54 jonalv yea humpty dumpty, bork, bork, bork
08:54 CIA-29 cheminfbenchmark: Egon Willighagen master * r4672126 / README : Added note that test-config.xml must be replaced by an existing config file - http://bit.ly/TZK8p
08:55 jonalv nonotify seems best, but for some reason I feel that is not what we want?
08:57 jonalv egonw: the LICENSE.txt file is empty... What's up with that?
08:57 egonw dunno
08:57 egonw just noticed that too
08:57 jbrefort joined #cdk
08:57 egonw jonalv: indeed
08:58 egonw nonotify does not send around change events
08:58 egonw causing a speed up
08:58 egonw nonotify is always faster than the normal data classes
08:58 jonalv egonw: but is nonotify correct?
08:59 egonw correct?
08:59 * jonalv thought those change events were important
08:59 egonw depends
08:59 jonalv like if they didn't exist it broke
08:59 egonw some apps depend on it
09:05 jonalv egonw: okey but it does actually make sense to compare with it?
09:06 egonw of course
09:06 egonw there just might be an interaction effect
09:07 jonalv so, how does this tool work now? What do we need to do?
09:07 egonw speed testing
09:07 egonw we compare fingerprints
09:07 egonw types
09:07 jonalv it seems to be that we are going to use it for a bit more than it was built for
09:07 egonw this is just for part 1
09:07 jonalv so we need to add tests to it
09:08 jonalv part 1?
09:08 jonalv ah
09:08 egonw [10/13/09 10:03] <egonw> 1 = speed
09:08 egonw [10/13/09 10:03] <egonw> 2 = that ratio
09:08 jonalv okey
09:08 jonalv egonw: is it running Junit tests?
09:08 jonalv Does it measure time automagicly?
09:10 egonw not sure aobut the internals
09:10 jonalv egonw: donn't we need to change the internals?
09:10 jbrefort joined #cdk
09:10 egonw no
09:11 * jonalv is not so sure about what we need to do then...
09:11 egonw see src/net/openmolecules/benchmark/​driver/FingerprintCDKBench.java
09:12 egonw there is *some* coding involved
09:12 egonw but you see that Japex takes care of all the timing
09:12 egonw we just follow a certain API
09:12 egonw prepare() and run()
09:12 egonw the last is used for the timing
09:13 egonw all code that must be run to do the test but which should not be timed goes into prepare()
09:17 jonalv okey
09:17 jonalv egonw: it feels like everything is already done. What is left to do?
09:18 egonw I don't think it compares the differnt FP types
09:18 egonw or does it?
09:19 jonalv maybe not
09:19 jonalv you want to run our tests on different CDK versions as well?
09:21 egonw no, let's stick to master
09:22 CIA-29 cheminfbenchmark: Egon Willighagen master * r059b77a / mx-cdk-sdf.xml : Added running of CDK trunk, with the SDF reading performance speed up - http://bit.ly/d1hFf
09:22 jonalv egonw: so,  we start by creating a new config and then a couple of Benches then?
09:22 egonw yes, sounds like a plan
09:22 jonalv who does what?
09:23 egonw if you go ahead with this, I'll write that new FingerPrinter ?
09:23 egonw (which uses the sp2 instead of aromaticity)
09:23 jonalv eg which one would that be?
09:23 egonw which removed the need of ring detection
09:23 egonw and hence is faster
09:24 jonalv 0 that sounds like a good idea :)
09:24 egonw well, we have yet to show that the ratio is not affected to much :)
09:24 jonalv yup
09:24 jonalv so what do I call my new config?
09:24 egonw fptype-config.xml ?
09:25 egonw ok, git is downloading the new cdk.jar from the repos
09:25 egonw will get some breakfast now...
09:25 egonw will be back in 5 minutes
09:25 jonalv oki
09:36 egonw back
09:36 egonw wow, was that 10 minutes ?!
09:36 jonalv apharenlty
09:36 jonalv do we use one workbench or mulitple?
09:37 jonalv there seems to be possible to divide things at many levels and I am not sure what difference it make...
09:38 egonw I think one bench for each FP
09:39 jonalv since one can pass apramaters it would be possible to just use one bencha for all...
