Camelia, the Perl 6 bug

IRC log for #cdk, 2008-09-24

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

All times shown according to UTC.

Time Nick Message
06:11 carsten joined #cdk
06:55 rojasm joined #cdk
08:05 gilleain joined #cdk
08:18 egonw joined #cdk
08:48 CIA-14 cdk: egonw * r12453 /cdk/branches/cdk-1.2.x/src/main/org​/openscience/cdk/CDKConstants.java: Fixed JavaDoc: bond orders have been IBond.Order typed for a while now
08:49 CIA-14 cdk: egonw * r12454 /cdk/branches/cdk-1.2.x/src/main/org​/openscience/cdk/CDKConstants.java: The lazyCreation patch has been long default part of the CDK API
08:50 egonw OK, going to find a better place to sit...
08:50 egonw brb
08:55 egonw joined #cdk
09:29 gilleain joined #cdk
09:52 shk3 joined #cdk
09:54 egonw hi
09:54 zarah hi egonw
09:54 egonw hi shk3
09:54 egonw shk3: trying to merge cdk1.2.x into trunk
09:59 shk3 let me know when finished
10:00 shk3 i will then try the modelbuilder
10:00 egonw ok, issued the git svn dcommit now
10:18 CIA-14 cdk: egonw * r12456 /cdk/trunk/ (109 files in 49 dirs): Merged with branch cdk-1.2.x
10:19 egonw ha, that took a while...
10:22 gilleain 109 files updated...
10:24 egonw yeah, big merge :)
10:26 gilleain Hmmm. errors.
10:27 gilleain '"customs" cannot be resolved'...
10:29 egonw hang on
10:29 egonw planned to do a check, but forgot in the excitement ... :(
10:30 gilleain Also, the test for bug 956923 uses "assertEquals" instead of "Assert.assertEquals"
10:30 gilleain in SmilesGeneratorTest
10:30 egonw I'm on it
10:31 shk3 hm, seems to work for me
10:33 gilleain Well, NewCDKTestCase does not have an assertEquals(String, String) method.
10:33 gilleain or, I should say, my copy doesn't.
10:39 egonw and it should not have
10:39 egonw that test method sneaked in from a patch where the class was not yet updated to JUnit4
10:39 gilleain Ahh.
10:41 egonw not sure what this customs is about
10:41 gilleain Well, I remember when PMR was here at the same time as you, and he gave a mini-talk on CML.
10:42 gilleain He mentioned custom shemas or something?
10:42 gilleain *schemas
10:42 egonw right
10:42 egonw but this is something else
10:43 gilleain Ah, I meant conventions.
10:43 gilleain And, I suppose that means that is really something else.
10:44 egonw indeed :)
10:48 CIA-14 cdk: egonw * r12457 /cdk/trunk/src/ (2 files in 2 dirs): Compile fixed, resulting from the merge with cdk-1.2.x, which I forgot to fix before committing :(
10:53 shk3 egonw: all tests for modelbuilder3d work
10:54 shk3 that's suspicious
10:54 egonw hahahaha
10:54 egonw :)
10:54 egonw well, just add more of them then...
10:55 egonw you'll bound to hit a bug at some stage...
10:55 shk3 shouldn't rajarshi's changes be in trunk now?
10:55 egonw if that makes you feel better, tht is :)
10:55 egonw dunno...
10:55 egonw his changes to cdk1.2.x ?
10:55 egonw yes, those should be
10:56 shk3 hm, the regression tests for the fingerprints fail
10:57 shk3 i will recreate fingerprints and see what happens
10:57 shk3 just for curiosity
11:00 egonw ok
11:01 shk3 did you merge trunk to 1.2 as well?
11:08 egonw no
11:08 egonw shk3: ping me if you want certain patches to reach the branch
11:10 egonw I'll try to scan the commit messages of trunk to find interesting commits
