Camelia, the Perl 6 bug

IRC log for #bioclipse, 2008-12-04

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

All times shown according to UTC.

Time Nick Message
06:23 Gpox joined #bioclipse
08:10 egonw joined #bioclipse
08:11 egonw moin
08:11 zarah hi egonw
09:21 shk3 joined #bioclipse
09:23 jonalv joined #bioclipse
09:23 jonalv hi
09:23 zarah hi jonalv
09:26 jonalv egonw: oh hai
09:27 egonw hej
09:28 jonalv egonw: Did I manage to change the Fingerprint calculations?
09:29 egonw yes, apparently :)
09:29 jonalv How can that be...
09:29 egonw going to finish up the grant application now...
09:29 egonw we'll check it out this afternoon...
09:29 jonalv oh okey
09:29 egonw do you already have CDK SVN write access?
09:29 jonalv egonw: yes
09:29 egonw please set up a branch of cdk1.2.x there with your patch applied...
09:30 egonw named: jonalv-fingerprint
09:30 egonw with something like:
09:30 jonalv uhm okey I will try.
09:30 egonw svn cp http://cdk.svn.bla/.../branches/cdk1.2.x http://cdk.svn.bla/.../branches/jonalv-fingerprint
09:30 zarah egonw's link is also http://tinyurl.com/5tmsu5
09:30 egonw and then apply the patch and commit
09:34 ever joined #bioclipse
09:35 jonalv hi ever
09:35 ever hi.
09:35 ever i'm just here for moral support :)
09:35 jonalv okey? :)
09:35 ever (friend of edrin's)
09:36 ever but he seems to not be here a the moment
09:36 jonalv ah
09:36 egonw hi ever
09:36 egonw ever: you are hosting the ejabberd server?
09:37 ever i was hosting one, but edrin never used it for his webservice
09:37 egonw oh...
09:37 ever now its a jabberd2
09:37 egonw ic
09:37 egonw how is that different from ejabberd?
09:37 egonw that is, can xws services run via jabberd2 too?
09:38 ever ejabberd is written in an abscure language called erlang, and for some reason i never got it to work again :/
09:38 ever yes sure
09:38 egonw hahaha :)
09:38 ever both support services
09:43 jonalv zarah: seen masak?
09:43 zarah masak was last seen 15 h 41 m 19 s ago in #november-wiki saying 'aloha'
09:44 jonalv zarah: dz
09:44 zarah jonalv: no, YOU'RE the cute one! :)
09:44 jonalv :)
09:44 jonalv botbeer
09:44 zarah ;)
10:48 jonalv egonw: Do you want to eat with us?
10:50 egonw nej tack
10:50 egonw need to get that grant app done
10:55 egonw mmm... amarok *really* rocks!
10:59 jonalv yup
10:59 jonalv amarok++
11:00 egonw http://henrikan.wordpress.com/2008/11​/24/lagg-till-och-lyssna-pa-svenska-r​adiokanaler-i-amarok-2/#comment-101
11:00 zarah egonw's link is also http://tinyurl.com/58uqf3
11:03 jonalv right, all I need now is getting amarok2 to run in Mac OS X... :)
11:11 gilleain joined #bioclipse
11:25 jonalv lunch
11:30 egonw yeah, Nigerian SPAM SCAM in swedish radio :)
11:30 egonw "all I need is your credit card numbers, and those last three digits in particular"
12:22 jonalv hehe
12:22 jonalv egonw: what program are you listening to?
12:23 egonw Star FM
12:23 egonw did not do an extensive search of swedish channels yet...
12:23 egonw but did was sort of OK... bit too Sky Radio to me...
12:29 jonalv I think something is wrong with my computer... :(
12:30 jonalv perhaps I shoujld reboot and restart these benchmark experiments...
12:31 * jonalv reboots... brb
12:35 jonalv joined #bioclipse
12:35 jonalv oh hai
12:35 zarah hi jonalv
13:34 thomas_ku joined #bioclipse
13:46 shk3 hi thomas_ku
13:47 shk3 how is your thesis going?