09:40 egonw yeah, good possibility actually
09:40 egonw since they all implement IFingerPrint
09:40 egonw you can make the actual FP type to use a parameter
09:40 jonalv egonw: yes, but the prepare method would start with a big switch. I don't like that
09:40 egonw true
09:40 egonw ah, no
09:41 jonalv better one bench for each and then a Runner class that contains the run code
09:41 egonw not if you have the parameter give the class name
09:41 jonalv egonw: I also wonder about instantiating of the fingerprinters. Aren't there different params to set for them all?
09:42 egonw type="org.openscience.cdk.​fingerprint.Fingerprinter"
09:42 egonw some have parameters indeed
09:42 jonalv egonw: so I think we want a bench for each so we can set those params from xml
09:42 egonw ack
09:43 CIA-29 cdk: mark_rynbeek * r14971 /jchempaint/trunk/src/main/​org/openscience/jchempaint/ (3 files in 3 dirs):
09:43 CIA-29 cdk: Pasting does not create a new AtomContainer.
09:43 CIA-29 cdk: Spacing between existing and newly pasted structures imrpoved
09:43 jonalv do we get statistics for each testCase?
09:44 jonalv yes we do
09:45 egonw http://apple.slashdot.org/story/09/10/12/2259228​/Major-Snow-Leopard-Bug-Said-To-Delete-User-Data
09:45 zarah egonw's link is also http://tinyurl.com/ygj74wp
09:45 jonalv hm this needs some planning in order to get a nice report
09:49 jonalv CDK code is wider than 80 cols, I have to scroll in two directions. :(
09:49 egonw file bug reports
09:49 egonw the more, the marrier
09:49 jonalv egonw: is there a policy about 80 cols?
09:49 egonw yes
09:49 egonw but not from the start I guess
09:50 egonw now you see what I am trying to maintain
09:50 egonw the CDK is not the most maintainer friendly code :)
09:50 jonalv egonw: then it's ok if I refactor files to within 80 cols?
09:50 egonw yes, and please do that 100% separate from any code changes
09:51 jonalv oki
10:00 CIA-29 cdk: mark_rynbeek * r14972 /jchempaint/trunk/src/main/org/openscienc​e/jchempaint/action/CopyPasteAction.java:
10:16 * jonalv @ lunch
10:19 egonw slyrus_: adding E and Z stereochemistry items to IBond:Stereo
10:38 CIA-29 cdk: shk3 * r14973 /jchempaint/trunk/src/main/​org/openscience/jchempaint/ (3 files in 2 dirs): the copy&paste with inchi works, generate inchi removed from applet
11:04 CIA-29 cdk: shk3 * r14974 /jchempaint/trunk/src/test/org/openscience​/jchempaint/JCPEditorAppletBugsTest.java: thanks to now having the UP_OR_DOWN_INVERTED, one test less fails.
11:38 bag joined #cdk
11:39 mgerlich joined #cdk
11:48 * jonalv is back (finally)
11:49 jonalv egonw: are tab cahrs allowed in cdk code?
11:49 egonw strongly discouraged AFAIK
11:49 egonw keep in mind, that there policies were not in place fromt he start
11:49 jonalv I just want the go n removing them
11:49 egonw but feel free to use 'git blame' if you feel communicative...
11:49 jonalv |go on
12:23 CIA-29 cdk: mark_rynbeek * r14975 /jchempaint/trunk/src/main/​org/openscience/jchempaint/ (InsertTextPanel.java action/CopyPasteAction.java): added undo redo for pasting
12:24 CIA-29 cdk: mark_rynbeek * r14976 /jchempaint/trunk/src/main/org/openscienc​e/jchempaint/action/CopyPasteAction.java: correct indentation
12:28 CIA-29 cdk: mark_rynbeek * r14977 /jchempaint/trunk/src/main/​org/openscience/jchempaint/ (3 files in 3 dirs): Moved method generateModel to JChemPaintAbstractApplet
12:51 * jonalv is back again (He keeps getting sidetracked today...)
12:52 egonw same here :)
13:10 jonalv http://github.com/egonw/cdk/blob/mast​er/src/main/org/openscience/cdk/finge​rprint/PubchemFingerprinter.java#L602
13:10 zarah jonalv's link is also http://tinyurl.com/yzowuq5
13:10 jonalv gah my eyes!
