Camelia, the Perl 6 bug

IRC log for #bioclipse, 2009-07-01

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

All times shown according to UTC.

Time Nick Message
05:52 olass joined #bioclipse
05:55 olass @seen egonw
05:55 zarah egonw was last seen 16 h 14 m 24 s ago in #bioclipse saying 'masak++'
06:07 edrin joined #bioclipse
06:07 edrin hi
06:07 zarah hi edrin
06:07 edrin hi olass
06:09 olass hi edrin
06:09 edrin olass: i copy pasted the letter to the google docs and marked things I already did
06:09 olass ok
06:43 Gpox joined #bioclipse
07:07 kmajoor joined #bioclipse
07:07 olass kmajoor: hi
07:07 olass kmajoor: have you seen egonw?
07:08 olass seems not
07:12 kmajoor joined #bioclipse
07:12 egonw moin
07:12 zarah privet egonw
07:12 * egonw is trying to wake up
07:13 egonw I guess what masak calls that massive wakeup fail
07:22 olass egonw: did you come to a conclusion about the aromaticity and hydrogens yesterday?
07:22 egonw conclusion, sort of I guess...
07:22 egonw solution, no
07:22 olass :(
07:23 egonw will continue today
07:23 olass ok
07:23 olass did you get my bugs?
07:23 egonw when I manage to think more clearly
07:23 olass ;)
07:23 egonw no
07:23 egonw not yet on my laptop
07:23 egonw but the computer with a working skype
07:24 olass I have different behaviour when drawing molecules from scratch than when saving the molecule and loading it back
07:24 egonw http://www.opentox.org/projects/show/toxtreeb
07:24 olass this is really confusing
07:24 zarah egonw's link is also http://tinyurl.com/mh7eku
07:24 egonw sounds like a nice project for bc2 plugins
07:24 egonw for your DS support
07:24 olass I know about it
07:24 egonw triggered by this report (sorry, it's in blog form):
07:24 egonw http://barryhardy.blogs.com/cheminfostr​eam/2009/06/overcoming-the-challenges-t​o-reducing-animal-experimentation-throu​gh-an-open-source-communitybased-d.html
07:25 zarah egonw's link is also http://tinyurl.com/m4d6gp
07:25 egonw yes, I think you met the ambit people
07:25 olass nina etc
07:25 olass yes
07:25 egonw ok, going to fetch breakfast
07:25 egonw and will boot laptop
07:25 olass ok
07:26 egonw I guess edrin left for work now?
07:26 egonw really sorry about that
07:26 olass yes, he left
07:27 olass we talked for an hour and got pretty far anyway
07:27 egonw brb
07:27 egonw ok, good
07:27 egonw will read the gdoc
07:36 egonw joined #bioclipse
08:03 olass egonw: let me know if I should explain 1382 better, maybe we can skype about it if needed
08:03 egonw no, seems clear enough
08:06 egonw olass: I need an executive call here...
08:06 egonw there are two options, two camps, if you like
08:06 olass ok...
08:06 egonw 1. follow file content, and show what is in the file, not what Bioclipse thinks is in the file
08:07 egonw 2. use the file content as starting point, and do some magic on it trying to understand possibly misinterpreting file content
08:07 olass I would suggest 2, since that is best for end users
08:08 olass or we could have a preference for this?
08:08 egonw no, I would not use 'best' in this context
08:08 egonw it's merely two options
08:08 olass right
08:08 egonw both will likely cause bug reports
08:08 olass I'd say 2, but I see the issue
08:09 olass we should document the reasoning on a wiki page
08:09 egonw yes, we should
08:09 olass and be open for the community reponses
08:09 egonw OK, I will add arom detection to MDL v2000 molfile reading
08:09 egonw but not to CML
08:09 olass ok
08:09 egonw the latter can serialize aromaticity props
08:09 egonw and should show up anyway
08:09 egonw MDL no...
08:10 olass about drwaing fom scratch, how handle aromaticity and hydrogens?
08:10 egonw that's the next step
08:10 olass ok
08:10 egonw each drawing edit triggers a reculatation of implicit hydrogens and aromaticity
08:11 egonw which requires atom typing...
08:11 olass yes
08:11 egonw I saw fishy things happening here
08:11 olass here too
08:11 egonw so, wrote a patch to show the atom type in the Outline
08:11 olass cool
08:11 olass did you see byg 1384?
08:11 egonw so that I can quickly expect when what goes wrong...
08:11 egonw while editing...
08:11 egonw and will try to figure out what is causing this trouble
08:11 olass great
08:12 egonw there is no reason it should fail...
