Camelia, the Perl 6 bug

IRC log for #cdk, 2008-02-02

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

All times shown according to UTC.

Time Nick Message
00:10 edrin joined #cdk
04:10 jpansanel joined #cdk
04:55 CIA-18 jmol: hansonr * r8944 /trunk/Jmol/src/org/jmol/ (3 files in 3 dirs): 11.5.4 # new feature: reading of Spartan spinput files (uses Odyssey reader)
05:22 jbrefort joined #cdk
07:27 rojasm joined #cdk
07:50 egonw joined #cdk
07:54 egonw moin
07:54 egonw hi rojasm
07:55 rojasm moin egonw
07:55 rojasm I sent just a seconds before a mail
07:55 egonw ah good
07:56 egonw rojasm: what about 18-19
07:56 egonw mom...
07:57 rojasm Freb?
07:57 egonw yes
07:57 egonw but just remember I have that currently as back up
07:57 egonw for another meeting...
07:57 egonw hang on
07:58 egonw btw, the CDK article is at 40 cites now :)
07:58 rojasm :)
07:58 rojasm that is good news
07:58 rojasm I will check too
07:58 egonw no, can't use 18-19 yet
07:59 egonw 25-26?
07:59 egonw still feb
08:00 rojasm a moment I check
08:00 rojasm perfect! then 25-26 febr
08:01 egonw OK, set
08:01 rojasm :)
08:02 rojasm I will sent a mail around the group
08:02 egonw OK
08:03 egonw feel free to schedule a seminar
08:03 egonw (to be given by me)
08:03 egonw if there is interest in that
08:03 egonw is optional
08:03 egonw but possible
08:03 egonw ok, back to mf/ and reaction/ ?
08:04 rojasm I will sent a mail about the idea for reactionTree which I think sound better to create a class called
08:04 rojasm System
08:04 rojasm which contains reactions linked between them
08:04 egonw rojasm: please read the CMLReact paper
08:04 egonw that will give you some good details of the semantics involved
08:05 rojasm ok
08:13 egonw got some secs for discussion mf/ and reaction/ ?
08:13 rojasm of course
08:13 egonw and whether you want to make the 10000 commit or not...
08:14 rojasm I am working now with the reaction staff
08:14 egonw 10 left... so that will happen this weekend
08:14 egonw about mf/
08:14 egonw please check nightly for PMD warnings
08:14 egonw acout reaction...
08:14 egonw about...
08:15 egonw the MreactionTests cannot currently be run
08:15 egonw I can help with fixing that
08:15 egonw but was not sure what you already did
08:15 egonw that is... what uncommited patches you have hanging around
08:15 rojasm I say you email. I did but not yet submitted
08:16 rojasm let me time until today aftermoon. And I will submit all
08:16 egonw sure
09:24 rojasm egonw: question please. What class give the element. I remember some thing like Element.Carbon
09:24 rojasm but I don't find it
09:25 rojasm Element.Elements
09:25 rojasm found
09:26 egonw good
09:33 carsten joined #cdk
09:37 egonw hi carsten
09:37 egonw you're at the PIM meeting?
09:38 carsten nope, wrong city :)
09:38 egonw ah, ack
09:41 carsten but many commits
09:47 egonw yeah, good to hear that...
09:48 egonw 4.0.1 was tagged two days ago...
09:48 egonw so in five days the tgz's and deb's should be available, right?
09:48 egonw that is, one week for the packagers, right?
09:49 egonw these pims won't be ready until 4.1, though... I assume
09:51 carsten [10:48] <egonw> so in five days the tgz's and deb's should be available, right?         <--- right
09:51 carsten no, all work on pim is for 4.1. KDEPIM isn't even a part of 4.0!
09:52 carsten egonw: Mailody has just been ported to use Akonadi, for example (just == 2 weeks ago)
09:52 carsten That port already uncovered bugs und missing stuff in Akonadi.
09:52 carsten Now the pimsters are also talking with some gnomes
09:53 egonw right
09:54 egonw it's a shame to see that in KDE4 development, the release early, release often is less common
09:54 egonw these demo programs should be out...
09:54 egonw for people to test, to talk about...
09:54 egonw not file bug reports...
09:54 egonw but get the buzz going...
09:54 egonw that attracts developers...
09:54 egonw it
09:55 egonw 's now too much hidden...
09:55 egonw closed development...
09:55 egonw semi-open rellay.