13:55 CIA-32 bioclipse: egonw * r7952 /bioclipse2/trunk/cdk-externals/trunk/ (41 files in 10 dirs): Synchronized with CDK's jchempaint-primary branch (rev 13462)
14:02 CIA-32 bioclipse: goglepox * r7953 /bioclipse2/trunk/plugins/ (8 files in 5 dirs):
14:02 CIA-32 bioclipse: Search replace Renderer2DModel RendererModel,
14:02 CIA-32 bioclipse: for revision jchempaint-primary@13448
14:03 CIA-32 bioclipse: goglepox * r7954 /bioclipse2/trunk/plugins/ (2 files in 2 dirs):
14:03 CIA-32 bioclipse: Move generator creation to a method to allow overriding.
14:03 CIA-32 bioclipse: And adds selection generator for the editor widget
14:03 CIA-32 bioclipse: goglepox * r7955 /bioclipse2/trunk/plugins/net.bioclipse.cdk.jchem​paint/src/net/bioclipse/cdk/jchempaint/handlers/ (AbstractJChemPaintHandler.java RemoveHandler.java): Remove handler now works on the whole selection
14:03 CIA-32 bioclipse: goglepox * r7956 /bioclipse2/trunk/plugins/net.bioclipse.cdk.jchem​paint/src/net/bioclipse/cdk/jchempaint/business/ (IJChemPaintManager.java JChemPaintManager.java): Added removeBond to the JChemPaintManager
14:04 CIA-32 bioclipse: goglepox * r7957 /bioclipse2/trunk/plugins/net.bioclipse​.cdk.jchempaint.view/src/net/bioclipse/​cdk/jchempaint/view/SWTRenderer.java:
14:04 CIA-32 bioclipse: SWTRenderer now renders PathElement,
14:04 CIA-32 bioclipse: fixed infinit look in reflection code
14:04 CIA-32 bioclipse: goglepox * r7958 /bioclipse2/trunk/plugins/ne​t.bioclipse.cdk.jchempaint/ (2 files in 2 dirs):
14:04 CIA-32 bioclipse: Added a few parameters to ChangeModeHandler
14:04 CIA-32 bioclipse: and added more menu alternatives to JCP-editor
14:07 egonw jonalv: check cell C10
14:07 egonw that has a pretty URL
14:07 egonw I actually get to see where you are...
14:07 egonw your active cell is green :)
14:07 jonalv oh
14:07 jonalv nice link
14:08 jonalv egonw: you think I should add those too? :)
14:08 egonw I will
14:08 jonalv I guess I need to find myself somewhere to blog about this... :)
14:09 thomas_ku hi shk3: the thesis is going quite well!
14:09 thomas_ku I hope I will print it tomorrow!
14:09 shk3 great to hear that
14:10 thomas_ku ;-)
14:11 egonw same here! congrats!
14:12 gilleain indeed, printing is a great step to be at.
14:16 shk3 so when will the verteidigung be?
14:16 thomas_ku it is planned for the afternoon of the 12.02.2009
14:17 shk3 and the party?
14:21 thomas_ku Up to now I haven't planned a party
14:21 thomas_ku but I start the planning after printed the thesis
14:22 shk3 ok, that's accepted
14:22 thomas_ku Not sure the party should be at the same day... but I think so...
14:22 shk3 you want to do it the day before?
14:22 shk3 well, you can try.
14:23 thomas_ku that would properbly clam me down if I drink... the day before... ;-)
14:23 thomas_ku thought also about the day after...
14:36 egonw shk3: did you get bioclipse to boot again?
14:36 shk3 no
14:37 egonw shk3, gilleain: is Christoph back already?
14:39 shk3 no
14:39 shk3 next week
14:55 egonw I just realize Bioclipse can do with a "New MDL molfile" wizard
14:55 egonw gilleain: ping
14:56 gilleain pang
14:56 egonw lasso select should not auto pick bonds, I think...
14:56 egonw but should possibly just focus on the lasso'd atoms...
14:56 gilleain aaaaaaaaaaaargh
14:57 egonw you spent a lot of time on getting bonds included?
14:57 gilleain sigh no, not really.
14:58 egonw you just dropper a hammer on your toes?
14:58 gilleain what do you mean by "auto pick bonds"?
14:58 egonw well... right now it selects atoms which are actually outside the lasso area
14:58 gilleain yes, ask stefan about that.
14:58 egonw because the atoms are involved in bonds which are inside...
14:59 egonw did he?
14:59 egonw shk3: why?
14:59 egonw shk3: it's different from JCP2.x behaviour...
14:59 gilleain he requested that all bonds in the selection have atoms that are also in the selection.