08:12 olass egonw++ //seems like you are on top of things
08:12 egonw but that is merely theoretically
08:12 olass egonw: I want to show you some things tomorrow at bmc
08:12 egonw oh
08:12 olass regarding m2d
08:12 egonw I just realized I dreamed about the pele server !
08:12 egonw that it had extra memory
08:12 olass bad dream?
08:13 egonw which was there all along
08:13 olass hmm
08:13 egonw but only as option
08:13 olass lol
08:13 olass forgot to order again :(
08:13 egonw I really think we should add some memory... I guess :)
08:13 olass will do it now
08:13 olass yes
08:13 olass I'll email Magnus
08:13 olass we have 2Gb now?
08:13 olass right?
08:13 egonw yes
08:14 olass we should extend to 4Gb
08:14 olass ok, will CC you
08:15 egonw great, thanx
08:28 egonw ok, bond edits are not causing a change event
08:28 egonw that is one problem
08:29 egonw neither does changing an element symbol, it seems
08:30 Gpox egonw: structurePropertiesChanged()
08:31 egonw will verify
08:33 Gpox egonw: what should the new controller work on as model IChemModel, IAtomContainer?
08:34 egonw I think we should have two
08:34 egonw one controller for molecule only
08:34 egonw and a second one extending the first, which also does ChemModel's
08:35 Gpox what about partitioning and unconnected structures in an atomcontainer
08:36 egonw yes in ChemModel version
08:36 egonw no in AtomContainer version
08:36 Gpox so you should not be able to create a unconnected structure in molecule mode?
08:38 egonw no, unconnected stuff in the AtomContainer version is fine
08:39 egonw or is your 'molecule mode' != AtomContainer model
08:41 Gpox no they are the same
08:41 shk3 joined #bioclipse
09:15 CIA-52 bioclipse: egonw * r11566 /bioclipse2/trunk/plugins/net.bioclipse.cd​k.jchempaint/src/net/bioclipse/cdk/jchempa​int/outline/StructureContentProvider.java: Also output atom type info, to allow quicker diagnostics (implementes #1384).
09:16 sneumann_ joined #bioclipse
09:38 egonw olass: ping
09:48 olass egonw: pong
09:49 egonw I'm going to commit a rather intruisive clean up of cdk.loadMolecule()
09:49 egonw needed for further work
09:49 egonw no functional changes
09:49 egonw but code clean up
09:49 egonw tested on all relevant sample data
09:49 egonw looks fine
09:50 egonw wanted you to be aware of it
09:52 CIA-52 bioclipse: egonw * r11567 /bioclipse2/trunk/plugins/net.bioclipse.cdk.busine​ss/src/net/bioclipse/cdk/business/CDKManager.java:
09:52 CIA-52 bioclipse: Code clean up of cdk.loadMolecule().
09:52 CIA-52 bioclipse: * integrated different situations
09:52 CIA-52 bioclipse: * no longer reads all molecules in the file, and then only returns the first
09:52 CIA-52 bioclipse:  but instead just returns the first directly
09:52 CIA-52 bioclipse: * some more informative debug information to the logger
09:53 olass egonw: sounds good
09:54 olass will update and test
09:54 egonw yes, please do
09:54 olass but first: lunch
09:57 olass egonw: about 1382, is that resolved then?
09:58 olass or is this next step?
09:58 egonw next step
09:58 olass ok
10:05 egonw Gpox: the mol table is not showing a message in the editor window itself that it is indexing the file...
10:05 egonw is that intentional?
10:12 egonw rojas1, rojasm: ping
10:42 olass egonw: new loading mechanism looks fine to me
10:42 olass after you add logic to editing I will build RC5
10:43 olass (if hings go well)
10:43 masak joined #bioclipse
10:43 olass hi masak
10:43 masak olass: oh hai
10:45 Gpox egonw: no, it is just not implementd
10:48 egonw Gpox: ack, you already have a feature request for that?
10:48 Gpox don't think so
10:51 olass Gpox: bug 1386 is actually rather urgent and hopefully quite trivial
10:56 CIA-52 bioclipse: egonw * r11568 /bioclipse2/trunk/plugins/net.bioclipse.cdk.busine​ss/src/net/bioclipse/cdk/business/CDKManager.java: Moved code to add implicit hydrogens for MDL V2000 molfiles into separate method (cleaner code), and called from lower level manager method (more consistent behavior among loadMolecule() methods).
