Camelia, the Perl 6 bug

IRC log for #cdk, 2009-04-29

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

All times shown according to UTC.

Time Nick Message
05:08 bag joined #cdk
05:11 egonw joined #cdk
05:15 jbrefort joined #cdk
06:11 rojasm joined #cdk
06:29 Gpox joined #cdk
06:40 egonw moin
06:40 zarah hi egonw
06:40 sneumann mornin
06:41 Andreas joined #cdk
06:41 Andreas hi
06:41 zarah oh hai Andreas
06:41 egonw hi Andreas
06:41 egonw Andreas: there are a few bots around here...
06:41 egonw one that logs what is being said here
06:41 egonw CIA who reports commit messages
06:41 egonw and zarah
06:42 egonw who's function is to greet people and relay messages and keep karma
06:42 egonw the rest is genuine human, but some in leach mode (which is fine)
06:43 Andreas its a very nice bot
06:43 egonw right now, people are online from DE, NL, SE, TW, FR
06:44 Andreas so i hope that the genuine humans can help me with a problem :)
06:46 Andreas i want to start a single unit test of the MDLRXNV3000ReaderTest class
06:46 egonw we can try :)
06:46 egonw ok
06:46 Andreas i have tried the following:
06:47 Andreas ant -Dtestclass=MDLRXNV3000ReaderTest junit-test
06:47 Andreas and i always get the error:
06:47 egonw try:
06:47 Andreas junit-test:
06:47 Andreas [echo] JUnit tests class: testclass=MDLRXNV3000ReaderTest.java
06:47 Andreas [junit] Running org.openscience.cdk.MDLRXNV3000ReaderTest.java
06:47 Andreas [junit] Tests run: 1, Failures: 0, Errors: 1, Time elapsed: 0 sec
06:47 Andreas [junit] Test org.openscience.cdk.MDLRXNV3000ReaderTest.java FAILED
06:48 egonw ant -Dtestclass=io.MDLRXNV3000ReaderTest junit-test
06:48 Andreas ok
06:48 egonw note the second line of your paste
06:48 Andreas i see
06:48 egonw that should match the actual location
06:49 egonw the testclass param needs all after the prefix org.openscience.cdk.
06:50 Andreas hmm now there are 4 errors:
06:50 Andreas junit-test:
06:50 Andreas [echo] JUnit tests class: testclass=io.MDLRXNV3000ReaderTest
06:50 Andreas [junit] Running org.openscience.cdk.io.MDLRXNV3000ReaderTest
06:50 Andreas [junit] Tests run: 15, Failures: 0, Errors: 4, Time elapsed: 0,062 sec
06:50 Andreas [junit] Test org.openscience.cdk.io.MDLRXNV3000ReaderTest FAILED
06:51 Andreas and i have changed nothing in the code
06:51 egonw can you list one?
06:51 egonw I might have a clue
06:53 egonw try: ant jarTestdata
06:53 egonw and then run the test again
06:53 Andreas the errors were not displayed.
06:54 egonw they are stored in reports/
06:55 Andreas Testcase: testReadReactions1(org.openscience.c​dk.io.MDLRXNV3000ReaderTest):Caused an ERROR
06:55 Andreas null
06:55 Andreas java.lang.NullPointerException
06:55 Andreas at java.io.Reader.<init>(Unknown Source)
06:55 Andreas at java.io.InputStreamReader.<init>(Unknown Source)
06:55 Andreas at org.openscience.cdk.io.MDLRXNV3000Rea​der.<init>(MDLRXNV3000Reader.java:84)
06:55 Andreas at org.openscience.cdk.io.MDLRXNV3000ReaderTest.te​stReadReactions1(MDLRXNV3000ReaderTest.java:72)
06:55 egonw OK, yes, I expected that
06:56 egonw [08:53:18] <egonw> try: ant jarTestdata
06:56 egonw that will solve that
06:56 Andreas still running...
06:56 Andreas and it runs without any errors. Thanks..
06:57 Andreas you saved my day
06:58 egonw Andreas: please note that this class is likely not been used much...
06:58 Andreas i have noticed that and i have found a bug there.
06:59 Andreas i will try to write a unit test and create a patch with git
07:00 egonw great
07:14 rojasm egown: about the email in dev-list. should we write a patcher tracker in http://sourceforge.net/tracker/?func=a​dd&amp;group_id=20024&amp;atid=320024 before we create a patch with git?
