Camelia, the Perl 6 bug

IRC log for #bioclipse, 2009-03-03

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

All times shown according to UTC.

Time Nick Message
05:11 masak joined #bioclipse
06:25 olass joined #bioclipse
07:33 Gpox joined #bioclipse
07:38 Gpoks joined #bioclipse
08:10 egonw joined #bioclipse
08:37 jonalv joined #bioclipse
08:40 jonalv oh hai
08:40 zarah hi jonalv
08:41 rojasm joined #bioclipse
08:54 edrin joined #bioclipse
08:54 edrin hi
08:54 zarah hi edrin
09:11 shk3 joined #bioclipse
09:39 CIA-55 bioclipse: carl_masak * r9194 /bioclipse2/trunk/plugins/net.bioclipse​.scripting/src/net/bioclipse/scripting/ (JsEnvironment.java JsThread.java):
09:39 CIA-55 bioclipse: [scripting] tweaks having to do with exceptions
09:39 CIA-55 bioclipse: We currently have a problem in which scripting sometimes hangs on tab
09:39 CIA-55 bioclipse: completion. (#628)
09:47 shk3 egonw: shall I go ahead with a 3d+2d coords content type
09:47 shk3 ?
09:49 egonw make it a subtype of the top level CML
09:57 shk3 yes
09:57 shk3 that's what I planned
11:30 edrin left #bioclipse
11:31 egonw joined #bioclipse
11:42 egonw umm... guys, I found a cheaper way to get LEGO building blocks for that game:
11:42 egonw http://gadgets.boingboing.net/200​9/03/02/video-using-lego-to.html
11:42 zarah egonw's link is also http://tinyurl.com/c68rdw
11:42 egonw check the last slide...
11:43 egonw all we have to do is come up how to explain proteochemometrics with LEGO bricks
11:44 shk3 egonw: with respect to create2d/3dcoordinates
11:44 egonw yes
11:44 shk3 making this create coordinates in existing mol seems tricky
11:44 jonalv egonw: sounds easy... uhm perhaps not...
11:44 shk3 this is an Imolecule
11:44 shk3 so effectitvly
11:44 shk3 we get a smiles or cml from it
11:45 shk3 not (necessarily) the internal data structure
11:45 shk3 I can create an icdkmolecule from it
11:45 egonw or cast
11:45 egonw in many cases that will work...
11:45 shk3 but setting coords in there won't change the original
11:45 egonw best to do that via getAdapter() I think
11:45 shk3 in many
11:45 shk3 yes
11:45 shk3 if it is an icdkmolecule
11:45 shk3 but else
11:46 egonw [12:45:29] <shk3> but setting coords in there won't change the original
11:46 egonw what do you mean?
11:46 egonw these actions have to change the ICDKMolecule, not a derivative
11:46 shk3 but it must not be an icdkmolecule
11:46 shk3 it least not now
11:46 shk3 the method has IMolecule as parameter
11:46 shk3 not ICDKMolecule
11:47 egonw I do not understand what you are trying to get at
11:47 shk3 wll
11:47 shk3 input is an IMolecule.
11:47 shk3 right?
11:48 shk3 please say yes or no
11:48 egonw of the manager methods?
11:48 shk3 yes
11:48 shk3 that's what we are taling about, isn't it?
11:49 egonw I think use is inconsistent
11:49 shk3 why?
11:49 egonw but I agree it would be good (tm) if they all took IMolecule
11:49 egonw no clue
11:49 egonw checking...
11:49 shk3 they do
11:49 egonw anyway, regarding 2d/3d...
11:50 egonw CML does both...
11:50 egonw so getCML() should not be a problem regarding 2D/3D coords, or... ?
11:50 shk3 well, it could
11:50 egonw could or is ?
11:50 shk3 is
11:50 shk3 let me explain
11:50 egonw yes, please :)
11:50 shk3 we do not know internals if IMolecules
11:50 shk3 right?
11:50 egonw no
11:51 shk3 of IMolecules, I mean
11:51 egonw and the problem is that there is not setCML() or so?
11:51 shk3 so getCML gives as a representation of the content
11:51 shk3 not the content as a such
11:51 shk3 if I change this representation,
11:51 egonw explain
11:51 shk3 the conent won't change,
11:51 egonw what is not serialized?
11:51 shk3 necessaritly
11:51 egonw ah, in that way...
11:51 egonw ok, indeed
11:51 shk3 even in ICDKMolecule
11:52 shk3 changing the cml you get
11:52 shk3 does not change the IAtomcontainer
11:52 shk3 so If you do getcml
11:52 shk3 change it
11:52 shk3 and do getcml again
11:52 shk3 you get the same as the first time
11:52 egonw right
11:52 egonw so, setCML is missing()
11:52 egonw ?
11:52 shk3 since both times the Iatomcontainer is written out
11:52 shk3 in a way
11:52 shk3 depends on what one wants
11:53 shk3 it is okay right now