10:56 CIA-52 bioclipse: egonw * r11569 /bioclipse2/trunk/plugins/net.bioclipse.cdk.busine​ss/src/net/bioclipse/cdk/business/CDKManager.java: Also do aromaticity detection when MDL molfiles are loaded
11:00 CIA-52 bioclipse: goglepox * r11570 /bioclipse2/trunk/plugins/net.bioclipse.cd​k.jchempaint/src/net/bioclipse/cdk/jchempa​int/widgets/JChemPaintEditorWidget.java:
11:00 CIA-52 bioclipse: Added selection firing when model is first set
11:00 CIA-52 bioclipse: Solves bug #1386
11:00 zarah bug #1386 | http://tinyurl.com/mg3skl
11:10 egonw olass: ping
11:10 olass pong
11:12 egonw looking at 1382 now
11:12 egonw not sure what is going on there...
11:12 egonw teh CML counts 11 hydrogens but the prop view not...
11:13 olass yes, this is important to sort out
11:15 egonw OK, found the problem
11:15 egonw was disguisted...
11:15 egonw an atom typing problem, when starting from the CML
11:20 olass oh
11:20 olass can you solve it?
11:21 olass right now, problems with atom typing etc is not really warned to the user
11:21 olass which it should be IMO
11:21 egonw no, that problems is generally unsolvable...
11:21 egonw yes
11:21 egonw we have an Error framework scheduled for 2.1
11:22 egonw I need to figure out why the created CML is wrong
11:22 egonw might be a bug in the SMILES parser
11:22 olass oh
11:22 egonw or maybe the SMILES is bad
11:22 egonw dunno yet
11:22 olass I experience problems when drawing from scratch too
11:22 olass you could try that also
11:23 egonw drawing from scratch worked fine
11:25 olass ok, good
11:25 olass egonw: is there a content type for sdf0d?
11:26 olass an sdfile with no coordinates?
11:26 olass I can't seem to find it...
11:26 egonw check net.bioclipse.doc.developers
11:26 olass can't run that
11:26 olass some perl issue
11:26 olass xpath version etc
11:26 olass that's why I ask
11:27 egonw ok, so we have not resolved that yet either... :(
11:27 olass btw, would like to to generate both doc and doc.devel and commit that
11:27 olass would you like to do that?
11:27 olass since it works on your machine...
11:27 olass anyway
11:27 egonw crappy OS/X
11:27 olass an SDF0D contenttype is a blocker
11:28 olass if it does not exist
11:28 olass will file a bug, hopefully it is easy to set up
11:28 olass (it probably is)
11:33 olass Gpox: the .tmp file is now shown in the Navigator when saving an SDF after e.g. removal of properties. Is that intentional?
11:34 Gpox no, but it has probably always been like that
11:34 egonw Gpox: can it use File.getTempFile() ?
11:34 olass it probably needs a workspace IFile
11:35 olass Gpox: we could add a filter to hide all .tmp files
11:35 olass not critical though
11:35 olass Gpox: Bug 1388 is critical however
11:35 egonw :)
11:35 egonw damn
11:35 olass or at least much more important
11:35 egonw there are many critical bugs
11:36 olass yes, some
11:36 olass but not THAT many
11:36 olass and better we find them now than after 2.0 release
11:37 * egonw is writing a CDK unit test
11:54 egonw olass: no, there does not seem to be a SDF 0D content type
11:54 olass I filed a bug
11:54 egonw I can add it, but where should it be used?
11:54 egonw who should I ping for an update?
11:54 egonw Gpox: ?
11:54 olass I have a use case
11:54 egonw sure you do :)
11:54 olass update?
11:54 olass what do you mean?
11:55 olass I just need it implemented
11:55 olass or?
11:55 egonw ok
11:55 egonw so, I only need to ping you when done
11:56 olass well, I sort of worked around the problem
11:56 olass using the sdf content type
11:56 olass but IMO we should have an sdf0D
11:57 olass I don't mind if you make it a 2.1 bug for now since I was able to resolve an issue here
11:57 egonw being commited
11:57 olass oh :)
11:57 olass egonw++
11:57 CIA-52 bioclipse: egonw * r11571 /bioclipse2/trunk/plugins/ne​t.bioclipse.cml/plugin.xml: Fixed human readable string to be human readable
11:58 olass I can't open target.platform anymore, it crashes eclipse!
11:58 olass tried reinstalling it
11:58 olass same behaviour
11:58 CIA-52 bioclipse: egonw * r11572 /bioclipse2/trunk/plugins/net.bioc​lipse.chemoinformatics/plugin.xml: Added content type for MDL SD files without coordinates given in the file (fixes #1387).