13:20 jonalv egonw: some further sudies od the fingerprinters reveal that only GingerPrinter, ExtendedFinger (the same paramaters) and PubchemFingerPrinter (String[] SMARTS) takes paramaters
13:21 jonalv so I guess I can make do with a bit fewer 'benches'
13:21 egonw yes perhaps...
13:21 egonw the last should do with the default params
13:22 egonw for the first two, I know that Rajarshi has explored the number of bits too
13:22 egonw but that should not matter for speed too much, I guess
13:28 egonw jonalv: it's just one patch, correct?
13:28 jonalv egonw: to a couple of files yea
13:29 egonw yes, yes
13:29 egonw meant not to miss a second separate commit when doing the fork queue :)
13:29 jonalv k, no it was jsut easier to fix it than to file a bug repport
13:30 egonw jonalv: you checked it all still compiles with 'ant clean dist-all test-dist-all' ?
13:30 jonalv but the PubchemFingerprinter looks TERRIBLE
13:30 jonalv egonw: nope
13:30 * jonalv feels like masak
13:30 jonalv "it's jsut a harmless chagne, what could possible go wrong" :)
13:31 egonw right :)
13:31 egonw did I mention this before: http://chem-bla-ics.blogspot.com/2009/07/​things-to-check-before-you-consider.html
13:31 zarah egonw's link is also http://tinyurl.com/ykv7c4t
13:31 egonw oh wait...
13:31 jonalv egonw: no honestly I am not familiar with how toi test things. I trust the anarchy
13:31 egonw I just sent that to cdk-devel
13:31 egonw jonalv: can you send me the 'git format-patch' ?
13:31 egonw I'll test compiling then
13:32 jonalv egonw: sure (how did I do that again?)
13:32 egonw git log
13:32 egonw pick the last commit hash *before* your commit
13:32 egonw then:
13:32 egonw git format-patch HASH
13:32 egonw which will create one or more 00**.patch
13:33 jonalv egonw: so you prefer patches before github?
13:34 egonw depends on the situation
13:34 egonw for the CDK, yes
13:34 egonw as I want to make sure it still compiles after the patch
13:34 egonw do you mind if I 'improve' the patch commit first line?
13:34 egonw make it
13:34 egonw layout clean ups
13:35 jonalv what's 'ups'?
13:35 egonw cleanups ?
13:35 egonw plural of up
13:36 jonalv (no, off course not you make it whatever you want I just want cred :)
13:36 egonw "clean up of layout" ?
13:36 jonalv oki
13:38 * egonw is making progress with the HybridOnlyFingerprinter
13:38 egonw written it
13:38 egonw now writing unit tests
13:39 jonalv sounds to me like you are not doing it right :)
13:39 jonalv first tests then code :)
13:39 egonw well, I am copy/pasting the unit tests from the Fingerprinter, so I know the tests are properly testing the contract
13:39 egonw but, in principle you're correct
13:40 jonalv k
13:44 egonw ah, crap...
13:44 * egonw is not happy with copy/pasting
13:44 jonalv good
13:44 egonw and will prepare a better patch: AbstractFingerprinterTest
13:45 egonw any test should really apply to *all* IFingerprinter's
13:45 egonw and gradually the CDK improves...
13:45 jonalv why can't I seem to get a ref to the PubchemFingerpritner?
13:45 jonalv egonw: is that not exported?
13:45 egonw do you have the fingerprint module ?
13:47 egonw jonalv: one question
13:47 egonw have you deliberately not set:
13:47 jonalv egonw: I have no idea. I am in the cheminfbenchmark project
13:47 egonw [user]
13:47 egonw name = Jonathan Alversson
13:47 egonw in ~/.gitconfig
13:47 egonw or ./.git/config ?
13:48 jonalv egonw: I am pretty sure I did that when I installed git the first time
13:48 egonw well, not according to your CDK patch
13:48 egonw which has a mere 'jonalv'
13:48 egonw shall I update that?
13:49 jonalv yes please
13:49 jonalv egonw: or well me name is jonalv and my email adress is set
13:50 jonalv egonw: is there a 'nick' property too?
13:56 jonalv egonw: I can use the MACCSFingerPrinter, EStateFingerprinter and the SubstructureFingerprinter but not the PubchemFingerprinter. Is it special somehow?
13:59 egonw should not
14:02 slyrus_ egonw: cool!
