Camelia, the Perl 6 bug

IRC log for #bioclipse, 2009-02-20

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

All times shown according to UTC.

Time Nick Message
00:16 CIA-55 bioclipse: gilleain * r8929 /bioclipse2/trunk/plugins/net.bio​clipse.cdk.jchempaint/plugin.xml: Increas->Increase, Decreas->Decrease.
05:38 masak joined #bioclipse
06:55 egonw joined #bioclipse
07:05 egonw @tell jonalv please review that springsource.org.mozilla.javascript patch... it does not work here, as my Eclipse install does not have that plugin... so no clue where the 'set target platform' should get it from... neither does Eclipse...
07:05 zarah Consider it noted.
07:28 CIA-55 bioclipse: egonw * r8930 /bioclipse2/trunk/plugins/net.bioclipse.cd​k.business/src/net/bioclipse/cdk/business/ (CDKManager.java ICDKManager.java): Fixed spelling error in API (fixes #529)
07:28 zarah bug #529 | http://tinyurl.com/alrvf4
07:28 CIA-55 bioclipse: egonw * r8931 /bioclipse2/trunk/plugins/net.bioclipse.cdk.busine​ss/src/net/bioclipse/cdk/business/CDKManager.java: Replaced use of the deprecated StringBufferInputStream with the ByteArrayInputStream
08:04 olass joined #bioclipse
08:04 egonw olass: moin
08:05 egonw doing a major clean up of CDKManager.save...
08:05 egonw it's bad code
08:05 egonw nasty code
08:05 olass egonw++
08:05 egonw and breaks important functionality in the JCP editor
08:05 egonw I hope it will not break new things...
08:05 egonw so a new beta is 100% in order...
08:06 olass ok
08:06 egonw good news is...
08:06 egonw the PubChem plugin is getting along
08:06 olass nice
08:06 egonw it downloads not the mol files
08:06 egonw as in BC2
08:06 egonw as in BC1
08:06 egonw not 2
08:06 olass ok
08:06 egonw but the ASN-XML
08:06 egonw which is more accurate
08:06 egonw *and* they already open in JChemPaint
08:06 egonw but we cannot save in that format...
08:07 olass saving in CML is OK I guess
08:07 egonw that's OK, as we can refer to saveAs()
08:07 olass yup
08:07 egonw but this got me into this stupid saveMolecule() methods...
08:07 egonw basically the current JCPEditor is broken, when:
08:07 egonw a MDL molfile does not end in .mol
08:07 egonw and a CML molfile in .cml
08:07 olass oh, I have one request for saveAs, to have a combobox for chosing the file format to save to
08:07 egonw those are the majority...
08:08 egonw but particularly, it breaks for CML
08:08 egonw mol too though...
08:08 egonw as .mdl is not uncommon either
08:08 olass right
08:08 egonw olass: yes, I know...
08:08 egonw but not sure if I have time to write a new SaveAs dialog from scratch
08:08 olass ok
08:08 olass strange that the SaveAsDialog does not support that
08:10 egonw yes
08:13 olass egonw: the FileDialog with SWT.Save, what was the thing with that?
08:13 egonw saving outside the workspace
08:13 olass it allows for extensions...
08:13 egonw which will be wildly confusing...
08:13 olass ok
08:13 Gpox joined #bioclipse
08:13 olass agreed
08:14 egonw hi Gpox
08:14 olass hi Gpox
08:14 egonw Gpox: doing that saveAs() stuff for JCPeditor now...
08:14 olass hi Gpox
08:14 zarah hi Gpox
08:14 Gpox hi
08:14 zarah ni hao Gpox
08:19 masak ni hao Gpox
08:27 egonw 552 bug reports already...
08:28 egonw I think we like bugzilla :)
08:28 egonw #552
08:28 egonw bug #552 ?
08:29 olass bug #552
08:29 olass @pz 552
08:29 zarah http://pele.farmbio.uu.se/cgi-b​in/bugzilla/show_bug.cgi?id=552
08:30 masak I made that function intentionally narrow.
08:30 masak I don't want her to blabber URLs as soon as people talk about bugs.
08:30 olass ok
08:30 egonw yes, makes sense
08:30 olass so what works?
08:30 masak zarah: bug 552
08:30 zarah http://pele.farmbio.uu.se/cgi-b​in/bugzilla/show_bug.cgi?id=552
08:30 egonw #bla in commit messages
08:30 egonw now I remember
08:30 olass in commot messages, how do I write?
08:31 masak olass: bug #nnn
08:31 masak with or without the '#'
08:31 olass ok
08:31 masak olass: or fixes #nnn
08:31 olass solves #xxx too?
