Camelia, the Perl 6 bug

IRC log for #cdk, 2008-09-08

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

All times shown according to UTC.

Time Nick Message
02:50 CIA-14 cdk: rajarshi * r12220 /cdk/trunk/tools/nightly.py: Added code to include test coverage counts. Also updates to improve robustness
03:13 CIA-14 cdk: rajarshi * r12221 /cdk/trunk/tools/nightly.py: More robustness. Added a cli keyword to indicate whether th script is being run for the first time or not
04:00 CIA-14 cdk: rajarshi * r12222 /cdk/trunk/tools/nightly.py: More robustness checks
04:02 CIA-14 cdk: rajarshi * r12223 /cdk/trunk/tools/nightly.py: More robustness checks
04:09 CIA-14 cdk: rajarshi * r12224 /cdk/trunk/tools/nightly.py: Fixed usage doc
04:11 CIA-14 cdk: rajarshi * r12225 /cdk/trunk/tools/nightly.py: Aargh. Fixed a bug in a loop. Should go to sleep
05:04 bag joined #cdk
05:07 CIA-14 cdk: rajarshi * r12226 /cdk/branches/cdk-1.2.x/src/ (39 files in 6 dirs): Updated and added copyright notices to the appropriate classes
05:11 CIA-14 cdk: rajarshi * r12227 /cdk/branches/cdk-1.2.x/src/main/org/openscience/​cdk/isomorphism/UniversalIsomorphismTester.java: Fixed the return value in getIsomorphAtomsMap. Fixes bug #2028685
05:43 jbrefort joined #cdk
07:26 malick joined #cdk
07:58 jonalv joined #cdk
08:03 maclean joined #cdk
08:04 shk3 joined #cdk
08:25 thomas_ku joined #cdk
08:29 thomas_ku moin
08:29 zarah hi thomas_ku
08:54 maclean Would it be a bad thing to add a no-argument constructor to the SmilesParser that made the call to DefaultChemObjectBuilder.getInstance() itself?
08:54 maclean So that you could just say "new SmilesParser()"...
08:55 thomas_ku I would think that this would be ok
08:56 maclean I started writing it as a feature-request, but it's kind of tiny :)
08:56 maclean Oh, and I tried to get a patched version of BSD java6 working on my mac, to run CDK-Taverna, but I couldn't get it to work.
08:57 thomas_ku :-(
08:58 thomas_ku It is really stupit that Apple do not support java 6 for all osx
08:58 maclean I still find it weird that it won't run the plugin under java5
08:59 maclean Apple's java always lags.
08:59 maclean I mean, the development.
08:59 thomas_ku that because some algorithm within the plugin needs java 6 and it is compile with java 6 so would say that it will only works with java 6
08:59 maclean Ahh. I see.
09:01 thomas_ku the main problem is that it is more or less one algorithm which uses java 6 for some seralization and I did not have the time to do some rewriting of this code to support java 5
09:01 maclean What aspect of java6 does it use?
09:03 thomas_ku the xmlstreamreader and writer mainly
09:05 maclean Ahh.
09:09 CIA-14 cdk: shk3 * r12228 /cdk/trunk/src/test/ (2 files in 2 dirs): added a test for bug #1930029
09:25 maclean Hmmm. RingSetTest is broken. there's no size() method in IRingSet.
09:26 CIA-14 cdk: gilleain * r12229 /cdk/trunk/src/main/org/openscien​ce/cdk/smiles/SmilesParser.java: Convenience constructor added.
09:36 CIA-14 cdk: shk3 * r12230 /cdk/trunk/src/main/org/openscien​ce/cdk/io/cml/CMLCoreModule.java: fixes bug #1930029
10:14 egonw joined #cdk
10:16 CIA-14 cdk: sithmein * r12231 /cdk/branches/cdk-1.2.x-KNIME/: changes necessary or useful for the KNIME integration
10:19 CIA-14 cdk: gilleain * r12232 /cdk/trunk/src/main/org/openscien​ce/cdk/smiles/SmilesParser.java: reverted.
10:24 egonw maclean: it's good for commit statistics :)
10:26 egonw so, a central API to check module deps etc
10:26 maclean I see.
10:27 egonw and list all java packages in a module (which I need to the eclipse plugins MANIFEST.MF Export: bits
10:27 maclean Is this for management tools?
10:27 egonw and for creating things like:
10:27 maclean Or for run-time pluggin of modules?