07:14 zarah rojasm's link is also http://tinyurl.com/cs8h5l
07:17 Andreas left #cdk
07:21 Andreas joined #cdk
07:25 olass joined #cdk
07:26 egonw rojasm: no, you registed finished patches there
07:26 egonw but that is not needed for bug fixes, but still possible
07:47 jonalv joined #cdk
08:08 shk3 joined #cdk
08:12 olass left #cdk
08:21 Andreas egonw: I have made some changes to the MDLRXNV3000Reader class and rebuilt all with "ant clean dist-all test-dist-all jarTestdata". if i use now "ant -Dtestclass=io.MDLRXNV3000ReaderTest junit-test" it seems that the changes not take place.
08:22 egonw mom
08:24 egonw Andreas: can you put the diff on cdk.pastebin.com ?
08:24 egonw or any other pastebin service
08:24 egonw e.g. gist.github.com
08:25 egonw ah, maybe better to that
08:25 Andreas please not so fast because i am new to that
08:26 egonw go to gist.github.com
08:26 egonw and paster your diff in the big text box
08:26 egonw and hit the 'Paste' button
08:28 Andreas done
08:28 egonw URL?
08:29 Andreas git clone git://gist.github.com/103663.git gist-103663
08:34 egonw Andreas: please do combine layout changes with code changes
08:35 egonw as that makes it impossible to see accurately what code has changed
08:36 egonw I cannot see what changed now
08:36 egonw too many layout changes
08:36 egonw and it even seems reordering of location of methods... ?
08:36 Andreas may i put comments on the changed code
08:37 Andreas i have not reordered anything
08:37 Andreas i only made some code changes within an existing class
08:37 Andreas *method
08:37 egonw mmm....
08:37 egonw which version?
08:37 egonw those layout changes really change the complete file
08:38 Andreas the actual branch. Loaded with $ git clone git://cdk.git.sourceforge.net/gitroot/cdk
08:39 Andreas ah i know what you mean
08:39 Andreas i have pressed ctrl shift f to format the code
08:39 Andreas thats the problem
08:47 egonw ok, please never do that :)
08:47 egonw makes reviewing impossible
08:48 Andreas i will reload the branch and change it again
08:50 egonw ok, thanx
09:07 Andreas so now it should be right
09:07 Andreas git clone git://gist.github.com/103676.git gist-103676
09:15 egonw ok, first comment:
09:15 egonw when you patch a file, you are entitled and encouraged to update the copyright line
09:15 egonw so, on line 4, you would add
09:16 egonw 2009  Andreas Foo <your.email@gmail.com>
09:16 egonw or something like that
09:16 egonw it's good to make that a habit
09:17 Andreas ok
09:18 egonw second comment
09:18 egonw please add unit tests
09:18 egonw preferably send those *before* the patch to fix it
09:19 egonw it helps the reviewer under stand what is being fixed
09:19 Andreas first the unit test which shows all the errors ant then the whole patch
09:20 egonw yes
09:20 egonw that can still be one email or so
09:20 egonw but as separate patch it helps people verify the problem first
09:20 egonw see that the unit test actually is failing
09:21 egonw have you tried: git format-patch ?
09:23 Andreas i was just developing all and my problem was that the unit test does not use the new code although  i have rebuilt all with "ant clean dist-all test-dist-all jarTestdata"
09:27 egonw why do you think it is not using your new code?
09:27 egonw you can check the jar timestamps
09:27 egonw but pretty sure the ant clean deleted all old ones
09:28 Andreas i have changed the output text and it was shown the old text
09:29 egonw what output text?
09:29 Andreas error logging
09:35 jonalv anyone here ever resolved a "Deleted merge conflict"?
09:36 egonw a what?
10:15 maclean joined #cdk
10:44 maclean egonw : I assigned the fully-qualifed class names 'bug' to you, if that's okay. It's really low priority, though :)
10:45 egonw assigning bugs to me is always ok
10:45 egonw if I feel I am not to blame, I reassig
10:45 egonw :)
10:46 maclean oh, you're not to blame (probably). In, fact I have an idea who is responsible for the java.lang ones...
10:49 egonw yeah, no excuses there
11:06 jpansanel moin
11:06 zarah hi jpansanel
11:11 steinbeck joined #cdk
11:22 CIA-58 cdk: shk3 * r14577 /cdk/branches/jchempaint-primary/src/​main/org/openscience/cdk/controller/ (ControllerHub.java IChemModelRelay.java): added comments
12:14 Andreas egonw: I have fixed all problems and created a .patch file. Shall I mail the patch to the dev mailing LIS
12:15 Andreas *list
12:25 maclean joined #cdk
14:02 maclean joined #cdk
14:05 shk3 joined #cdk
14:14 rojasm joined #cdk
14:16 egonw maclean, shk3: if you apply patches to classes in jchempaint-primary *outside* the modules control and render, please commit them as separate patch commits
14:16 egonw and send them in for patches for cdk-1.2.x
14:17 shk3 I was thinking about that
14:17 maclean okay.