14:04 egonw slyrus_: http://pele.farmbio.uu.se/cgi-bin/git​web.cgi?p=jchempaint-primary.git;a=sh​ortlog;h=refs/heads/bioclipse-2.1.x
14:04 zarah egonw's link is also http://tinyurl.com/l72447
14:04 egonw it's the mere concept now, but I start
14:07 slyrus_ Ok. I need to spend some time figuring out how this is all supposed to work, but the problem with E/Z is that it is dependent on the properties of the substituents. change an atom on a side chain and you've just changed the configuration across a (perhaps distant) double bond.
14:07 egonw yes
14:08 egonw I think we are a long way from seeing E/Z in the editor
14:08 slyrus_ the things that I see needing to happen are 1) representation of the configuration in the bond/atomcontainer 2) smiles parsing of the configuration metachars 3) the 2d structure gen algos need to respect this
14:08 slyrus_ that's my xmas wish list :)
14:08 egonw right 1) is now done :)
14:08 slyrus_ :)
14:09 slyrus_ thanks! (always best to tackle the trivial tasks first :) )
14:09 egonw yes
14:10 egonw BTW, you could consider writing one or two unit tests with the expected behavior
14:10 egonw that tends to speed up implementation
14:10 egonw the core developers hate failing unit tests
14:12 jonalv egonw: I think I have missed something...
14:15 egonw how so?
14:16 jonalv egonw: our test cases should probably be different data input not different fingerpritners
14:16 jonalv the different fingerpritners should be different drivers
14:16 egonw why?
14:16 jonalv otherwise we only have one driver
14:17 slyrus_ egonw: ok, will do. maybe not for a couple days though.
14:17 egonw no rush
14:18 jonalv egonw: or do we always run default flavour fingerpritners on the same testdata?
14:18 egonw we can/should do both I guess
14:18 jonalv brb
14:22 jonalv egonw: was that for me?
14:22 egonw the last one, yes
14:23 egonw did someone notice an internet fail yesterday? http://www.thelocal.se/22618/20091013/
14:23 zarah egonw's link is also http://tinyurl.com/yl2x4pj
14:23 jonalv so, hence we should have one driver for each fingerpritner
14:23 CIA-29 cdk: mark_rynbeek * r14978 /jchempaint/trunk/build.xml: further shrinking: io_cml jar and nu.xom
14:23 egonw yes, probably best thing to do
14:53 jonalv egonw: hm I have run into some ClassNotFoundExceptions now...
14:53 egonw details?
14:53 jonalv [java]   Warning: Unable to load driver 'Fingerprinter'
14:53 jonalv [java]            java.lang.ClassNotFoundException: net.openmolecules.benchmark.driver.​fingerPrinting.FingerPrinterDriver
14:53 jonalv
14:54 jonalv hm
14:54 jonalv strange
14:54 jonalv nm
14:57 egonw solved?
14:57 jonalv soon
15:07 jonalv egonw: done
15:07 jonalv egonw: pull
15:08 egonw from your fork?
15:08 jonalv egonw: yea
15:08 CIA-29 cheminfbenchmark: jonalv master * r6eb31fd / .gitignore : added bin to gitignore - http://bit.ly/198byf
15:09 egonw ah, still not updated your [user] name = I see :)
15:09 jonalv egonw: nah I want it to say jonalv so ERC goes ping becasue my name is mentioned
15:09 egonw true :)
15:10 jonalv when the CIA bots kick in :)
15:10 jonalv CIA++
15:10 CIA-29 cheminfbenchmark: jonalv master * ra9d577b / (8 files in 2 dirs):
15:10 CIA-29 cheminfbenchmark: created fptype-config.xml + Drivers
15:10 CIA-29 cheminfbenchmark: for compairing CDK-fingerprints - http://bit.ly/Y9inn
15:13 jonalv egonw: I could not get the PubchemFingerPrinter to work though. Maybe you can have a look at that?
15:13 egonw please email to cdk-devel and ask Rajarshi
15:13 egonw he's in on the paper too, and wrote the thing
15:13 egonw he knows about cheminfbench too
15:14 jonalv egonw: O so it's not jsut something easy that you can spot at once?
15:14 egonw no idea
15:14 egonw what is the problem?
15:14 egonw yes, I can have a look...