11:11 shk3 i think the modelbuilder changes should go to 1.2
11:11 shk3 can also do this, if you don't have time
11:12 egonw pleaes give me the revision numbers, than I'll merge them
11:16 shk3 of the 10 test fingerprints, only 5 were changed by rajarshis changes
11:16 shk3 might be all the test for the modelbuider just used fragments, where the fingerprints were not affected
11:16 shk3 I will commit the new ones
11:17 egonw to trunk of cdk-1.2.x?
11:17 egonw you're working in trunk not?
11:18 CIA-14 cdk: shk3 * r12458 /cdk/trunk/src/ (2 files in 2 dirs): new fingerprints because of rajarshis changes
11:18 egonw so, I'll merge that one too...
11:19 egonw shk3: I'll browse the trunk commits now...
11:20 egonw and merge appropriate ones to cdk1.2.x
11:21 egonw anyway... great to hear that the builder3d is working again :)
11:23 egonw shk3: is commit 12309 relevant to the branch?
11:24 shk3 mom
11:25 egonw http://cdk.svn.sourceforge.net/view​vc/cdk?view=rev&revision=12309
11:29 shk3 11350, 12166, 12286, 12288, 12290, 12458 should do it
11:29 shk3 12309 was just a fix in the test
11:29 egonw ok, merging those commits
11:30 shk3 calling configureatoms was necessary, without that you get different canonical numbers the second time
11:31 shk3 looks like setting point2d destroys the configuration
11:31 egonw ok
11:32 egonw the oldest two seem already applied
11:33 shk3 might well be
11:37 egonw shk3: what's the file size of the fingerprint .gz ?
11:41 egonw that is...
11:41 egonw the .gz in trunk is only 4kB
11:41 egonw where it used to be some 120kB
11:41 egonw tar: This does not look like a tar archive
11:50 egonw well, will commit anyway
11:51 CIA-14 cdk: egonw * r12459 /cdk/branches/cdk-1.2.x/src/main/org​/openscience/cdk/modeling/builder3d/ (TemplateExtractor.java TemplateHandler3D.java): all tests for modelbuilder3d work. I will improve the fallback and add comments, but I want this working state to be committed
11:51 CIA-14 cdk: egonw * r12460 /cdk/branches/cdk-1.2.x/src/main/org/openscience​/cdk/modeling/builder3d/TemplateHandler3D.java: this has a fallback, better comments and is cleaned up
11:52 egonw I really wish git-svn would make these commit messages like: Applied bla from bla:
11:52 egonw because these commits are really shk3 work...
11:52 CIA-14 cdk: egonw * r12461 /cdk/branches/cdk-1.2.x/src/main/org/openscience​/cdk/modeling/builder3d/TemplateHandler3D.java: more cleanup
11:54 CIA-14 cdk: egonw * r12462 /cdk/branches/cdk-1.2.x/src/ (2 files in 2 dirs): new fingerprints because of rajarshis changes
12:01 egonw shk3: that completes the sync
12:01 egonw shk3: but I do think you uploaded a broken fingerprint database...
12:06 shk3 how so?
12:07 shk3 hm, test doesn't work
12:07 egonw because I can't inspect the tar.gz file with tar tzvf
12:07 shk3 looking at it
12:07 egonw and the file is about 120kB smaller
12:07 shk3 it's not a tar.gz
12:08 shk3 just a gziped txt
12:09 egonw what about the file size?
12:10 jbrefort joined #cdk
12:10 egonw shk3: ok, I think I see the problem
12:10 shk3 ah
12:10 egonw check the actual fingerprints
12:11 egonw the are almost all the same
12:11 egonw 10751 lines
12:11 egonw but only 7 unique lines
12:11 egonw which explains the enormous efficient gzipping :)
12:11 * shk3 is confused
