Camelia, the Perl 6 bug

IRC log for #cdk, 2010-08-11

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

All times shown according to UTC.

Time Nick Message
03:02 azeem_ joined #cdk
04:00 azeem joined #cdk
05:57 jbrefort joined #cdk
06:07 egonw joined #cdk
07:01 Gpox joined #cdk
07:25 Gpox joined #cdk
07:28 egonw joined #cdk
07:33 egonw_ joined #cdk
08:06 s9asad joined #cdk
10:20 egonw azeem: did it work out with packaging Jmol in the end?
11:02 jonalv joined #cdk
11:02 jonalv is SMILES file a recognized IChemFormat in cdk?
11:26 egonw no
11:26 egonw the SMILES files are too hard to detect
11:26 jonalv that's sad
11:26 * egonw is assuming jonalv is refering to the format detction factories
11:27 jonalv egonw: yes
11:28 egonw you just need to specify behavior here manually?
11:33 jonalv egonw: sry?
11:33 jonalv egonw: what do you mean?
11:35 egonw if no format is detected, you need to manually check if it is a .smi file
11:35 egonw SMILESFormat.match() should work, I think...
11:35 egonw will check when eclipse is booted
12:26 s9asad joined #cdk
12:35 s9asad Hi all
12:35 zarah ni hao s9asad
12:45 s9asad I have a quick question
12:45 egonw how are you running the tests?
12:45 s9asad ant dist-all test-all
12:45 s9asad I get errors in this mode
12:46 s9asad but when I run via netbeans its error free (does not complain about dependency )
12:46 s9asad example src/test-smsd/org/openscience/cdk/smsd​/algorithm/rgraph/CDKMCSTest.java:57: cannot find symbol
12:46 s9asad [javac] symbol  : class TimeManager
12:46 s9asad [javac] location: package org.openscience.cdk.smsd.tools
12:46 s9asad [javac] import org.openscience.cdk.smsd.tools.TimeManager;
12:46 s9asad [javac]                                      ^
12:47 egonw are your patches online?
12:48 s9asad yes
12:48 s9asad I am using the code from http://github.com/asad/cdk-smsd
12:48 zarah s9asad's link is also http://tinyurl.com/2fn8ej9
12:50 egonw in which branch?
12:50 s9asad master
12:51 s9asad is it because smsd is not mentioned here http://github.com/asad/cdk-smsd/​blob/master/META-INF/MANIFEST.MF
12:51 zarah s9asad's link is also http://tinyurl.com/39d9myy
12:51 s9asad you might like to clone this git://github.com/asad/cdk-smsd.git
12:51 egonw don't think so
12:51 egonw that MANIFEST.MF is used by Eclipse
12:52 s9asad oh ic
12:53 s9asad something wrong here http://github.com/asad/cdk-smsd/blob/ma​ster/src/META-INF/test-smsd.libdepends?
12:53 zarah s9asad's link is also http://tinyurl.com/34txzs8
12:53 s9asad or here http://github.com/asad/cdk-smsd/blob/ma​ster/src/META-INF/test-smsd.cdkdepends
12:53 zarah s9asad's link is also http://tinyurl.com/2vrktfu
12:54 egonw those look fine to me
12:54 egonw no, don't spot the error right now
12:54 egonw but don't have time to test your patch right now
12:54 egonw but it it something like that...
12:54 egonw see if the cdk-smsd.jar is present
12:54 egonw and use jar tvf dist/jar/cdk-smsd.jar to see if the TimeManager is in it
12:55 s9asad okay will do it right away
12:55 s9asad by the way thanks for the initail review
12:55 s9asad initial *
12:56 s9asad I have submitted the patches with changes
12:56 egonw yes, I saw the email
12:59 s9asad jar tvf dist/jar/cdk-smsd.jar
12:59 s9asad 0 Wed Aug 11 13:49:02 BST 2010 META-INF/
12:59 s9asad 95 Wed Aug 11 13:49:00 BST 2010 META-INF/MANIFEST.MF
12:59 s9asad 56 Wed Aug 11 13:48:18 BST 2010 smsd.javafiles
12:59 egonw well, there's your problem then
13:00 s9asad aha
13:00 s9asad so the solution...
13:05 s9asad do I need to fix the build.xml?
13:07 egonw no, just rebuild the jars
13:07 egonw ant clean dist-all
13:10 s9asad just did that and I got this
13:10 s9asad Total time: 47 seconds
13:10 s9asad asad:cdksmsdgithub Asad$ jar tvf dist/jar/cdk-smsd.jar
13:10 s9asad 0 Wed Aug 11 14:10:10 BST 2010 META-INF/
13:10 s9asad 95 Wed Aug 11 14:10:08 BST 2010 META-INF/MANIFEST.MF
13:10 s9asad 56 Wed Aug 11 14:09:30 BST 2010 smsd.javafiles
13:21 egonw what's the content of build/smsd.javafiles ?
13:29 egonw s9asad: ?
13:29 s9asad I am compiling is again
13:29 s9asad how to check that?
13:30 egonw a text editor?
13:31 s9asad org/openscience/cdk/smsd/alg​orithm/vflib/Molecules.java
13:31 s9asad thats all
13:31 egonw huh?
13:32 egonw s9asad: please explain what you mean... you lost me