11:53 shk3 just don't rely on things which are not in the contract
11:53 jonalv http://img186.echo.cx/img186/8749/vi7kt.gif steinbeck++
11:53 zarah jonalv's link is also http://tinyurl.com/am975d
11:53 shk3 btw, the comment of the createcoordiante methods say
11:53 shk3 Create 2D coordinate for the given molecule
11:53 shk3 not
11:53 shk3 Create 2D coordinate in the given molecule
11:54 shk3 so that's consistent
11:54 egonw jonalv: cool !
11:54 jonalv yea from the steinblog :)
11:55 egonw shk3: OK, I think those semantics in those labels got lost
11:56 egonw shk3: but we still need an action to change the coordinates *in* an existing resource, agreed?
11:56 shk3 that can be done by overwriting
11:57 shk3 IMolecule is a read-only data type
11:57 shk3 that was always my understanding
11:58 egonw no, don't think that is the intention...
11:58 egonw what about JCP then??
11:59 egonw shk3: we have a meeting here in about 1 minute... I'll bring it up...
11:59 shk3 does that work on IMolecule?
11:59 shk3 don't think so
11:59 shk3 it works on files
11:59 shk3 and has save methods
11:59 shk3 save changes the content
11:59 shk3 that's for sure
11:59 egonw no, the JCP manager does not work on files
11:59 egonw but on IMolecule or ICDKMOlecule...
12:00 shk3 ah
12:00 egonw not sure, but should really be the first...
12:00 shk3 don't understand how that's supposed to work
12:00 egonw but thanx for bringing this up... it *is* an important issue...
12:00 egonw I'll make sure to define some clear decisions here
12:00 egonw and convert those into unit tests...
12:00 shk3 my understanding of the creaet stuff was:
12:00 shk3 read molecule
12:01 shk3 newmolelule = create2dcoords(oldmolecule)
12:01 shk3 save newmolecule
12:01 shk3 either overwriting oldmolecule
12:01 shk3 or writing new file
12:05 CIA-55 bioclipse: shk3 * r9195 /bioclipse2/trunk/plugins/net.bioclipse.ui/ (2 files in 2 dirs): created a 2d and 3d=5d coordinate type
12:06 egonw shk3: please add unit tests too
12:07 shk3 what for?
12:07 shk3 sorry
12:07 shk3 you refer to content type?
12:07 egonw yes
12:07 shk3 more will follow
12:07 egonw ok, good
12:08 shk3 doesn't look to good with respect to unit tests in that field, btw
12:13 CIA-55 bioclipse: shk3 * r9196 /bioclipse2/trunk/plugins/net​.bioclipse.cdk.ui/plugin.xml: the create coordinates menus fit with 2d/3d/5d conent types
12:30 jonalv @seen olass
12:30 zarah olass was last seen 21 h 40 m 14 s ago in #farmbio saying 'gick den bra?'
12:36 masak famous last words...
13:37 edrin joined #bioclipse
13:37 edrin hi again
13:48 grandm44 joined #bioclipse
13:49 grandm44 hi all
13:49 zarah saluton grandm44
13:57 CIA-55 bioclipse: shk3 * r9197 /bioclipse2/trunk/plugins/net.bioclipse.ui/ (2 files in 2 dirs): cleanup of coordinate types
14:06 CIA-55 bioclipse: shk3 * r9198 /bioclipse2/trunk/plugins/net.bioclipse.cdk.busine​ss/src/net/bioclipse/cdk/business/CDKManager.java: fixes bug #638 (Create 2D coordinates deletes 3D coordinates)
14:06 zarah bug #638 | http://tinyurl.com/cgh5b5
14:07 CIA-55 bioclipse: shk3 * r9199 /bioclipse2/trunk/plugins/net.bioclips​e.cdk.business.test/src/net/bioclipse/​cdk/business/test/CDKManagerTest.java: tests for bug #638 (Create 2D coordinates deletes 3D coordinates)
14:07 zarah bug #638 | http://tinyurl.com/cgh5b5
14:46 olass joined #bioclipse
14:50 jonalv egonw: are there plans for getting 2d coordinate generation from smiles to be able to hadnle this sort of SMILES nicely: CN1CCC23C4C1CC5=C2C(=C(C=C5)O)OC3C(C=C4)O ?
15:05 shk3 what's particular about it?
15:05 shk3 sdg should work for that, I would say.
15:32 egonw it does
15:33 egonw it's just that it does not do one of the loops nicely
15:33 egonw partly because the SMILES did not have stereochemistry
15:33 egonw BTW, 3d coord generation seems to crash on it
15:43 egonw shk3: did you commit the unit tests for the 5D content type detection? I did not see them, and interested to see how you did it
15:45 shk3 I sort of gave up
15:45 shk3 but I will come back on this
15:45 egonw please do
15:45 shk3 there is a ContentTypeHandler or so
15:45 shk3 but I could not yet figure out how to use it
15:45 egonw as I have to write/we need  unit tests for all other content types too