14:59 egonw ah, but that sounds different
14:59 egonw and actually the same as what I ask
15:00 gilleain another way to do it would be to exclude bonds with only one atom in the selection
15:00 egonw or maybe not...
15:00 egonw yes, that's what I'd prefer...
15:00 shk3 the problem was till now it could happen a bond was included, but not both it's atoms
15:00 gilleain personally, I don't care which it is
15:00 shk3 and that later causes trouble
15:00 egonw shk3: why?
15:00 egonw removeAtom() actually removes the atom and all bonds to it...
15:00 gilleain so I shall change it to excluding bonds that have atoms that are not in the selection?
15:00 shk3 well, many cdk classes don't work on that
15:00 egonw so, one should actually never end up with dangling bonds
15:01 shk3 if you add a bond, yes
15:01 egonw shk3: huh? adding a bond? where does that come in?
15:01 shk3 I suppose in the selection code
15:01 shk3 but I don
15:01 shk3 't know
15:02 shk3 sorry, went through all of this with giliean
15:02 shk3 fact is it happened
15:02 shk3 I don't know why
15:02 egonw gilleain: I guess it should not select atoms which are just not inside the select area
15:02 shk3 and I don't need to know why, do I?
15:03 egonw shk3: sorry... you completely lost me...
15:03 egonw shk3: what problem did you actually find?
15:03 shk3 I report a bug. Fullstop.
15:03 egonw sure, we can discuss it there
15:03 gilleain egonw: well, more precisely, it should not select bonds whose atoms are not in the selection area.
15:03 shk3 that the smiles parser failed on the selection because the selection contained bonds
15:03 shk3 the atoms of which where not in the selection.
15:04 egonw shk3: ummm... and why do you SMILES-parse a selection??
15:04 egonw sorry, I don't understand that use case
15:04 shk3 perhaps it was another case
15:04 shk3 don't know
15:04 egonw gilleain: I can live with that
15:05 gilleain cool, implementing...
15:05 shk3 in any case it is not good code, is it?
15:05 egonw what is not?
15:05 egonw SMILES parsing a slection?
15:05 egonw that indeed does not sounds goo
15:05 egonw that indeed does not sound good
15:06 shk3 grrrrrrrrrr
15:06 shk3 perhaps it wasn't the smiles parser (or generator, probably)
15:07 shk3 but in any case an atomcontainer where bonds point to atoms not in the atom container is bad, I would say
15:07 egonw yes, that is very bad too
15:07 shk3 and if you think it is fine, ok
15:07 shk3 but I don't want to discuss this a second time for hours
15:07 shk3 sorry
15:07 egonw did we before?
15:08 shk3 no, but i did with gilliean
15:09 egonw right, well I was not there, but I also do not need to redo that discussion...
15:10 egonw just tried to understand what problem you found, ok?
15:10 gilleain I didn't realise that the cdk has a very 'loose' graph model where you can have any combination of atoms in containers bonded to other containers etc.
15:10 gilleain so stefan had to explain :-/
15:10 shk3 I found: the selection counld have bonds in it with atoms which were not in the selection.
15:11 shk3 that's all
15:11 egonw gilleain: yeah, that's definately a point of improvement for the CDK...
15:11 shk3 nothing mor to say
15:11 shk3 if it was the smiles generator or anything else which made me realize this is of no importance
15:11 gilleain not necessarily, it makes it quite flexible.
15:11 egonw shk3: actually, that sounds not that unreasonable actually...
15:11 egonw shk3: it's not necessary, but one could select a number of bond, and delete those, but keep the atoms...
15:12 shk3 but that's the other way round
15:12 shk3 ah ok, I see what you mean
15:12 egonw ah, ok... then I did understand it the first time...
15:12 egonw that would mean that somewhere a wrong call is being made...
15:12 shk3 I don't know what gilliean did, but we had the idea
15:12 Julio joined #bioclipse
15:12 shk3 the shapeselection might have a
15:13 egonw removeAtom() and not removeAtomAnAllConnectedElectronCOntainers()
15:13 egonw @ the CDK level
15:13 shk3 getselection which gives an atomcontainer
15:13 egonw not ControllerHub
15:13 egonw that sounds like a serious bug
15:13 shk3 this should not have bonds-without-atoms
15:13 Julio left #bioclipse
15:13 shk3 the internal model of the selection might
15:13 egonw as long as you are not doing SMILES generation on getSelection()
15:13 egonw I don't see why one would like to do that
15:14 egonw that is... for the kind of selection we are currently looking at
15:14 egonw ah... ok
15:14 egonw well, this is why we were having a partial rerun of your earlier discussion
15:15 egonw the distinction between selection types is critical...