13:33 s9asad more build/smsd.javafiles
13:33 s9asad org/openscience/cdk/smsd/alg​orithm/vflib/Molecules.java
13:33 s9asad this is what I get
13:34 s9asad ore src/META-INF/smsd.libdepends
13:34 s9asad vecmath1.2-1.14.jar
13:34 s9asad commons-cli-1.0.jar
13:34 s9asad signatures-1.0.jar
13:34 s9asad and
13:34 s9asad more src/META-INF/smsd.cdkdepends
13:34 s9asad cdk-annotation.jar
13:34 s9asad cdk-interfaces.jar
13:34 s9asad cdk-core.jar
13:34 s9asad cdk-standard.jar
13:34 s9asad cdk-data.jar
13:34 s9asad cdk-io.jar
13:34 s9asad cdk-sdg.jar
13:34 s9asad cdk-smiles.jar
13:34 s9asad cdk-valencycheck.jar
13:34 s9asad cdk-signature.jar
13:35 s9asad so thats all
13:35 s9asad now I did ant clean dist-all dist-large
13:35 egonw ah...
13:35 s9asad more build/smsd.javafiles
13:35 egonw well, there should be plenty more classes in the smsd.javafiles
13:35 s9asad indeed
13:35 s9asad but it not creating :-(
13:35 egonw try this:
13:36 egonw ant clean runDoclet
13:36 egonw and check the file content again
13:36 egonw is there is still one file, then your patches managed to break something seriously
13:36 s9asad more build/smsd.javafiles
13:36 s9asad org/openscience/cdk/smsd/alg​orithm/vflib/Molecules.java
13:37 _ilbot2 joined #cdk
13:37 Topic for #cdkis now The Chemistry Development Kit (http://cdk.sf.net) | JUnit CDK 1.2.6: 15673, 76 fails + 15 errors (0.57%) | Logs at http://moritz.faui2k3.org/​irclog/out.pl?channel=cdk | CDK News Feed: http://friendfeed.com/rooms​/chemistry-development-kit
13:37 s9asad on the other hand with netbeans its working
13:37 egonw sure
13:37 egonw netbeans does not use the CDK modules concept
13:38 s9asad how can I trace, the bug!
13:38 s9asad is there a way to compile just the SMSD
13:39 s9asad and check if the job is done
13:39 egonw that's no use
13:39 egonw the code compiles fine
13:39 s9asad yes
13:39 egonw you need to look at the runDoclet taks
13:39 egonw task
13:39 s9asad okay
13:39 egonw and check why it fails on your classes
13:40 egonw that is
13:40 egonw why it does not put them in the build/smsd.javafiles
13:40 s9asad sorry i am naive to this but where is this runDoclet file?
13:41 egonw it's a task in the Ant build.xml
13:41 egonw as in: ant clean runDoclet
13:41 s9asad ok
13:44 egonw s9asad: btw, please review Jules patch for SMSD
13:44 egonw s9asad: or is that already included in the new SMSD 1.02 patch?
13:44 s9asad I have added his changes in the 1.02
13:45 egonw ok, can you update patch #3037820 please then?
13:45 s9asad should I wait until my patches are accepted?
13:46 egonw you could just mention that his patch is now part of the patches in your report
13:47 s9asad Sorry I have not review this patch but another one where he has added finally file,close
13:48 egonw ok, please review this one too
13:52 egonw please do check regularly if there are bugs and patches related to your work...
13:53 egonw that's part of maintaining a CDK module
13:53 egonw there are also still 3 bugs related to the SMSD code...
13:53 egonw please check those too, and respond, fix, whatever else you think is suitable :)
13:54 egonw sorry... make that 2
13:54 egonw one was closed
13:54 * egonw hates the SF default to show also Closed bugs :(
13:56 s9asad sure, will do that
13:56 s9asad but presently how to solve this build error!
13:58 egonw git bisect
13:58 egonw you know that it works fine in master
13:58 egonw but not with your latest version
13:58 egonw so, one of the patches in between made it broken
13:58 egonw that's where 'git bisect' comes in
13:59 egonw http://chem-bla-ics.blogspot.com/2008/11/f​inding-commit-that-causes-regressions.html
13:59 zarah egonw's link is also http://tinyurl.com/683mhn
14:05 s9asad okay
14:05 s9asad how to accept a patch on SF?
14:14 s9asad so this is what I did
14:14 s9asad git bisect start
14:14 s9asad git bisect bad
14:14 s9asad git bisect good upstream/master
14:14 s9asad ant clean dist-all test-dist-all jarTestdata
14:14 s9asad ant -Dmodule=smsd test-module
14:15 s9asad which gave me
14:15 s9asad test-module:
14:15 s9asad [echo] Testing classes for CDK's smsd module.
14:15 s9asad [echo] CDK dependencies defined: true
14:15 s9asad [echo] Library dependencies defined: true
14:15 s9asad [echo] Developer Library dependencies defined: true