09:55 carsten well, on the pimside there is simply a lack of manpower
09:55 egonw but so closed that new-comers don't see a clear invitation to hack on whatever they are hacking on...
09:55 egonw right
09:55 egonw so, release-soon, release-often
09:55 carsten and people simply don't start porting their apps until it is called stable
09:55 egonw rigth...
09:55 egonw that's the culprit
09:56 egonw that's wonr
09:56 egonw wrong
09:56 carsten there was a loooooooong discussion about releasing "only libs" and similar models
09:56 egonw bad choice, IMHO :)
09:56 egonw people don't get enthousiastic about libs
09:56 egonw they good amazed by applications
09:56 egonw which wrap around those cool libs
09:56 carsten indeed.
09:57 egonw only *existing* developers get enthousiastic about cool new libs
09:57 carsten Well, we already have quite a few "TODO KDE5" comments in our code
09:57 carsten [10:54] <CIA-5> osterfeld * r769870 akregator/trunk/KDE/kdepim/akregator/src/ (subscriptionlistview.cpp subscriptionlistview.h): Implement next feed and next unread feed actions
09:57 egonw but attracting *new* developers is much easier than getting *existing* developers to route manpower to the development of the new libs
09:57 carsten ^^^ those are the cool commits :-)
09:57 egonw KDE5 ???
09:57 egonw that's plain stupid
09:58 egonw that's saying: don't work on this, because we already plan to do this...
09:58 egonw yes, not until 2012 has come...
09:58 egonw but no use starting hacking something yourself now
09:58 carsten nono, that is for example remobing deprecating methods!
09:58 carsten we cannot remove them in libs until kde5
09:59 egonw arghh... KDE4 is as unpolished as can be...
09:59 egonw do people really require KDE 4.x to be compatible with 4.1?
09:59 carsten you mean 4.1 with 4.0?
10:00 egonw well, no, because PIM won't be around before 4.1 :)
10:00 carsten Well, if I were "king for a day" I would say: Get rid of this crap
10:00 egonw but yeah...
10:00 egonw indeed
10:00 egonw you'll have to maintain more code
10:00 egonw without adding functionality...
10:00 carsten /**
10:00 carsten -   * Returns whether the next action shall be initialized.
10:00 carsten +   * @deprecated not useful in public API
10:00 egonw I would say... deprecated code is dropped in the next .x release...
10:00 carsten */
10:00 carsten -  bool initializeNextAction() const;
10:00 carsten +  KDE_DEPRECATED bool initializeNextAction() const;
10:00 carsten look at that commit for example, from this morning
10:39 rojasm egonw: could you prove that. It is some thing strange in addImplicitHydrogen
10:39 rojasm IMolecule molecule2 = builder.newMolecule();
10:39 rojasm molecule2.addAtom(builder.newAtom("C"));
10:39 rojasm molecule2.getAtom(0).setFormalCharge(1);
10:39 rojasm
10:39 rojasm addExplicitHydrogens(molecule2);
10:39 rojasm result with SMILES: [H][C+][H]
10:40 rojasm I checked and the problem is with addImplicitHydrogens(container);
10:40 rojasm as well
10:41 egonw that's one H missing, right?
10:41 egonw mom
10:41 rojasm yes!
10:42 egonw weird indeed
10:42 egonw ok, will make that a unit test...
10:43 rojasm and wenn this atom exactly comes with formalcharge 2 then AtomTypeManipulator doesn't recognize
10:43 rojasm as well
10:46 egonw C+2??
10:47 egonw does that exist
10:47 egonw ?
10:47 egonw that's indeed not an atom type at this moment...
10:51 egonw rojasm: weird indeed...
10:52 egonw check CDKAtomTypeMatcherTest
10:52 egonw line 709-ff
10:52 egonw so, having implicit hydrogens must be the problem...
10:53 rojasm yes! the problem with implicithydrogens
10:53 rojasm but
10:54 rojasm if you put [C+2]
10:54 rojasm there is the problem then with CDKAtomTypeMatcher
10:54 CIA-18 cdk: egonw * r9991 /trunk/cdk/src/org/openscience/cdk/test​/atomtype/CDKAtomTypeMatcherTest.java: Added test for C.plus.sp2 without explicit hydrogens...
10:54 egonw right, but is C+2 something really existing??
10:54 rojasm I don't know
10:55 egonw can't remember ever having seen it...