08:31 masak I think so.
08:31 olass ok, good
08:31 masak cf source :)
08:31 olass then I know
08:32 masak zarah++
08:32 olass @gg zarah bioclipse
08:32 zarah olass: http://irclog.perlgeek.de/text.pl%3Fch​annel%3Dbioclipse%3Bdate%3D2008-07-04
08:32 masak hm :)
08:32 olass masak: do we have a zarah wiki page?
08:32 olass with commands etc?
08:32 masak olass: no.
08:32 masak good idea though.
08:32 olass maybe a short biography of zarah?
08:32 masak http://bioclipse.svn.sourceforge​.net/viewvc/bioclipse/bot/trunk/
08:32 zarah masak's link is also http://tinyurl.com/cs82dd
08:32 masak olass: sure.
08:40 jonalv joined #bioclipse
08:40 egonw everyone: ping
08:41 egonw major commit comming up...
08:41 egonw will possible break things with saving chem files...
08:41 egonw please file directly if you note something odd...
08:41 egonw jonalv: you got messages
08:42 jonalv @tell me
08:42 zarah jonalv: 1 h 36 m 23 s ago, egonw said please review that springsource.org.mozilla.javascript patch... it does not work here, as my Eclipse install does not have that plugin... so no clue where the 'set target platform' should get it from... neither does Eclipse...
08:42 egonw crap...
08:42 egonw umm... OK, the first commit was not intended to hit SVN...
08:42 egonw sorry about that...
08:42 jonalv egonw: I added that jar file to our target platform
08:42 egonw 8932, that is...
08:42 CIA-55 bioclipse: egonw * r8932 /bioclipse2/trunk/ (7 files in 7 dirs):
08:42 CIA-55 bioclipse: Revert "switched to bundled version of rhino (1.7.0RC1) which resides in the target platform."
08:42 CIA-55 bioclipse: This reverts commit 08fa6c8a5095d866dc6e1f871c6f7013c2682d1e.
08:42 egonw jonalv: yes... I think I understand what it is supposed to do...
08:42 egonw but it does not work here...
08:43 egonw so, I have to revert that one all the time...
08:43 CIA-55 bioclipse: egonw * r8933 /bioclipse2/trunk/plugins/net.bioclipse.cd​k.business/src/net/bioclipse/cdk/business/ (CDKManager.java ICDKManager.java): Added a method to convert IContentType into IChemFormat
08:43 jonalv egonw: you are not here :(
08:43 egonw and remove the revert upon 'git svn dcommit'... which I had not forgotten :(
08:44 egonw and remove the revert upon 'git svn dcommit'... which I had noW forgotten :(
08:44 egonw jonalv: I *an* here...
08:44 egonw jonalv: I *am* here...
08:44 CIA-55 bioclipse: egonw * r8934 /bioclipse2/trunk/plugins/ (11 files in 8 dirs):
08:44 CIA-55 bioclipse: Fixed chem format specs: never use file extensions, but IChemFormat instead Removed concept of file extensions. If
08:44 CIA-55 bioclipse: you want to know reasonable extensions, try use the IChemFormat API
08:45 jonalv egonw is Egon Willighagen (n=kvirc@81-234-87-76-no38.tbcn.telia.com)
08:45 jonalv you are on telia...
08:45 egonw yes, that's here
08:45 egonw jonalv: anyways...
08:46 jonalv so I can't just walk over see your problem now can I?
08:46 egonw the problem is that I get Access restriction problems
08:46 jonalv :)
08:46 egonw yes, please, very please
08:46 egonw but it's indeed a bit unrealistic...
08:46 jonalv egonw: please what?
08:46 egonw it's just 10km
08:46 jonalv right...
08:46 egonw ok...
08:46 egonw let's just do it via IRC, ok?
08:46 jonalv acceses restrictions?
08:47 jonalv egonw: yea :)
08:48 egonw so, what info do you need?
08:48 jonalv egonw: could you look if the jar file is in your target platform?
08:48 jonalv it should be there...
08:49 egonw yes, it seems to be...
08:50 egonw so, the access restriction is not about a missing jar...
08:50 egonw OK
08:50 egonw copy that
08:50 jonalv egonw: what exactly is the access restriction?
08:50 egonw on Context
08:50 egonw and another class
08:50 CIA-55 bioclipse: egonw * r8935 /bioclipse2/trunk/ (7 files in 7 dirs):
08:50 CIA-55 bioclipse: Revert "Revert "switched to bundled version of rhino (1.7.0RC1) which resides in the target platform.""
08:50 CIA-55 bioclipse: This reverts commit e970d3db62df83c85c129588e35901f1f9599061.