11:58 zarah bug #1387 | http://tinyurl.com/kwl6kb
11:58 olass this means I cannot run tests anymore :(
11:58 masak nice to see that the human-readable string is human-readable!
11:58 egonw that sucks
11:58 olass yes
12:01 olass ok, was able to set it using preferences
12:02 olass *phew*
12:02 * olass wants Galileo
12:04 olass oh no
12:04 olass things fail
12:04 olass argh!
12:04 masak fail--
12:04 egonw what fails?
12:05 egonw more hardware trouble?
12:05 egonw I hope not... this is a different machine!
12:05 olass this is eclipse
12:09 olass checking out a new target.platform seemd to resolve some things...
12:10 olass right, I'm back in business
12:22 CIA-52 bioclipse: goglepox * r11573 /bioclipse2/trunk/plugins/net.bioclipse.cdk​.ui.sdfeditor/src/net/bioclipse/cdk/ui/sdfe​ditor/business/IMoleculeTableManager.java:
12:22 CIA-52 bioclipse: Changed return type to void
12:22 CIA-52 bioclipse: To get it to run as a job
12:22 CIA-52 bioclipse: Solves bug #1388
12:22 zarah bug #1388 | http://tinyurl.com/msn64r
12:23 CIA-52 bioclipse: goglepox * r11574 /bioclipse2/trunk/plugins/net.bioclipse.cdk​.ui.sdfeditor/src/net/bioclipse/cdk/ui/sdfe​ditor/business/MoleculeTableManager.java:
12:23 CIA-52 bioclipse: Call setHidden(true) on temp file when calculating properties for file.
12:23 CIA-52 bioclipse: It doesn't work
12:25 egonw olass: found the atom typing bug source
12:25 olass great!
12:38 edrin1 joined #bioclipse
12:38 edrin1 left #bioclipse
12:38 edrin joined #bioclipse
12:39 edrin hi
12:39 zarah hi edrin
12:40 masak hi edrin
12:41 edrin hi masak
12:45 steinbeck joined #bioclipse
12:49 sneumann steinbeck@guest.wtgc.org ? You're sitting in the cafeteria ;-)
13:04 egonw joined #bioclipse
13:14 CIA-52 bioclipse: egonw * r11575 /bioclipse2/trunk/cdk-externals/trunk/or​g.openscience.cdk.core/src/org/openscien​ce/cdk/atomtype/CDKAtomTypeMatcher.java: Applied fix for CDK bug #2815172 solving no atom type perception for certain nitrogens (fixes #1382)
13:14 zarah bug #2815172 | http://tinyurl.com/ny52b4
13:20 egonw Gpox: would it be much to get more pronounced visual feedback on atom merging?
13:35 egonw oh god
13:35 egonw this is bad
13:35 egonw olass: ping
13:36 egonw those other fixes did not fix the other major issue
13:36 egonw updating properties upon edits
13:36 egonw when I drop a cyclohexane and modify the bond orders
13:36 egonw ...
13:36 steinbeck sneumann: Sorry, only now saw your posting :-) You get that address anywhere on the campus and in the buildings when you are on the Wireless.
13:36 egonw the bonds change, but not the atom types...
13:46 egonw Gpox: ping
13:46 egonw still around?
13:47 Gpox egonw: pong
13:47 egonw changing bond orders is not using the manager
13:47 egonw fair
13:47 egonw which module is this happening?
13:47 egonw as it does not seem to trigger a changedPropertyEvent()
13:49 olass egonw: pong
13:49 Gpox Do yo have add bond/atom selected that it is in AddBondDragModule
13:49 egonw olass: see above
13:50 egonw Gpox: sorry, I don't understand your question
13:50 Gpox Toolbar button
13:51 Gpox Do you have the first button in the jcp toolbar selected?
13:51 egonw yes, the "Add Bond" toolbar icon
13:51 egonw ok
13:51 Gpox then it is in AddBondDragModule
13:52 Gpox I have rewritten that module for the new controller structure so i don't have the code
13:53 egonw :)
13:56 egonw mmm...
13:57 olass Gpox: is there a unit test somewhere for moltable reading/parsing?
13:57 olass I'd like to extend it if exists
13:57 Gpox no there isn't
13:58 olass ok
14:00 olass time for one now
14:02 olass will add one
14:03 olass Gpox: should I create a new test-plugin for this or put it some other place?
14:03 olass I mean, things will be refactored anyway...
14:04 olass egonw: can I write the test in cdk.business.test and add deps to cdk.sdfeditor?