14:17 shk3 we should have a deadline (could be now) from which we work back
14:17 egonw shk3: some have, but at least three have not been brought up for cdk-1.2.x
14:17 maclean I don't think that I have done that, but okay.
14:17 shk3 I haven't done it either
14:17 maclean egonw : example?
14:17 shk3 i prepared a diff, but did not yet check it.
14:17 shk3 so I would say: From now on, we do it
14:17 shk3 and work backwards.
14:18 egonw maclean: GeometryTools, commits 14507, 14534, 14542
14:18 egonw shk3: already doing the work backwards
14:18 egonw from now on is great
14:18 egonw maclean: those are commits to jchempaint-primary
14:19 maclean right.
14:19 egonw maclean: and commit 14552
14:33 bag joined #cdk
14:42 shk3 joined #cdk
14:46 steinbeck_ joined #cdk
14:55 maclean_ joined #cdk
15:05 rajarshi joined #cdk
15:05 rajarshi hi
15:05 zarah hello rajarshi, you fantastic person you
15:06 egonw hej rajarshi
15:07 maclean_ joined #cdk
15:08 egonw rajarshi: has BioIT started yet?
15:08 rajarshi started on Monday :)
15:08 rajarshi Tony's talking now, then PMR
15:08 egonw ah, yes, right
15:09 egonw was wondering about that meeting
15:09 egonw PMRs talk should be nice
15:09 rajarshi yes
15:09 rajarshi Chem4Word looks pretty neat
15:09 egonw jim shows steinbeck_ and me the thing in action last week :)
15:09 egonw yes, indeed
15:09 egonw I might actually miss Word for once :)
15:10 rajarshi hopefully, non-Word users will be able to parse out chemistry without needing Windows or Visual Studio
15:10 egonw yes, that should be no problem
15:10 egonw here comes strigi-chemistry :)
15:10 rajarshi :)
15:10 rajarshi increase your todo by 1 more item
15:11 egonw it's on it :)
15:11 egonw I'm so glad I have Frequence 3 again...
15:11 egonw really need to fix the audio on my laptop
15:11 egonw jaunty--
15:16 egonw rajarshi: looking forward to your slides
15:17 rajarshi on slideshare
15:17 rajarshi already
15:21 egonw URL?
15:22 rajarshi http://www.slideshare.net/rguh​a/open-source-cheminformatics
15:22 zarah rajarshi's link is also http://tinyurl.com/dbrc3p
15:40 jbrefort joined #cdk
15:41 egonw how's Tony's talk?
15:44 rajarshi nice as usual
15:44 rajarshi PMR speaking now
15:44 rajarshi semantic chemistry
15:44 rajarshi open data
15:44 egonw sure :)
15:44 rajarshi nice open source session
15:47 egonw grmp...
15:47 egonw one would hope that 1.2.0 -> 1.2.1 would fix things
15:47 rajarshi ?
15:47 egonw instead, I find new bugs in the CDK
15:47 egonw public org.openscience.cdk.interfaces.IMolecule getMolecule(int number)
15:47 egonw {
15:47 egonw return (Molecule)super.getAtomContainer(number);
15:47 egonw }
15:47 egonw and not refering to maclean_ bug report
15:48 rajarshi why is it a bug?
15:48 egonw but the getAtomContainer() should be casted into an IMolecule !
15:48 egonw not Molecule
15:48 maclean_ indeed.
15:48 rajarshi oh right
15:48 maclean_ but if we didn't have (I)Molecule...
15:49 maclean_ :)
15:49 rajarshi ban the IMolecule :)
15:49 * egonw is getting headaches
15:49 maclean_ :( sorry
15:49 egonw too much crap code today
15:49 maclean_ that can get you down
15:50 maclean_ oh, egonw , rajarshi: quick question - there isn't a convertExplicitHydrogensToImplicitHydrogens mehod is there?
15:50 maclean_ there is the opposite method in AtomContainerManipulator.
15:50 rajarshi no I don't think so
15:51 maclean_ oh, ok
15:51 rajarshi makes sense to have one though
15:52 maclean_ in AtomContainerManipulator, I guess.
15:54 egonw yes, that certainly makes sense
15:54 rajarshi right
15:55 egonw rajarshi: was already wondering about your generous travel funds :)
15:55 maclean_ I did it with ACM.countExplicitHydrogens, and IAtom.setHydrogenCount
15:55 egonw would have been rather ironic :)
15:55 egonw plenty of funding, but no visum :)
15:55 rajarshi what is ironic?