08:50 CIA-55 bioclipse: Oops... that first revert was not intended for SVN, just for getting BC2 booting here...
08:50 egonw coming from org.mozilla.javascript
08:50 egonw Scriptable
08:53 jonalv egonw: I don't understand
08:53 jonalv egonw: where do you get the error and what is the error?
08:53 egonw OK, the compile errors actually seem to be gone now, but I still get:
08:53 egonw !SUBENTRY 2 net.bioclipse.scripting 2 0 2009-02-20 09:53:17.832
08:53 egonw !MESSAGE Missing required bundle com.springsource.org.mozilla.javascript_1.7.0.
08:54 jonalv egonw: did you just revert my commit?
08:54 egonw yes
08:54 egonw and reverted the revert...
08:54 * jonalv just realised that it's a bit complicated to search for error now...
08:54 egonw did you read what I typed here on IRC?
08:55 jonalv egonw: yea and I thougth you said you reverted the revert
08:55 egonw [09:42:06] <egonw> crap...
08:55 egonw [09:42:16] <egonw> umm... OK, the first commit was not intended to hit SVN...
08:55 egonw [09:42:19] <egonw> sorry about that...
08:55 jonalv egonw: we seem to have some lag in c'est system...
08:55 egonw yes 8935
08:55 masak jonalv: please never use "c'est" that way again...
08:56 jonalv egonw: You have of coursed reset target p[latform?
08:56 jonalv masak: did I spell it wrong?
08:56 egonw sure
08:56 masak jonalv: no, you spelled it OK.
08:56 egonw a million times...
08:56 egonw actually went an hour later to bed, just because I could not boot BC2
08:56 egonw like now
08:56 jonalv egonw: okey trying the same thing on masak+++ s machine now
08:56 jonalv hold on
09:00 egonw Gpox, olass: you can boot the current SVN version too?
09:00 olass dunno
09:00 egonw Gpox, olass: 8935 that is...
09:01 masak "java.lang.RuntimeException: No application id has been found."
09:02 egonw !MESSAGE Product net.bioclipse.ui.product could not be found.
09:02 egonw similar...
09:03 egonw yeah, and further down in the log, I have that message too...
09:03 egonw masak: so, you can't boot either?
09:03 masak correct. I can't start Bioclipse.
09:04 * olass is updating
09:04 masak oh, now it can haz workz.
09:04 masak "Add missing plugins" did the trick.
09:04 jonalv Starting from features is _so_ broken
09:05 jonalv egonw: okey now it runs on masaks machine. Let's try yours now
09:05 olass jonalv: probably you did not update the bioclipse_feature with your added deps
09:05 egonw ok, good
09:05 jonalv olass: probably not no
09:06 olass please do and commit
09:06 jonalv but it works on my machine...
09:06 olass from features?
09:06 olass on clear workspace?
09:06 jonalv on clear workspace I don't know
09:06 olass with new run config?
09:06 jonalv I am gonna try from features now
09:06 olass else you run an old run config
09:06 olass (if it exists)
09:07 olass they are not synced
09:07 jonalv it might be that I didn't save when I removed all features...
09:07 jonalv borken shit
09:07 olass jonalv: that is not what Im saying
09:07 jonalv olass: not?
09:07 olass run config does *no* know about features
09:07 jonalv olass: what are you saying then?
09:07 olass remove the rin config alltogether
09:07 olass then open product
09:07 olass run
09:07 olass that craetes a new run config
09:08 jonalv *cheminfo
09:08 egonw jonalv: no worries...
09:08 egonw I'm sure my patches break a thing here and there...
09:08 shk3 joined #bioclipse
09:08 olass it MUST create a fully working run config like this, else export will not work
09:08 egonw having the test suite properly back online would be nice
09:08 olass jonalv: you with me?
09:08 jonalv olass: I think so, hang on
09:09 olass if you have a run config, start from features will not overwrite it
09:09 olass the sync there is not 100%
09:09 olass but
09:09 olass it WILL tell you if some plugins have missing deps
09:09 olass then it is detective work and manual updating of the features
09:10 olass I've done this a lot...
09:11 CIA-55 bioclipse: jonalv * r8936 /bioclipse2/trunk/features/net.​bioclipse_feature/feature.xml: added com.springsource.org.mozilla.javascript to net.bioclipse_feature
09:11 * jonalv hopes thinks will work now
09:12 * egonw is trying
09:13 olass yes?
09:13 olass no?
09:13 jonalv all this just because edrin want's to use his js-editor in Eclipse
09:13 olass jonalv: works?
09:13 jonalv and Eclipse already have a js-ediotor right?