15:15 egonw just assumed that if you did not spot the problem, neither would I
15:15 egonw but perhaps I am smarter afterall :)
15:15 egonw hahahah
15:15 * egonw ducks
15:15 jonalv egonw: open the PubchemFingerprinterDriver and try uncomenting the line after the RuntimeException
15:15 egonw ack, hang on
15:16 jonalv zarah: slap egonw
15:16 * zarah decomposes egonw into several parts using the Banach-Tarski theorem and reassembles them to get two copies of egonw!
15:16 egonw oh, cool!
15:16 egonw zarah: that's just what I needed!
15:16 zarah egonw: please rephrase or stop trying to be witty :)
15:16 egonw zarah++
15:17 lely7 joined #cdk
15:17 lely7 left #cdk
15:21 egonw jonalv: what exception do you get?
15:21 jonalv egonw: none, it doesnät compile if I u ncomment that row
15:21 egonw that uncommented line looks fine to me
15:21 jonalv I can't find the class...
15:21 jonalv the one throwing an ecxeption_
15:21 egonw did you try Open Type in Eclipse ?
15:22 jonalv egonw: yes but that is not the same
15:22 egonw no, true
15:22 egonw did you check the depends?
15:22 jonalv these are not plugins and I ahve CDK checked out in the same workspace
15:23 egonw the PubChem FP is only available from trunk
15:23 egonw 1.3.x
15:23 egonw and the current .classpath only has 1.2.x listed
15:23 jonalv egonw: okey that's it then
15:24 jonalv egonw: But we should run against as late a version as posible or?
15:24 egonw for our application we are only interested in 1.3.x indeed
15:30 jonalv egonw: have you tried running it?
15:30 egonw no, not yet
15:30 egonw I actually get this exception for all configs:
15:30 egonw [java] Exception in thread "main" java.lang.IncompatibleClassChangeError: Found class com.sun.image.codec.jpeg.JPEGImageEncoder, but interface was expected
15:30 egonw [java]     at org.jfree.chart.encoders.SunJPEGEncoderAdap​ter.encode(SunJPEGEncoderAdapter.java:136)
15:30 egonw have not resolved that yet
15:32 egonw solved
15:32 egonw JFreeChart fails on OpenJDK6 :(
15:32 egonw will need to update that version, I guess
15:32 egonw I'm sure a recent version works
15:32 egonw as the JFreeChart authors is rather pro OpenJDK :)
15:33 jonalv sounds good
15:34 egonw OK, running your test now
15:46 jonalv egonw: ...and?
15:47 egonw great!
15:50 egonw http://sourceforge.net/tracker/inde​x.php?func=detail&amp;aid=2877961&a​mp;group_id=20024&amp;atid=320024
15:50 zarah egonw's link is also http://tinyurl.com/yllojzx
15:51 CIA-29 cdk: Egon Willighagen 73-rdf * r2666459 / (3 files): Added a few missing dependencies - http://bit.ly/9Kr85
15:51 CIA-29 cdk: Jonathan Alversson master * r5f7cb53 / (6 files):
15:51 CIA-29 cdk: Clean up of layout.
15:51 CIA-29 cdk: Changed 80 cols and no tabs where feasible
15:51 CIA-29 cdk: Signed-off-by: Egon Willighagen <egonw@users.sourceforge.net> - http://bit.ly/4lnIU
15:52 egonw OK, have to go make dinner now...
15:52 egonw I might finish the new fingerprinter tonite, and will be at BMC tomorrow
15:52 jonalv egonw: oki cya tomorrow
15:53 egonw I suggest we try to implement testing scripts (not using Japex) tomorrow for testing that ratio of true/false hits
15:53 egonw based on the FP filterting
15:53 egonw that ratio we talked about
15:54 jonalv yea
16:06 CIA-29 cdk: shk3 * r14979 /jchempaint/trunk/src/main/​org/openscience/jchempaint/ (9 files in 5 dirs): Got the radical stuff working
16:56 jbrefort joined #cdk
17:51 CIA-29 cdk: shk3 * r14980 /jchempaint/trunk/src/main/org/openscien​ce/jchempaint/action/CleanupAction.java: the last commit was unwanted
18:04 jbrefort joined #cdk
20:30 egonw_ joined #cdk
20:31 slyrus_ joined #cdk

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