14:04 * olass don't want a new plugin for just one test that will be deprecated in 1 month
14:04 Gpox go ahead and create it
14:04 olass where?
14:04 olass the plugin you mean?
14:04 Gpox yes
14:05 olass ok then
14:05 Gpox It will just be moved
14:05 egonw we don't release a Test Feature yet
14:05 olass why should we?
14:05 olass I see no use case
14:07 egonw a lot of JCP edits are not causing changes in the Outline...
14:07 egonw no fucking clue what is going on here...
14:07 egonw the change events seem to happen
14:07 egonw but the outline is updated for some things, like new atoms, but not for changes in atom type... :(
14:21 sneumann_ joined #bioclipse
14:37 CIA-52 bioclipse: egonw * r11576 /bioclipse2/trunk/cdk-externals/trunk/or​g.openscience.cdk.control/src/org/opensc​ience/cdk/controller/ControllerHub.java: The connectivity does not change upon a bond order change; therefore, send a structurePropertyChange event instead
14:38 CIA-52 bioclipse: egonw * r11577 /bioclipse2/trunk/plugins/net.bioclipse.cd​k.jchempaint/src/net/bioclipse/cdk/jchempa​int/widgets/JChemPaintEditorWidget.java: Percieve aromaticity upon edits (fixes #1380)
14:38 zarah bug #1380 | http://tinyurl.com/ly8ddm
14:43 olass egonw: how are things going? You seem to have progres...
14:44 egonw no, stuck
14:44 egonw see bug report I just created
14:44 olass for me?
14:44 CIA-52 bioclipse: ospjuth * r11578 /bioclipse2/trunk/plugins/net.bi​oclipse.cdk.ui.sdfeditor.tests/: Initial import.
14:44 egonw for arvid, cc-ed you
14:44 egonw @pz outline
14:44 olass ok
14:44 zarah [1391] Atom type info in Outline does... http://tinyurl.com/m365mk
14:44 zarah [1333] 70 source files without copyri... http://tinyurl.com/nz9aff
14:44 zarah [888] Jmol outlineshould not be expan... http://tinyurl.com/lrbcta
14:44 zarah [330] Outline for MoleculesTable         http://tinyurl.com/l3fj4g
14:44 zarah [1324] Context menu for outline          http://tinyurl.com/l7m5qt
14:44 zarah Entire list (13) at http://tinyurl.com/kvrdxj
14:44 egonw @pz 1391
14:45 zarah http://pele.farmbio.uu.se/cgi-bi​n/bugzilla/show_bug.cgi?id=1391
14:45 CIA-52 bioclipse: ospjuth * r11579 /bioclipse2/trunk/plugins/net.bi​oclipse.cdk.ui.sdfeditor.tests/ (18 files in 11 dirs): Added test for reading and parsing MolTable.
14:58 olass egonw: Can you do this exercise?
14:58 olass 1) new emptymolecule
14:58 olass 2 place benzene, note outline has atom type
14:58 olass 3) save to cml
14:58 olass close editopr
14:58 olass open saved cml file
14:59 olass note outline has no atom types
14:59 olass egonw: sjould I file a bug?
14:59 steinbec1 joined #bioclipse
14:59 olass or is this in the bug report?
15:03 egonw bug
15:03 egonw JCP editor should save atom type info
15:04 egonw fuck...
15:07 olass ok, filed as bug 1393
15:07 egonw thanx
15:08 olass egonw++ //you are doing a good job on this difficult stuff
15:08 olass I'll help you test as much as I can
16:45 edrin left #bioclipse
18:05 sneumann joined #bioclipse
19:19 sneumann_ joined #bioclipse
19:36 egonw wow... SourceForge has changed the layout
19:36 egonw no more orange...
19:36 egonw it's blue now
20:15 CIA-52 bioclipse: egonw * r11580 /bioclipse2/trunk/features/net.biocl​ipse.pellet_feature/pellet.product: Added product for Bioclipse+RDF+Pellet
20:33 sneumann__ joined #bioclipse
20:48 CIA-52 bioclipse: egonw * r11581 /bioclipse2/trunk/plugins/ (5 files in 2 dirs):
20:48 CIA-52 bioclipse: Factored out a new interface defining the getModel() method, to allow pellet and
20:48 CIA-52 bioclipse: rdf managers to work on the stores the both create (which share the new
20:48 CIA-52 bioclipse: interfaces). In the future, the managers can have a back up route via an RDF
20:48 CIA-52 bioclipse: serialization.

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