09:13 olass well, yes and no
09:13 olass yes: yes they have
09:14 olass no: it's not suited for RCP
09:14 jonalv olass: what works?
09:14 olass runing from features?
09:14 jonalv olass: for me it does, but I am not sure that's eniough...
09:14 olass egonw?
09:14 olass works?
09:15 egonw ...
09:15 olass seems to work here anyways
09:15 egonw yes, seems to work here
09:15 olass ok, good
09:15 egonw jonalv: thanx!
09:15 jonalv phew :)
09:16 olass features is a bit manual work indeed
09:16 * jonalv does not like it when people turn up here and say that they can't start Bioclipse...
09:16 egonw jonalv: :)
09:16 olass jonakv: then please keep features in sync with run config
09:16 jonalv olass: yea I learned that today
09:16 egonw OK, save is confirmed broken now...
09:16 olass matchingAtoms.add( 1 );
09:16 olass good
09:17 * egonw goes into bug fix mode
09:17 olass ?
09:17 * olass needs a haircut
09:17 * jonalv too
09:19 edrin joined #bioclipse
09:19 edrin hi
09:19 zarah hello edrin, you fantastic person you
09:19 egonw hi edrin
09:19 edrin hi egonw
09:19 jonalv hi edrin
09:19 edrin hi jonalv
09:19 masak hi edirn
09:19 masak s/ir/ri/
09:19 jonalv edrin: ETA on removal of "run in main thread"
09:19 jonalv ?
09:19 jonalv :)
09:19 edrin hi masak
09:20 edrin jonalv: the run in main thread must be removed, right
09:20 jonalv edrin: when do you think you will have time?
09:21 edrin jonalv: at weekend
09:21 jonalv edrin: I am planning to give your js-editor a test run today and report everything I find as bugs
09:22 jonalv edrin: does that sound okey?
09:22 edrin i will disabled the run in main thread function, instead I will add a "run with its own console"?
09:22 edrin jonalv:  yes, sure
09:22 jonalv edrin: I honestly only want one way of running scripts
09:22 edrin jonalv: ic
09:23 jonalv scripts are complicated enough as they are. Adding different ways of running them just makes people wonder how to use it
09:23 jonalv edrin: And I don't see any good reasons for doing that
09:24 edrin jonalv: but in any case the way of running scripts should be really different from a .js file and jsconsole. a js file must run at once, in one context, a js console fires commands each line...
09:24 jonalv egonw: do you know anything about cdk.getSMILES?
09:24 egonw sure
09:24 jonalv egonw: it seems to be gone...
09:24 jonalv edrin: here we go again... :)
09:24 egonw jonalv: sorry?
09:24 egonw what do you mean?
09:24 egonw done?
09:25 egonw I think it's part of CDKMolecule
09:25 egonw not CDKManager
09:25 egonw not?
09:25 jonalv egonw: I can't find the method cdk.getSMILES
09:25 edrin jonalv: to sum up, we must keep different run methods for js console and js editor, as the javascirpt commands are in fact run differently
09:25 jonalv egonw: it was on both earlier... There is a js-script in example data that depens on it...
09:26 egonw yeah, indeed...
09:26 egonw mom
09:26 jonalv edrin: take that up with masak. What I said is that I only want one way of running a js-file.
09:26 egonw jonalv: fromSMILES
09:26 edrin jonalv: in the js console you run each command line step by step in a rhino context, the editor must run it on one context at once, this is important to get the line number of errors
09:26 jonalv egonw: I could rewrite the js-script...
09:26 egonw cdk.fromSMILES("CCC")
09:26 edrin jonalv: they are both using rhino
09:27 masak edrin: I agree about the difference you just explained.
09:27 jonalv egonw: that is not what I want. I want cdk.getSMILES(cdkMolecule)
09:27 masak edrin: but I don't think that's reason enough to do it twice in almost the same way.
09:27 masak I think it should be done once, centrally.
09:27 egonw jonalv: don't know that one
09:27 egonw there should be cdkMolecule.getSMILES
09:28 jonalv egonw: yes there is. But once upon a time both existed. If you don't think that was good then let's not put it back...
09:29 egonw I have never ever seen cdk.getSMILES(ICDKMolecule)
09:29 jonalv edrin: print is not working from you editor...
09:29 egonw and I agree that it is not good to put back in
09:29 jonalv egonw: then let's not :)
09:31 jonalv edrin: This is _exactly_ why you shopuldn't copy code. Things go broken. The print() method is not in your version of the Bioclipse js-environment. Down this path lies madness...
09:32 edrin jonalv: to be honest, running a script and/or command with rhino is really no magic and code of "~3 lines", when you remove the specific error reporting.