10:27 egonw org.openscience.cdk.interfaces
10:27 egonw org.openscience.cdk.core
10:27 maclean *plugging
10:27 egonw org.openscience.cdk.standard
10:27 egonw org.openscience.cdk.extra
10:27 egonw org.openscience.cdk.builder3d
10:27 egonw etc, etc
10:27 egonw yeah, run-time
10:28 egonw source will just be what we have right now...
10:28 maclean I see.
10:28 egonw so, other projects (viz Bioclipse2) can depend on particular CDK plugins
10:28 maclean that's cool.
10:28 egonw likewise... POM, for Jumbo and CDK-Taverna
10:28 maclean CDK's a bit too big otherwise.
10:28 egonw also autogenerated
10:28 egonw maclean: yes
10:29 egonw 15MB is nothing compared to nowadays software, really... but for frequent updating...
10:29 maclean Sure
10:30 maclean Did you see the monster package diagram, btw?
10:30 egonw yes, I think I did...
10:30 egonw did you see the diagrams on Nightly?
10:30 maclean yeees
10:30 egonw made with graphviz
10:30 maclean they don't seem to work on my mac
10:30 egonw ah :(
10:30 maclean Prefiew only shows me part of them
10:30 egonw file a bug report
10:30 maclean *Preview
10:30 maclean OK
10:31 egonw always file bug reports
10:31 egonw otherwise it gets lost...
10:31 * maclean is finding that lots of things don't work on macs...
10:31 maclean git didn't build for example
10:31 egonw and it gives a good overview of what remains to be done, and where people might jump in to help etc
10:31 maclean sure, sure.
10:31 shk3 hi egonw
10:33 egonw hi shk3
10:33 shk3 I saw you started a cdk 1.2 branc h
10:33 shk3 will things get ported from/to there?
10:33 egonw shk3: the thing with the fingerprints is... that the depend not just on the algorithms, but also on the input...
10:33 egonw shk3: yes
10:33 egonw no worries about your patches to trunk
10:33 shk3 i just commited to trunk
10:33 shk3 ah ok
10:33 egonw I'll apply them to cdk1.2.x
10:34 shk3 great
10:34 egonw git rulez!
10:34 shk3 don't understand this
10:34 shk3 sure, fingerprints depend on input
10:34 shk3 different structures give different fingerpritns
10:35 egonw right, but I mean something else...
10:35 shk3 they should at least
10:35 egonw I mean, that is now also depends on aromaticity...
10:35 egonw which is relatively subjective...
10:35 shk3 hm, aromaticity detection is currently called in fingerprinter
10:36 shk3 so as long as the algorithm doesn't change
10:36 egonw and with every bug fix or just more tuned atom type, these things might slightly vary
10:36 shk3 it should be all right
10:36 shk3 yes, buth this is the algorithm
10:36 egonw which would explain why 'we' had to regenerate the index all the time
10:36 shk3 so fingerprinter depends an aromaticity algorithm, yes
10:37 egonw which really makes sense
10:37 egonw as it is an indicator for flat ring systems...
10:37 shk3 the input file the ring templates has aromaticities in it
10:37 shk3 i. e. bond order 4
10:38 shk3 so we could take out detection for this
10:38 shk3 then just the user needs to make sure  he gets aromaticity right
10:40 egonw I need to think about this, but I think we can suffice with detecting planar rings...
10:40 egonw and say, forget aromaticity, let's consider any ring system with SP2 + PLANAR3 only atoms as planar
10:41 egonw ok, got to work on our house again...
10:41 shk3 ok
10:41 shk3 I would like to fix the last failing test for model builder, but this is connected to aromaticity, it seems.
10:42 egonw yeah, no worries...
10:42 shk3 so please discuss your conclusions with me
10:42 egonw on every chance in atom type perception, I'll make sure to check the tests for builder3d
10:43 egonw and when tests fail, we know we have to regenerate the index...
10:43 egonw not optimal, but certainly workable
10:43 shk3 and those for template handler as well, please