14:15 s9asad [junit] Running org.openscience.cdk.modulesuites.MsmsdTests
14:15 s9asad [junit] Tests run: 1, Failures: 0, Errors: 1, Time elapsed: 0 sec
14:15 s9asad [junit] Test org.openscience.cdk.modulesuites.MsmsdTests FAILED
14:15 s9asad git bisect bad
14:15 s9asad Bisecting: 2 revisions left to test after this (roughly 1 step)
14:15 s9asad error: You have local changes to 'build.xml'; cannot switch branches.
14:16 s9asad git bisect good
14:16 s9asad e0c938bf1349426c4a89ff4a16c9e029b9b9703d was both good and bad
14:16 s9asad so now
14:16 s9asad what can we do
14:18 egonw you have made changes to the build.xml while doing the bisecting...
14:18 egonw check them with: git diff
14:18 egonw and/or: git diff --cached
14:18 egonw (depending on whether those changes are staged or not)
14:19 egonw and probably remove those changes
14:19 s9asad oh
14:23 egonw or: git stash
14:52 s9asad Hi
14:52 zarah hi s9asad
14:53 s9asad git bisect bad
14:53 s9asad Bisecting: 2 revisions left to test after this (roughly 1 step)
14:53 s9asad [7f3d314d70564029b6ef67f17a1f3f6811df9929] Added new SMSD code a) This code comes with matchers (bond, atoms) for each algorithm b) SMSD.java was renamed as Isomorphism.java c) init interface was improved to add hydrogens, clean molecule etc. d) Molhandler was improved e) Java doc and code suggestions by Egon introduced. f) bond match flags adjusted with matcher g) Major update speeds up VF lib in the SMSD h) Ability to match chem moles with chem knowledge
14:53 s9asad improved, although     Atom typing need to improve for some mols j) Chemfilters score changed k) Gilleain's Signature based labelling code add to make canonical atomcontainer for each mol before subgraph search (no sure if it improves speed or not!). l) VF lib substructure to return more matches, not just first map m) Over all speed omptimization n) SMSD algorithm rolled back to the one in the SMSD publication. o) code cleaning and optimization
14:53 s9asad asad:cdksmsdgithub Asad$ git bisect good
14:53 s9asad Bisecting: 0 revisions left to test after this (roughly 1 step)
14:53 s9asad [20b878b578d4969f3d1b010684b39a392a420264] cyclic dependency resolved in the build.xml as SMSD now requires Signature.jar
14:53 s9asad any thoughts
14:58 egonw fix the patch?
15:00 s9asad 20b878b578d4969f3d1b010684b39a392a420264] cyclic dependency resolved in the build.xml as SMSD now requires Signature.jar
15:00 s9asad ?
15:01 egonw if that patch broke the compiling, yes
15:04 s9asad hmm, let me try
15:22 s9asad is license type necessary to compile ?
15:23 egonw license type?
15:30 egonw_ joined #cdk
15:39 s9asad compile-module:
15:39 s9asad [echo] Compiling classes for CDK's smsd module from src/main...
15:39 s9asad [echo] Datafiles defined: ${module.datafiles.present}
15:39 s9asad [echo] Extra files defined: ${module.extrafiles.present}
15:39 s9asad [echo] CDK dependencies defined: true
15:39 s9asad [echo] Library dependencies defined: true
15:39 s9asad [echo] Developer Library dependencies defined: ${module.devellibdepends.present}
15:39 s9asad [echo] Autogenerated files specified: ${module.autogenerated.present}
15:39 s9asad [mkdir] Created dir: /Users/Asad/Software/GITROOT/​cdksmsdgithub/build/src/smsd
15:39 s9asad [mkdir] Created dir: /Users/Asad/Software/GITROO​T/cdksmsdgithub/build/smsd
15:39 s9asad [jar] Building jar: /Users/Asad/Software/GITROOT/cdk​smsdgithub/dist/jar/cdk-smsd.jar
15:39 s9asad [jar] Building MANIFEST-only jar: /Users/Asad/Software/GITROOT/cdksmsd​github/dist/jar/cdk-smsd-sources.jar
15:39 s9asad why is is just compiling
15:39 s9asad ..
15:39 s9asad ...
15:39 s9asad [jar] Building MANIFEST-only jar: /Users/Asad/Software/GITROOT/cdksmsd​github/dist/jar/cdk-smsd-sources.jar
15:40 egonw_ s9asad: I'll be back later... dinner first now
15:40 egonw_ s9asad: keep in mind if there are no files in build/smsd.javafiles ...
15:40 egonw_ then the compiling will not be a problem
15:40 egonw_ so, when bisecting, look at the smsd.javafiles content
15:40 egonw_ bbl
15:41 s9asad okies
16:14 s9asad joined #cdk
16:24 egonw_ joined #cdk
16:25 s9asad joined #cdk
16:25 s9asad @ egon fixed the error and updated the patches on the SF
16:26 egonw_ excellent... well done
21:46 bag_ joined #cdk

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