15:46 shk3 supply it with conenttypes, that is
15:46 shk3 indeed, none is tested
15:46 shk3 ContentTypeMatcher
15:47 shk3 that's what I think one should use
15:47 shk3 but I can't see how it gets its types
15:47 shk3 feel free to find out
15:51 CIA-55 bioclipse: ospjuth * r9200 /bioclipse2/trunk/plugins/net.bioclipse.​cdk.business.test/src/net/bioclipse/cdk/​business/test/CDKManagerPluginTest.java: Extended properties test for IAtomContainer props on 2D generation. More rigorous test can never be wrong, right?
15:51 olass shk3: one idea would be to look at how Eclipse test their content types
15:51 olass I'm sure they have junit tests
15:52 egonw shk3: you could ask on #eclipse
15:52 shk3 you could ask ;-)
15:52 egonw I will when I'll the unit tests for the yet existing content types...
15:53 egonw and I would have when I had done the 5D content type too
15:53 egonw but you decided to take that one over, and consequently took responsibility for the unit test for that code too
15:54 olass shk3: does the net.bioclipse.contenttypes.cml.singleMolecule5d not work?
15:54 egonw shk3: for now, please leave the bug open, reassign to me, with a comment about having added an implementation, but that you could not work out a unit test
15:54 egonw olass: it probably does
15:54 olass ok
15:54 shk3 it works
15:54 olass good
15:54 egonw but we need to have means to see if it works in 3 months too
15:54 egonw that's what the unit tests is for
15:54 shk3 it is as good as all other content types
15:55 olass :-)
15:55 egonw not for testing if something works now
15:55 egonw but to prevent regressions
15:55 shk3 guys, I might look at the unit tests
15:55 egonw shk3: which are untested too, so a rather meaningless statement
15:55 shk3 but I can't see why everybody else is out of this business
15:55 egonw I'm not
15:55 egonw I have a bug report about it
15:55 shk3 the person who did the first conent type
15:55 shk3 should have done it in a way
15:56 egonw yes, but that does not remove any responsibility from those who add code too
15:56 egonw just because someone jumps in a river, does not mean that everyone has to
15:56 egonw o
15:56 shk3 just because I volunteer for some bit,
15:56 shk3 doesn't mean everything can be thrown at me
15:56 egonw right... you picked to fix do it...
15:57 egonw so you took responsibility for it...
15:57 egonw you voluntered...
15:57 egonw the bug report was there *not* because it is difficult to add those lines to the plugin.xml
15:57 egonw but because of a proper way to test it...
15:58 egonw just because people forgot to complete code with unit tests in the past, does not mean that you do not have to make them either
15:58 egonw go to catch my bus now
15:58 egonw [16:54:25] <egonw> shk3: for now, please leave the bug open, reassign to me, with a comment about having added an implementation, but that you could not work out a unit test
15:59 shk3 done
16:17 olass @pz
16:17 zarah olass: Pelezilla is at http://bugs.bioclipse.net
16:44 CIA-55 bioclipse: carl_masak * r9201 /bioclipse2/trunk/plugins/net.bioclipse.scrip​ting.ui/src/net/bioclipse/scripting/ui/views/ (JmolConsoleView.java RConsoleView.java):
16:44 CIA-55 bioclipse: [scripting] added R Console and Jmol Console
16:44 CIA-55 bioclipse: These don't do anything yet. Still need to plug them in.
16:46 CIA-55 bioclipse: carl_masak * r9202 /bioclipse2/trunk/plugins/n​et.bioclipse.scripting.ui/ (3 files in 3 dirs):
16:46 CIA-55 bioclipse: [scripting] wired up the Jmol console
16:46 CIA-55 bioclipse: Theoretically, it shoudl work now.
16:46 masak ouch, typo. :/
16:46 masak ok, hold on and I'll add the Jmol view to the plugin.xml too. then we "shoudl" be up and running.
16:56 masak oh noes, a cycle! :(
17:00 CIA-55 bioclipse: carl_masak * r9203 /bioclipse2/trunk/plugins/n​et.bioclipse.scripting.ui/ (3 files in 3 dirs):
17:00 CIA-55 bioclipse: Revert "[scripting] wired up the Jmol console"
17:00 CIA-55 bioclipse: This reverts r9202.
17:00 CIA-55 bioclipse: Turned out I wasn't allowed to make scripting.ui depend on jmol... that created
17:00 CIA-55 bioclipse: a cycle.
18:00 ilogger2_ joined #bioclipse
19:54 rojasm left #bioclipse

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