09:32 edrin As the error reporting + rhino context usage is different, as it is for jsconsole and jseditor, there is almost no code duplication.
09:33 edrin And I also really think that the jseditor should be responsible to run its scripts and the jsconsole should be responsible to run its commands.
09:33 edrin There is the possibility to make both things together one plugin, and I do not really have a problem merging my code with the javascript plugin, but i truly think that you are building one big core that aims to be the "one only" API to be used, everything must depend on it.
09:33 jonalv edrin: I hear what you are telling me but it does not change thae fact that you forgot to add the print method. I just wonder what other stuff will turn up...
09:34 edrin jonalv: but i agree that the way of adding managers must be done unified, to exclude this
09:34 edrin jonalv: the print method prints to the jsconsole, right
09:34 edrin ?
09:34 jonalv edrin: I guess
09:34 jonalv edrin: most importantly it prints
09:35 edrin jonalv: what is the method, just print() ?
09:35 edrin or js.print() ?
09:35 jonalv edrin: right now the example js-scripts in example data are not working because they are using print...
09:35 edrin jonalv: oh, i did not know about these examples...
09:35 jonalv edrin: the method I am missing is print() but it is linked to js.print()
09:35 edrin ic
09:36 edrin js.print() should work
09:36 jonalv but I am afraid that more things might pop up...
09:36 edrin jonalv: yes
09:37 jonalv you and masak really should write one base for setting up the Bioclispe JS-environment
09:37 edrin jonalv: I wasn't aware about that you have these test scripts. where are they located?
09:37 jonalv in the Sample Data
09:37 jonalv edrin: there is a bug assigned to you regarding them...
09:38 jonalv edrin: do you read your bugs?
09:38 jonalv at least it has been assigned to you...
09:38 edrin jonalv: yes, we can do this of course, but i want one thing taken care of: the plugins architecture should not end up in a circulus vitiosus of dependencies.
09:39 jonalv edrin: circular deps are not just something that is unwanted they are strictly forbidden by Eclipse so don't worry about that
09:39 edrin jonalv: yes, i interpeted this i should write test scripts, not to test the scripts
09:39 jonalv edrin: ah yes sorry...
09:40 jonalv edrin: anyways, I am going to fix those two scripts no I think. But please add a few more to that example folder :)
09:40 olass edrin: it would be great if sampledata contained many sample scripts
09:40 jonalv egonw: ping
09:40 olass so users (like me) can cut and paste
09:40 edrin jonalv: that's exactly what I was saying 6 months ago, pls give up the "core" thing when it comes to specialized implementations that must not depend on each other
09:40 egonw jonalv: pong
09:40 jonalv egonw:
09:40 jonalv @pz 552
09:40 zarah http://pele.farmbio.uu.se/cgi-b​in/bugzilla/show_bug.cgi?id=552
09:41 jonalv egonw: I don't understand...
09:41 olass edrin: can you also add content assistance to the JSEditor?
09:41 olass I really would like to get ctrl+space showing commands...
09:41 egonw jonalv: which bit?
09:41 edrin olass: like . and then functions pop up?
09:41 olass yes
09:41 olass that would be REALLY good
09:42 olass with JS commands and manager commands showing up
09:42 jonalv egonw: most of it actually...
09:42 olass edrin: I demoed the console/edior/manager soup for the Cambridge people. They were very impressed!
09:42 egonw jonalv: who did the scripting EP then?
09:42 jonalv egonw: me..
09:42 egonw thought it was you
09:42 jonalv egonw: but I don't understand the problem. Do you?
09:42 edrin olass: this is almost impossible with javascript from an editor (as an editor per se), sorry. the way we have to go here is to add a record function to the jsconsole and then display the record in the editor...
09:43 egonw jonalv: yes
09:43 egonw the EP has two scriptContribution elements
09:43 edrin olass: good to hear
09:43 edrin masak: available?
09:43 masak edrin: yes.
09:44 * olass will be away to get a haircut
09:44 olass only time they had :-(
09:44 masak hair--
09:44 jonalv egonw: and that is not allowed?
09:44 egonw jonalv: and your EP does not allow it... though it does work
09:44 jonalv egonw: but it works?
09:44 egonw right
09:44 egonw indeed
09:44 jonalv egonw: so it works even thoiugh it's not allowed?
09:45 egonw yes
09:45 * jonalv finds that a very strange behavior...
09:45 jonalv egonw: please do the change if you know how to...
09:45 egonw olass: can you check @bug #552
09:45 egonw olass: does that plugin.xml snippet look OK to you?
09:46 edrin masak: I will tell you know how I think things must be done: in fact the way of getting a list with managers should be unified. this plugin should be named for example .businessprovider for example. other plugins can request a <List> of managers there for example...