10:55 rojasm ok, I see what you mean
10:55 egonw BTW, 9 commits to go...
10:55 rojasm the commit r1000 is already here :)
10:55 rojasm :)
10:56 egonw right... do you want to make it?
10:56 egonw gets you some CDK fame :)
10:56 rojasm I think, you should do it
10:56 rojasm I will make the r1001 :)
10:56 egonw Christoph wondered too why I did not want to make it...
10:57 egonw the idea is that I already make so many commits...
10:57 egonw it's nice if someone else gets some commit fame too
10:57 egonw that's why I'd prefer to see someone else make the 10.000th commit
10:58 rojasm But I think that with the commits is not reallistic because you can allways pack the commits in one or split them in 10
10:58 egonw that's what I do...
10:58 egonw that actually is the right thing to do...
10:58 egonw make commits as atomic as possible...
10:58 rojasm you mean split them
10:58 rojasm or pack
10:58 egonw don't commit 10 bug fixes in one commit...
10:58 egonw but make those 10 commits instead
10:59 egonw so that people can apply fix-by-fix without having to remove unwanted bug fixes
10:59 rojasm ok, I see
11:10 CIA-18 cdk: egonw * r9992 /trunk/cdk/ (42 files in 8 dirs): Fixed libiocml module: remove '-' and added a test suite for the module (closes PMT #1748306)
11:15 egonw lunch time
12:19 egonw back from lunch
13:38 CIA-18 jmol: hansonr * r8945 /trunk/Jmol/src/org/ (8 files in 5 dirs): (log message trimmed)
13:38 CIA-18 jmol: 11.5.4
13:38 CIA-18 jmol: # new feature: getProperty("appletInfo") includes a key "registry" for which
13:38 CIA-18 jmol: # the value is the static Hashtable htRegistry
13:38 CIA-18 jmol: # //note that the following JavaScript retrieves the registry:
13:38 CIA-18 jmol: # var registry = jmolGetPropertyAsJavaObject("​appletInfo").get("registry")
13:38 CIA-18 jmol: # // and the following code then retrieves an array of applets:
13:54 egonw rojasm: how are things getting along?
13:54 rojasm just finishing
13:54 rojasm only two reactions
15:12 egonw doing some shopping
15:18 CIA-18 cdk: egonw * r9993 /branches/egonw/ifragment/src/org/openscience​/cdk/test/DefaultChemObjectBuilderTest.java: Added missing unit test
15:19 rojasm ant -D=reaction module-test
15:20 egonw -Dmodule=reaction
15:20 rojasm thanks
15:20 egonw and test-module
15:20 rojasm I forgot allways
15:20 egonw :)
15:25 CIA-18 cdk: egonw * r9994 /trunk/cdk/src/org/openscience/cdk/ (8 files in 5 dirs): Moved IFragmentAtom to the interfaces module (closes #1872765)
15:25 egonw OK, that was easier than anticipated
15:25 CIA-18 cdk: egonw * r9995 /branches/egonw/ifragment/: Patch applied to trunk/. Branch closed.
15:26 egonw ok, really going for shopping now
16:06 rojasm egonw arround?
16:06 egonw yes
16:06 egonw just back
16:06 egonw but not for long
16:06 egonw because I need to start making dinner soon
16:06 egonw what's up?
16:06 rojasm ok, only a point :)
16:06 rojasm IMolecule molecule = (new SmilesParser(org.openscience.cdk.DefaultChemObject​Builder.getInstance())).parseSmiles("Fc1ccccc1");
16:06 rojasm result
16:06 rojasm SmilesGenerator sg = new SmilesGenerator();
16:06 rojasm System.out.println(sg.createSMILES(molecule));
16:06 rojasm FC1CCCCC1
16:07 rojasm It doesn't create the resonance anymore!
16:07 rojasm am I making some error?