15:55 egonw [17:55:19] <egonw> plenty of funding, but no visum :)
15:56 egonw see email steinbeck_
15:56 egonw of a few minutes back
15:56 rajarshi aah
15:56 rajarshi :)
15:56 egonw maclean_: make sure to to count the existing implicit Hs too!
15:56 maclean_ why would there be any
15:57 rajarshi if someone reads in a 'partially hydrogenated' structure
15:57 egonw maclean_: don't get me started...
15:57 maclean_ oh, right.
15:57 * egonw murmers ...SMILES...
15:58 maclean_ I know that these structures all have only explicit hydrogens
15:58 maclean_ as they are made with molgen
15:58 egonw maclean_: I tend to say....
15:58 egonw Murphy's law is peanuts...
15:58 egonw try cheminformatics
15:58 maclean_ do partially hydrogenated smiles spread better straight from the fridge?
15:59 egonw no, they definately need some day light first
15:59 egonw hahahahhaa
15:59 rajarshi :)
16:00 rajarshi battery is dying
16:00 rajarshi bbl
16:00 egonw ok
16:00 egonw regards to PMR and Tony
16:06 egonw dinner
16:06 egonw bbl
16:22 maclean_ @tell rajarshi it seems that the method AtomContainerManipulator.removeHydrogens also updates the hydrogen count!
16:22 zarah Consider it noted.
16:22 maclean_ most kind
16:33 uli_ joined #cdk
16:37 uli_ hi, just wanted to install cdk and play around a bit with it with groovy, but still got stuck with an "unreported exception" error, is this a known problem?
16:39 maclean_ hmmm. not that I know of.
16:41 maclean_ what was the code? - could you put it on cdk.pastebin.com
16:41 uli_ cdk-1.2.1, jumbo5.4.2,  java-6-sun-1.6.0.10
16:41 uli_ ok
16:41 maclean_ or any other online copy/paste thing
16:42 maclean_ the groovy code, that is
16:43 uli_ no groovy yet, ant failed
16:43 uli_ http://cdk.pastebin.com/db09e494
16:43 zarah uli_'s link is also http://tinyurl.com/ckpvfk
16:43 maclean_ oh, right I see it
16:45 Andreas joined #cdk
16:46 maclean_ hmmm. I guess  that it is jumbo (which is the CML library) that has changed somehow, that the CDK has not.
16:47 maclean_ if you are building cdk from source, there should be a jar file that is compatible in the jars directory
16:47 uli_ can you recommend a working jumbo version? tried three already
16:48 maclean_ do you need stuff that is in jumbo that isn't in the jarfile in the cdk distribution?
16:48 uli_ my jar directory did not exist, picked all the dependencies in a very lengthy procedure by myself
16:49 maclean_ ahh
16:50 maclean_ weird - where did you get the cdk copy from? (I'm not blaming you :) I'm just curious...)
16:51 uli_ sf
16:52 uli_ wait...
16:53 uli_ ooops, there's a cdk.jar .... omg
16:53 uli_ thought i'd had to built it by myself
16:54 maclean_ :) no, not necessary.
16:54 uli_ spent hours, but learnt a lot about ant ;)
16:54 maclean_ I've done that kind of thing for packages.
16:55 maclean_ the worst is when it doesn't work at the end of a lengthy dependancy-hunt
16:55 maclean_ but yes, learning about ant is good.
16:58 uli_ :) so thanks anyway, i'll be back later.
16:59 maclean_ no problem
17:09 egonw hi uli_
17:09 egonw uli_: who? uli_ f?
17:10 maclean_ he's gone, I think.
17:10 maclean_ he was trying to get the cdk working with groovy - I blame you :)
17:12 uli_ no, still here, I saw it ;)
17:14 * maclean_ looks innocent
17:16 maclean_ I'm not a big fan of groovy, but I am finding using a dynamic language (javascript, at the moment) wrapper around the CDK is quite useful.
17:16 * egonw is listening to Billy Joel, and bought last week keyboard music of several of his hits last week...
17:16 egonw in Cambridge
17:17 egonw .. on a wednesday afternoon ..
17:17 egonw (oh no, he starts singing again...)
17:17 uli_ i do some ruby stuff at the moment, so i think groovy is the best way to try out cdk
17:18 maclean_ oh for sure.
17:18 * egonw will now take kids to bed... bbl
17:19 uli_ "Groovy is now executing" ...how frightening
17:21 maclean_ :)
17:24 Andreas joined #cdk
20:22 Andreas left #cdk

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