09:46 masak ok.
09:46 olass egonw: cannot say
09:46 edrin masak: Jsconsole may then request stuff from it
09:46 olass this is jonalv's doing
09:46 egonw olass: how so? why not?
09:46 edrin the managers
09:46 masak edrin: or you could just make use of the js environment like the rest of us.
09:46 egonw sorry... no, not like that...
09:47 olass egonw: I don't know how the EP is read in code
09:47 egonw ola: like: is it uncommon to have two EP within one <extionsion> element
09:47 egonw ?
09:47 egonw I don't think so...
09:47 egonw Gpox: you are doing that too, not?
09:47 olass that is very common
09:47 edrin masak: this truly ends up in a circulus vitiosus of dependencies. if we do everything like this all the plugins we develop will becomes useless without the "core" stuff
09:47 egonw ok, good
09:48 egonw that's all I wanted to know...
09:48 olass but you can set the bounds of the sequence in teh EP
09:48 egonw jonalv: I'll 'fix' the scripting EP schemata then...
09:48 egonw olass: yes, I know
09:48 masak edrin: that's right, but I don't think it's as bad as you seem to think.
09:48 olass it is by default set to 1-1
09:48 edrin masak: i want to keep the possibility to run plugins in eclipse only, too
09:48 jonalv egonw: thanks
09:48 egonw right
09:48 olass ok, I'm off
09:48 egonw my suggestion is to remove that limit
09:48 edrin masak: I think it is really bad.
09:48 masak edrin: I know you do.
09:48 olass be back in 1 hour I guess
09:49 egonw jonalv: please comment that there is no reason to not have more elements, and reassign to me
09:49 edrin masak: bioclipse should extend eclipse with useful plugins that bring in helpful stuff for bio/info, it should not replace eclipse. If you start to try this
09:49 edrin this will fail
09:49 masak edrin: I hear you.
09:50 masak edrin: you're questioning a very fundamental design decision in Bioclipse.
09:50 jonalv edrin: I just don't think that you can run Bioclisep with out the net.bioclispe.core plugin
09:50 masak if you're right, we'll have to go make and undo it.
09:50 masak but we won't do that lightly.
09:50 edrin masak: at the moment we are a group of ~10? ppl that contribute to bc project. we are not an eclipse team of >100 developers?
09:50 masak (because I, for one, believe that it's worth it)
09:50 masak edrin: I know.
09:51 jonalv edrin: if we are to do things your way we would probably have to restart and release Bioclispe 3 in about a year from now...
09:51 * olass is looking forward to a summary of this discussion
09:51 egonw jonalv: make that 2
09:51 olass edrin, masak: please be pragmatic
09:51 olass this is NOT the last version of Bioclipse
09:51 masak it's not pragmatic to undo deep-rooted design decisions now.
09:51 edrin jonalv: no, code lives,
09:51 olass I agree
09:51 masak also, we _want_ those features.
09:51 edrin olass: exactly what i wanted to say
09:52 olass so be pragmatic for Bioclipse 2.0 right now
09:52 edrin masak: just because there was once this core idea with bc this does not mean we must insist on it
09:52 olass after 2.0, we can plan teh rest of the development
09:52 olass edrin: you could be right, but this is NOT the time to change core
09:52 edrin olass: yes, i just want to avoid to make everything more worse now
09:52 olass but I agree in some ways
09:53 masak edrin: I wish there was a way for you to suggest features and additions without uprooting what we already have. I don't think it's as bad as you say.
09:53 olass masak: I agree
09:53 * olass REALLY is gone now
09:53 masak it's perfectly normal for an application to have a core on which everything depends.
09:53 masak if we leave that design, we can't do spring, and we want that.
09:53 edrin if you inweave everything much more now, it gets worse and more difficult to undp
09:54 masak edrin: branch and do your own Bioclipse, then.
09:54 olass edrin: I think what you are discussing requires a branch
09:54 egonw edrin: the thing is... just eclipse does not do what we want to do in bioclipse
09:54 olass :-)
09:54 egonw so, we cannot depend on just eclipse
09:54 egonw and patching eclipse is non-trivial
09:54 edrin egonw: yes
09:54 olass there is no problem to depend on eclipse + bioclipse core
09:54 olass IMO
09:55 olass why should that be a problem?
09:55 masak olass++
09:55 olass edrin: I'd like a use case where this is a problem
09:55 egonw I think edrin's point is that JChemPaint cannot be run as a eclipse plugin now
09:55 egonw which would be very good
09:55 edrin i know what you mean, but please consider that there might be plugins that could run standalone without bc... but would also be nice to run them in bc
09:55 egonw actually, one of the Eclipse GSoC ideas I emailed...