10:44 egonw might you write up your findings? maybe as CDK News write up?
10:45 egonw and what the exact steps need to be done if tests fail again?
10:47 shk3 not much to say about this, I am afraid...
10:49 CIA-14 cdk: egonw * r12233 /cdk/branches/cdk-1.2.x/src/test/ (2 files in 2 dirs): added a test for bug #1930029
10:49 CIA-14 cdk: egonw * r12234 /cdk/branches/cdk-1.2.x/src/main/org/ope​nscience/cdk/io/cml/CMLCoreModule.java: fixes bug #1930029
10:51 egonw this is what I typed to acchieve this:
10:51 egonw git cherry-pick 1040ed4e47b04d7a93e4098d54933a34edc5cacb
10:51 egonw git cherry-pick 8b96b230b33ebbbd3759ec953f18a5f648bd7b33
10:51 egonw git svn dcommit
10:51 egonw (git uses different commit identifiers...)
10:52 egonw merging is fun again :)
10:52 maclean hmm yes. I tried out that git stuff, but it took 1hr to download the first 1,000 revisions of 12,000 so I gave up.
10:52 egonw maclean: yes, I did mention the overnight(s), not?
10:52 maclean you did.
10:53 maclean Is it worthwhile to make a repository from a later revision than r1?
10:53 egonw yes, I think it is
10:53 maclean perhaps I'll do that instead then.
10:53 egonw revision 10000 would be a reasonable one
10:53 maclean Ok.
10:54 egonw btw, just make one for the EBI, then just copy from each other
10:54 egonw so, that not all of you have to wait X nights :)
10:54 maclean Hmmm.
10:54 maclean That reminds me.
10:54 maclean I should ask support about the group space.
11:01 egonw bye for now
11:23 jpansane1 moin
11:23 zarah hi jpansane1
12:04 jonalv_ joined #cdk
12:26 malick joined #cdk
13:50 CIA-14 cdk: sithmein * r12235 /cdk/branches/cdk-1.2.x-KNIME/build.xml: built jar files contain sources (to make debugging easier)
13:50 CIA-14 cdk: sithmein * r12236 /cdk/branches/cdk-1.2.x-KNIME/src/main/org/o​penscience/cdk/renderer/Java2DRenderer.java: improvements in the 2D renderer
14:32 malick left #cdk
15:03 shk3 note I just closed the oldest open cdk bug
15:07 thomas_ku shk3: great
15:08 jonalv shk3++
15:53 jbrefort joined #cdk
16:22 carsten joined #cdk
17:16 CIA-14 cdk: shk3 * r12237 /cdk/trunk/src/test/org/openscience​/cdk/smiles/SmilesParserTest.java: added a test for bug #1274464 adn #1963731
19:53 bag joined #cdk
23:07 CIA-14 cdk: rajarshi * r12238 /cdk/branches/cdk-1.2.x/src/main/​org/openscience/cdk/RingSet.java: Removed redundant casts
23:36 CIA-14 cdk: rajarshi * r12239 /cdk/branches/cdk-1.2.x/src/ (3 files in 3 dirs):
23:36 CIA-14 cdk: Added a method to IRingSet that allows you to add a single ring to the set.
23:36 CIA-14 cdk: Implemented the method in RingSet and updated the test to properly test for
23:36 CIA-14 cdk: clone, rather than just assertTrue (which was not really a test). Fixes bug
23:36 CIA-14 cdk: #2099919
23:41 CIA-14 cdk: rajarshi * r12240 /cdk/branches/cdk-1.2.x/src/main/​org/openscience/cdk/RingSet.java:
23:41 CIA-14 cdk: Updated the clone method since it was duplicating the rings. This is because
23:41 CIA-14 cdk: IRingSet is a child of IAtomContainerSet so super.clone() already clones all the
23:41 CIA-14 cdk: atom containers in the set. Thus when RingSet copied over all the atom
23:41 CIA-14 cdk: containers, we end up with duplicates.
23:44 CIA-14 cdk: rajarshi * r12241 /cdk/branches/cdk-1.2.x/src/ (3 files in 3 dirs): The commit that added a method to the IRingSet interface to add a single ring was unnecessary since we can use addAtomContainer from the super class to add a single ring. Removed the extraneous method
23:46 CIA-14 cdk: rajarshi * r12242 /cdk/branches/cdk-1.2.x/src/main/​org/openscience/cdk/RingSet.java: Updated Javadocs
23:49 CIA-14 cdk: rajarshi * r12243 /cdk/branches/cdk-1.2.x/src/main/org/op​enscience/cdk/interfaces/IRingSet.java: Updated Javadocs

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