15:16 shk3 it was copy/paste
15:16 shk3 the selection needs to go into clipboard
15:16 shk3 as e. g. a smiles
15:16 shk3 or a molfile
15:17 shk3 so I need to generate smiles on selection
15:17 shk3 or use the mdl writer
15:17 egonw ah... interesing use case...
15:17 egonw thanx... that was informative...
15:18 egonw IMO, this is a problem with the SMILES and MDL molfile format...
15:18 egonw unless we require selections to not have dangling bonds...
15:18 egonw which directly means, that the JCP editor will never allow removing of just bonds with lasso mode...
15:18 egonw trade off...
15:19 jonalv egonw: I can not get this caching idea to increase speed. Actually it's the other way around... :)
15:19 egonw jonalv: :)
15:20 jonalv this is why you should only to optimisation when you have some way to measure things... :)
15:20 egonw right
15:26 shk3 so why not have a getSelection() and getValidAtomContainerSelection() or so in the shapeselection?
15:27 egonw because the code that puts stuff on the clipboard can perfectly well do it itself, so that we don't have to clutter the API of shapeselection
15:29 egonw shk3, gilleain: I'd like to change the IChemModelRelay to have addAtom() and sort return the new IAtom
15:29 egonw any objections?
15:29 egonw you can always ignore it..
15:29 gilleain no, fine
15:29 egonw but it would clean up code like this:
15:29 shk3 sure
15:29 egonw addAtomTo(worldCoordinate);
15:29 egonw newAtom = chemModelRelay.getClosestAtom(worldCoordinate);
15:29 egonw which actually I am looking at because of a NPE bug
15:29 gilleain that was my code :/
15:29 gilleain sorry about the bug
15:29 egonw no worries...
15:30 egonw crappy IChemModelRelay API
15:32 egonw oh... that looks bad...
15:32 egonw who wrote that code??
15:33 * egonw boots git blame
15:33 egonw and suspects himself or gilleain
15:33 egonw and hopes it was not the first :)
15:33 gilleain which code?
15:33 egonw public IRing addRing(int ringSize, Point2d worldcoord) {
15:34 egonw well... maybe it's not so bad in itself...
15:34 egonw expect there is other code not expecting it, I think...
15:34 egonw ControllerHub 463-472
15:34 gilleain that would be me
15:34 gilleain only I have comments like //err...
15:34 egonw :)
15:35 gilleain it's for creating a ring not attached to anything.
15:35 egonw yeah, that makes sense
15:36 egonw the problem is, that I am sure there is other code around, which only looks at the first IAtomContainer...
15:36 gilleain hmmm
15:36 egonw worse is that it overwrites the IMoleculeSet
15:36 egonw so, this should go fine when nothing is drawn yet
15:36 gilleain it does?!
15:36 gilleain ...
15:37 egonw but I guess everything drawn elsewhere will dissappear
15:37 gilleain ahhh.
15:37 gilleain that explains a bug.
15:37 egonw hehehe :)
15:37 gilleain shall I change it?
15:38 egonw yes, please do...
15:38 gilleain ok
15:38 egonw for now, please just add it to the first atom container...
15:38 egonw or, alternatively... use chemModel.getMoleculeSet()
15:40 gilleain sure
15:41 gilleain IMoleculeSet set = chemModel.getMoleculeSet();
15:41 gilleain if (set == null) {
15:41 gilleain set = chemModel.getBuilder().newMoleculeSet();
15:41 gilleain }
15:41 gilleain set.addAtomContainer(ring);
15:41 gilleain ??
15:42 egonw yes
15:42 egonw though set should never be null
15:42 gilleain hmmm. 'should'.
15:42 egonw because something would be seriously broken then...
15:42 egonw rigth
15:42 egonw maybe make some noise in the else clause?
15:42 egonw because whatever led to that situation sounds really bad
15:43 gilleain if you say new ChemModel()...
15:43 egonw ah...
15:44 egonw good argument for not having added setChemModel() ! :)
15:44 CIA-32 bioclipse: egonw * r7959 /bioclipse2/trunk/cdk-externals/trunk/ (2 files in 2 dirs): Return what has been created
15:44 egonw I knew there must be a reason why I was unsure about the method :)
15:44 gilleain but you generally start with one somewhere.