09:55 olass egonw: agreed
09:55 olass that is because we require SPRING and a new target platform
09:56 olass I REALLY agree with this as well
09:56 olass it is problematic
09:56 edrin olass: egonw just said one
09:56 olass but should be solvable
09:56 olass we do not throw away eclipse
09:56 olass we extend it
09:56 masak agreed.
09:56 olass we just extend it a lot
09:56 edrin another plugin that would be cool to have available in eclipse per se, too is the jsconsole and/or the jseditor (if they are mature)
09:57 olass edrin: I think there already exists such a project in Eclipse
09:57 olass or similar at least
09:57 olass something with "eclipse monkey"
09:57 edrin olass: yes, but it is discontinues
09:57 edrin d
09:57 masak edrin: depends on what you mean by mature
09:57 olass right, now I'm late
09:57 edrin yes
09:57 egonw edrin: not all plugins we have in bioclipse do depend on bioclipse.core, I think...
09:57 olass edrin: we are working the last weeks of 2.0 now
09:57 egonw would be nice to have an overview of deps again...
09:57 olass edrin: please kep this in mind
09:57 olass ok, bye
09:57 egonw will look into some graphviz hacking
09:57 egonw olass: bye :)
09:57 edrin olass: i will take care everything will work
10:01 edrin masak: I wrote several program in the past. several of them had version 1, 2, 3... just because I realized at some point, there is something suboptimal on the base
10:01 masak edrin: I hear you.
10:02 masak I don't think Bioclipse is perfect in its current form.
10:02 edrin masak: from this i learned to do stuff in small chunks. changing the base does not kill everything then
10:02 masak but I also don't agree with your reasons for changing things.
10:02 edrin masak: what reason to change what thing?
10:03 masak edrin: your reasons to throw out spring and managers just to make Bioclipse like Eclipse.
10:04 edrin masak: btw. I think I remember that in the beginning you wanted to have the javascript thing to reside in the core plugin?
10:04 masak aye.
10:04 edrin but in fact it is good to have it in its own plugin
10:05 masak if you say so. :)
10:06 edrin masak: it's a dogma
10:06 edrin (from eclipse)
10:06 masak ok. then it must be right.
10:06 edrin believe me
10:06 masak I'm doing my best trying.
10:07 edrin masak: wouldn't it be cool to run your console in eclipse standalone?
10:07 masak no.
10:07 masak it would be js without all the managers.
10:08 edrin masak: this is different for the jseditor
10:08 masak yes.
10:08 edrin it would work pretty well
10:08 masak indeed.
10:08 masak but it's a slightly different product, then.
10:08 edrin this sentence i did not get
10:09 masak the js editor would have to be slightly different in Eclipse, because Eclipse doesn't make use of Spring and managers like Bioclipse does. this is what we're talking about.
10:09 masak so you'd have to package it differently, or something.
10:09 masak make it possible to turn off the Bioclipse-specific parts.
10:10 edrin masak: for example the webservice plugin, ok, i dont see a reason to have this in eclipse standalone, but currently it depends on *one* cdk object. i think i had to checkout ~100 plugins just to download a fasta file from the web via webservices
10:10 edrin masak: right now the jseditor runs standalone
10:10 masak edrin: yes, and that already causes problems.
10:11 edrin masak: not tested, and it would need two supporter plugins that just export js.jar and log4j
10:12 masak edrin: what you want to do is fundamentally against the Bioclipse (2.0) idea.
10:12 edrin but not against the eclipse idea
10:13 edrin and i think it is not the fundament of bioclipse either
10:13 jonalv edrin: no you need the new js bundle that we are using now
10:14 edrin jonalv: that's a supporter plugin
10:14 jonalv edrin: I managed to find a bundled js version
10:14 jonalv edrin: a "supporter plugin"?
10:14 edrin but that's what I am saying, we should do it like this
10:14 jonalv edrin: That bundle contains all js. I wouldn't call it a supporter. It's highly needed not just support
10:15 edrin we should have .jsconsole, .jseditor, .rhinojarprovider, .managerprovider
10:15 edrin this way you can just take .jseditor + rhinojarprovder plugin and run them standalone
10:15 masak edrin: have you looked at what the js env currently does?
10:15 jonalv edrin: I don't follow now. What's .jsconsole?
10:16 masak i.e. the class JsEnvironment
10:16 edrin jonalv: a plugin that uses rhino and provides a console for it. the managers it fetchs from the managerprovider plugin (if it is available)
10:16 jonalv edrin: aha
10:17 edrin you could run jsconsole without jseditor plugin
10:17 edrin and vice versa
10:17 jonalv edrin: constructing this lvl of abstraction would of course be very good. But it's not something that is easy to do. But it's probably a worthy goal to work towards
10:17 edrin the manager provider plugin is an optional plugin
10:18 jonalv edrin: that is a bit more than just one plugin btw...