16:10 egonw right
16:10 egonw because it was not in the input
16:11 egonw you need to do a DeduceBondOrders run
16:11 egonw but this is exactly why you should not use the SmilesParser in tests
16:11 rojasm I know. But it was an old test
16:12 rojasm that I wanted to look first the results
16:12 rojasm ok
16:14 egonw rojasm: use the cdk1.0.x/ trunk for that
16:14 rojasm Not necessary I am removing all test with smilesParser
16:15 rojasm I am not in StructuerGenerator
16:15 rojasm tha last
16:15 egonw :)
16:15 rojasm StructureResonanceGenerator
16:15 egonw ok, gonna make dinner now
16:16 egonw will be online tonight, cu later (after 20:00)
16:16 rojasm cu later
17:45 CIA-18 cdk: egonw * r9997 /branches/egonw/jchempaint/.project: Keep Eclipse happy
17:49 CIA-18 cdk: egonw * r9998 /branches/egonw/jchempaint/ (29 files in 14 dirs): Removed first bits of old code
17:57 edrin joined #cdk
18:17 CIA-18 cdk: miguelrojasch * r9999 /branches/miguelrojasch/react​ion/src/org/openscience/cdk/ (43 files in 5 dirs): adapted tests to junit4. Correction of bugs. Not finished the revision yet.
18:26 rojasm egonw: if I don't use QueryAtomContainerCreator in UniversalIsomorphismTester.isIsomorph to compore two structure the method looks for all possible differences?
18:55 egonw the default only looks at element symbol and bond order... and if both are aromatic, that's a match too
18:55 CIA-18 cdk: miguelrojasch * r10000 /branches/miguelrojasch/react​ion/src/org/openscience/cdk/ (2 files in 2 dirs): Removed Flags. They were not used anymore.
18:55 rojasm ups!
18:55 rojasm ok
18:56 rojasm oooo r1000
18:56 rojasm ok. Than I need to created someone that match all
18:57 egonw yohoo !
18:57 egonw there's the big 1 oh oh oh oh !
18:57 rojasm Finally. Nobody wanted to do it :)
18:57 egonw :)
19:00 rojasm And one of the test I wrote "C=N-C=C => {(no possible) C=[N+]=C-[C-]}"
19:00 rojasm that is true?
19:00 rojasm is it not possible to have N sp2
19:00 rojasm sorry N sp1
19:00 egonw sounds acceptable
19:01 rojasm ok
19:02 CIA-18 cdk: miguelrojasch * r10001 /branches/miguelrojasch/reaction/src/org/​openscience/cdk/test/reaction/type/Displa​cementChargeFromDonorReactionTest.java: Changed the test which confirmed that C=N-C=C => {(no possible) C=[N+]=C-[C-]}
19:20 egonw http://chem-bla-ics.blogspot.com​/2008/02/10000-cdk-commits.html
19:23 rojasm Nice :)
19:29 egonw I'll update the branch, and see what fails...
19:29 egonw I've had a look at launchpad.net today
19:29 egonw is from Canonical, the guys who make (*)Ubuntu
19:29 CIA-18 cdk: miguelrojasch * r10002 /branches/miguelrojasch/reaction/src/o​rg/openscience/cdk/test/reaction/type/ (3 files): added FIXME REAC: to resolve conflicts
19:30 egonw which has a very nice tracking system for tasks...
19:30 egonw one can define dependencies...
19:30 egonw something I really miss with SF
19:30 egonw should we use that for the IP/reaction/charge related bugs...?
19:31 egonw define tasks, their deps (what must be fixed first), etc...?
19:31 egonw and assign them between us...
19:31 egonw each task can have a different priority, person who works on it
19:31 egonw and the ability to indicate how far the progress is...
19:43 egonw rojasm: see https://blueprints.launchpad.net/+about
19:54 rojasm but this lauchpad project is for all types of project or ubuntu projects?
19:55 egonw any project, AFAIK
19:55 egonw https://code.launchpad.net/+about
19:56 edrin joined #cdk
19:58 CIA-18 cdk: miguelrojasch * r10003 /branches/miguelrojasch/react​ion/src/org/openscience/cdk/ (23 files in 2 dirs): added checker if resulting atom type is reasonable with the method atMatcher.findMatchingAtomType.
20:00 egonw can't find the definate answer though
20:00 egonw but any project will do, AFAIK...
20:01 egonw are you interested? then we could give it a test go
20:01 egonw on behalf of the CDK project
20:01 rojasm yes of course
20:02 rojasm But I have no idea how :(
20:02 egonw we'll see...
20:02 egonw launchpad is pretty new to me too...
20:02 egonw i'll register the cdk project
20:06 egonw https://launchpad.net/cdk
20:06 egonw there we go
20:06 rojasm perfect
20:06 rojasm I am also registred
20:06 egonw good
20:07 egonw what account?
20:07 rojasm miguelrojasch
20:07 egonw comfy :)
20:07 egonw ok, excellent
20:07 egonw now we need to define 'blueprints'
20:07 egonw shall we try doing this top down?