15:45 gilleain unless you have a hub method newChemModel() which makes one properly, internally.
15:46 egonw yes... that sounds like a good idea
15:46 egonw or
15:46 egonw throw a BadInputException
15:46 egonw when it does not look proper...
15:46 egonw then again...
15:46 egonw there are degrees or proper...
15:46 egonw same for fixing it...
15:46 jonalv egonw: I found another 4% optimisation + the code got cleaner with this one :)
15:46 egonw not on GDocs yet?
15:47 egonw that's 4% compared to -465?
15:47 egonw or to -473?
15:47 jonalv 473
15:53 CIA-32 bioclipse: egonw * r7960 /bioclipse2/trunk/cdk-externals/trunk/org​.openscience.cdk.control/src/org/openscie​nce/cdk/controller/DrawBondModule.java: 1. Make use of powers of the new and improved IChemModelRelay API; 2. no excetion thrown, so no need to catch it
16:01 edrin joined #bioclipse
16:01 edrin hi
16:01 zarah hi edrin
16:01 edrin hi justin
16:51 edirn joined #bioclipse
16:51 edirn egonw, shk3 is the webservice plugin working for you?
16:51 shk3 for me?
16:51 edirn shk3 yes
16:51 shk3 for me not even bioclipse works...
16:52 edirn bc2 ?
16:52 edirn why?
16:52 egonw edirn: ask me tonite or tomorrow
16:52 egonw got to catch the bus now...sorry
16:52 egonw will try!
16:52 edirn ok
16:54 shk3 edirn: since friday, I can't start bioclipse any more
16:54 edirn ok
16:54 edirn well, not ok
17:00 gilleain what's goes wrong with the webservices plugin?
17:01 edirn gilleain for me it did no work
17:01 edirn gilleain did it work for you?
17:01 edirn back in two min...
17:03 edirn joined #bioclipse
17:04 ever jo edrin
17:04 gilleain edrin: I haven't tried it recently.
17:04 gilleain but what didn't work?
17:10 edirn i think the wizard for retreiving stuff
17:11 edirn i'm trying to fix it
17:11 gilleain oh, right.
18:07 CIA-32 bioclipse: edrin_t * r7961 /bioclipse2/trunk/plugins/net.bioc​lipse.webservices/src-axis1/com/th​emindelectric/www/wsdl/GetEntry/:
18:09 CIA-32 bioclipse: edrin_t * r7962 /bioclipse2/trunk/plugins/net.bioclipse.w​ebservices/src/net/bioclipse/webservices/ (5 files in 4 dirs): fixed to adapt current stuff from embl<ebi
18:10 CIA-32 bioclipse: edrin_t * r7963 /bioclipse2/trunk/plugins/net.biocli​pse.webservices/src-axis1/uk/ac/ebi/ (wsdbfetch/server/ www/ws/services/urn/Dbfetch/):
18:14 CIA-32 bioclipse: edrin_t * r7964 /bioclipse2/trunk/plugins/net.biocli​pse.webservices/src-axis1/net/biocli​pse/webservices/clients/WSWUBlast/:
18:19 CIA-32 bioclipse: edrin_t * r7965 /bioclipse2/trunk/plugins/net.bioclipse.w​ebservices/src-axis1/uk/ac/ebi/jdbfetch/ (6 files in 2 dirs):
18:22 CIA-32 bioclipse: edrin_t * r7966 /bioclipse2/trunk/plugins/net.bioclipse.webservic​es/src-axis1/uk/ac/ebi/www/ws/services/WSDbfetch/ (6 files):
18:23 CIA-32 bioclipse: edrin_t * r7967 /bioclipse2/trunk/plugins/net​.bioclipse.webservices/jars/ (4 files in 2 dirs):
18:24 CIA-32 bioclipse: edrin_t * r7968 /bioclipse2/trunk/plugins/net.bioclip​se.webservices/META-INF/MANIFEST.MF:
18:27 CIA-32 bioclipse: edrin_t * r7969 /bioclipse2/trunk/plugins/net.bi​oclipse.webservices/.classpath:
19:29 egonw joined #bioclipse
20:07 edrin hi egonw
20:07 egonw hi edrin
20:08 edrin egonw: web service(s) should work again
20:08 egonw yes, just replied to your email
20:08 egonw happy to hear that!

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