12:12 shk3 why only 7 unique lines?
12:12 shk3 that can't be right
12:12 egonw something wrong with the FP generatino
12:12 egonw it seems that element symbol is not taken into account
12:13 shk3 no
12:13 shk3 that's part of the fallback
12:13 egonw the 7 FPs are for a 3 ring system, 4, 5, 6, etc
12:13 egonw $ cat src/main/org/openscience/cdk/modeling/bui​lder3d/data/ringTemplateFingerprints.txt | sort | uniq
12:13 egonw {148, 206, 392, 542, 637, 742, 752, 830}
12:13 egonw {148, 206, 392, 542, 637, 742, 752, 830, 906}
12:13 egonw {148, 392, 542, 637, 742, 752, 830}
12:13 egonw {148, 542, 637, 742}
12:13 egonw {148, 542, 637, 742, 752}
12:13 egonw {148, 542, 637, 742, 752, 830}
12:13 egonw {542, 637, 742}
12:13 egonw each longer FP just adds an integer
12:13 egonw so just a path of one node longer
12:14 shk3 hm, then the fingeprinting shouldn't help too much with filtering
12:14 egonw :)
12:14 egonw this is new
12:14 egonw because the previous file was much larger
12:14 shk3 yes and no
12:14 egonw indicating a high diversity in FPs
12:15 shk3 it was not used in the fingerprints,
12:15 shk3 but was intended to be used in the search
12:15 shk3 the fallback to atom symbol unspecific search, I mean
12:15 shk3 which was one of the reasons for problems
12:36 CIA-14 cdk: egonw * r12463 /cdk/branches/cdk-1.2.x/ (3 files in 3 dirs): Ignore autogenerated files in git too
13:05 CIA-14 cdk: egonw * r12464 /cdk/branches/egonw-cdkcmlcdk/: Created a branch for developing a test framework to back up CDK-CML-CDK roundtripping for all cdk.interfaces.
13:10 carsten joined #cdk
13:12 CIA-14 cdk: egonw * r12465 /cdk/branches/egonw-cdkcmlcdk/src​/test/org/openscience/cdk/io/cml/ (3 files): Merge branch 'addTestCDKCMLCDKRoundTripping' into my-cdkcmlcdk
13:22 shk3 joined #cdk
14:23 CIA-14 cdk: shk3 * r12466 /cdk/trunk/src/main/org/openscience/cdk/m​odeling/builder3d/TemplateHandler3D.java: fixes in modelbuilder
14:31 CIA-14 cdk: shk3 * r12467 /cdk/branches/cdk-1.2.x/src/main/org/openscience​/cdk/modeling/builder3d/TemplateHandler3D.java: fixes in modelbuilder
14:31 CIA-14 cdk: shk3 * r12468 /cdk/trunk/src/main/org/openscience/cdk/m​odeling/builder3d/TemplateHandler3D.java: fixes in modelbuilder
14:33 CIA-14 cdk: rajarshi * r12469 /cdk/branches/cdk-1.2.x/build.xml: Added some text to the info target to list useful targets. Also pdated some properties so that the output is more legible
14:37 jonalv joined #cdk
14:40 rojasm joined #cdk
16:06 rojasm joined #cdk
17:38 edrin joined #cdk
18:30 CIA-14 cdk: rajarshi * r12471 /cdk/branches/cdk-1.2.x/src/test/org/opensci​ence/cdk/graph/ConnectivityCheckerTest.java: Test cases to check for bug 2126904, in which the checker indicates the molecule is not connected when read from HIN but works OK when read from SMILES
18:51 CIA-14 cdk: rajarshi * r12472 /cdk/branches/cdk-1.2.x/src/main/org/open​science/cdk/renderer/Java2DRenderer.java: Commented out a lot of print statements, which mess up other applications using this class
18:58 CIA-14 cdk: rajarshi * r12473 /cdk/branches/cdk-1.2.x/src/main/org/opens​cience/cdk/graph/ConnectivityChecker.java: Updated to use 1.5 idioms and iterables
19:11 bag joined #cdk
19:12 jbrefort joined #cdk
20:42 carsten joined #cdk

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