20:07 egonw start with the ultimate goal?
20:07 egonw and work our way down to current tasks?
20:08 rojasm yes let try
20:08 egonw so, what is the ultimate goal?
20:08 egonw MS prediction?
20:08 rojasm ooo
20:08 rojasm to ambitious
20:09 egonw well... no requirements on how good it is...
20:09 egonw just that it works...
20:09 rojasm I would put prediction Ionization
20:09 rojasm more possible
20:09 egonw ok, let's start with that then :)
20:09 egonw that matches the current bug reports fine
20:11 egonw can you try to define that blueprint?
20:11 rojasm yes
20:11 egonw "IP prediction for molecules" or so
20:13 rojasm done
20:14 egonw mmm... don't see it...
20:14 rojasm https://blueprints.launchpad.net/cd​k/+spec/prediction-ionization/+edit
20:16 egonw oh, so it is...
20:16 egonw excellent :)
20:16 egonw apparently, one can define blueprints on a branch level too!
20:17 egonw ok, so what do we need for this to work, that is, what are the deps:
20:17 egonw 1. reaction-error-free
20:17 rojasm I have assigned you as well
20:17 egonw which depends on reaction-fully-unit-tested
20:17 egonw 3. charge-error-free
20:17 egonw which depends on:
20:18 egonw 4. charge-fully-unit-tested
20:18 egonw and we'll have 5. reaction-pmd
20:18 egonw and 6. charge-pmd
20:18 egonw does that make sense?
20:19 rojasm yes
20:20 egonw do you add the reaction ones?
20:20 egonw then I'll add the three charge ones
20:21 egonw "Make the charge module CDK stable"
20:21 rojasm I can not follow you. Because I don't see any thing of that!
20:22 egonw I'm using that as title for the new blueprint 'charge-error-free'
20:25 egonw https://blueprints.launchpad.net​/cdk/+spec/prediction-ionization
20:25 egonw added one of the deps
20:25 rojasm oo niche
20:25 rojasm nice
20:25 egonw yes, indeed :)
20:26 egonw title of charge-fully-unit-tested: "Implement unit tested for all code in the charge module"
20:28 egonw and the dependency tree grows
20:35 egonw rojasm: do a reload on the dependency page: https://blueprints.launchpad.n​et/cdk/+spec/charge-error-free
20:41 egonw https://launchpad.net/~cdk-developers/+members
20:45 rojasm egonw: I think I go at home. I am a bit tired
20:45 egonw ack, sure
20:45 egonw before you go...
20:45 egonw quick question...
20:45 rojasm yes
20:45 egonw what's the state of reaction?
20:46 egonw anything you want me to do at this moment?
20:46 rojasm yes
20:46 egonw tell me what I can do...
20:46 egonw I'll add the missing blueprints, at least...
20:47 egonw we'll do the assignment later...
20:47 rojasm I think I need to create the query for the complety molecule
20:47 rojasm with all properties
20:47 egonw why?
20:48 rojasm to be sure that the products are those I think
20:48 egonw that is... what do you need that for?
20:48 egonw ah... cehcking LP, single electrons, etc is not enough?
20:48 rojasm yes it is enought
20:48 rojasm enough
20:48 egonw thinking about it...
20:48 rojasm ok. Only this then
20:49 egonw not sure we can use the UIT for that...
20:49 egonw because at the atom-atom-matching level...
20:49 egonw it no longer knows about the full container, I think...
20:49 rojasm other is about the system class which linked differents reactions
20:49 egonw and since the atom does not refer to the LP, SE, or any IElectronContainer...
20:51 rojasm May be we could speak tomorrow abou that
20:51 rojasm cu tomorrow then
20:51 egonw not sure I'll be online all day...
20:51 egonw we'll see...
20:51 egonw have a good sleep!
20:51 rojasm ok
20:51 rojasm not problem
20:51 rojasm then other day
20:52 rojasm Thanks I have still 20 minunts to go at home
20:52 rojasm :(
20:52 egonw how so?
20:52 egonw you bike only works after ten?
20:52 egonw :)
20:52 rojasm I am in der university
20:53 rojasm but I go by walk
20:53 rojasm :(
20:53 egonw but I don't get why you can't leave right now...
20:53 egonw no bus involved?
20:53 rojasm I have no Idea
20:54 rojasm I will see
20:54 egonw never mind :)
20:54 rojasm cu

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