10:19 edrin jonalv: yes, thanks that you agree. and this is exactly why I want to avoid to mix editor with console right now
10:19 edrin because this will make it much more difficult afterwords
10:19 masak edrin: I'm sorry you don't like our design. but sometimes I also get the feeling that you don't fully grasp it, so I'm reluctant to let you change it around and possibly mess it up.
10:20 edrin jonalv: right now it seems you try to centralize everything to a core, but I think we should agree that this is very risky and should be avoided in future. rather bc should provide stuff that is cool and *can* be used, but it must not be a dogma
10:21 masak edrin: no, I don't agree.
10:21 jonalv edrin, masak: I think you two should be able to make the scripting editor and scritping console into stand alone things. You will probably need another base plugin doing the js stuff which both depends on...
10:22 edrin jonalv, masak: right now you are ~3-5 developers in uppsala, what is in 5 years?
10:22 masak edrin: I don't like your tone.
10:22 edrin jonalv: that is what I want to achieve in some way
10:23 jonalv edrin: I think it's a good vision but I think olass have a point in saying that we need to be a bit pragmatic. We have already spent very much time building the core parts of Bioclispe 2
10:23 masak I think the scripting/manager system should be kept for better reasons than "It's too late to change before 2.0"...
10:24 edrin jonalv: yes, let's be pragmatic, what must be done to get stuff working?
10:24 jonalv yea the scripting / manager system _is_ Bioclispe 2
10:24 masak indeed.
10:24 jonalv + I am qute proud of it...
10:25 edrin jonalv: that's good
10:36 * egonw is struggling with a CDK limitation...
10:37 egonw to make Bioclipse's code base cleaner
10:37 egonw and better
10:38 jonalv egonw: so when will we switch to git? :)
10:38 egonw well, Ola is not here...
10:38 egonw let's do it now :)
10:38 jonalv yey!
10:38 jonalv :)
10:39 masak once you go git...
10:43 maclean joined #bioclipse
10:44 egonw hi maclean
10:44 maclean hi egonw
10:49 * jonalv heads out hunting for food
10:50 masak maclean: HAI
10:50 maclean hola
10:51 maclean this properties-on-highlight in jcp is very nice, but can be very slooooow
10:53 masak first make it work, then make it fast :)
10:53 maclean true dat
11:04 Gpox jcp dose not select the highlighted element when doing actions so i need to add it to the selection
11:04 Gpox @pz 326
11:04 zarah http://pele.farmbio.uu.se/cgi-b​in/bugzilla/show_bug.cgi?id=326
11:06 maclean okay - so...
11:07 maclean shift-click adding to selection is desirable
11:07 maclean I think you would need to be in select mode, though
11:08 Gpox ctrl-click on win / linux, cmd-click on mac. that is how you selet multiple items
11:09 maclean yes
11:09 maclean there's a java Toolkit method to get the platform-specific 'control' button
11:11 maclean Toolkit.getDefaultToolkit()​.getMenuShortcutKeyMask();
11:12 Gpox it would be nice to have it in all modes that dont add stuff.
11:13 maclean I don't understand that
11:13 maclean have the ctrl/cmd thing?
11:13 Gpox at least in move mode
11:15 Gpox i don't know exactly how it should work in inc/dec charge or chage atom mode
11:16 Gpox but i do know that you don't want to selecti a bunch of atoms then go to the toolbar and click and back again to chage, maby you use the popup menu for that though
11:16 olass joined #bioclipse
11:17 olass hi
11:17 zarah privet olass
11:17 * olass is back from haircut and lunch
11:18 egonw was your hair tasty?
11:22 olass sort of
11:23 olass 10:38 egonw well, Ola is not here...
11:23 olass 10:38 let's do it now :) <- egonw: I read logs...
11:23 olass :-)
11:23 egonw :)
11:23 egonw good one...
11:24 olass I will forulate a letter to bioclipse-devel about edrins points
11:24 egonw next time I'll make sure to use the no log option
11:24 egonw better...
11:24 egonw put it in the blog too...
11:24 olass no
11:24 egonw with a good catchy title...
11:24 olass wrong forum
11:24 egonw looking forward to hear about the Planet Eclipse has to say about it...
11:24 olass bioclipse-devel it is
11:24 olass egonw: you'd hear absolutely nothing
11:24 egonw OK if I put it in my RCP blog then?
11:24 olass whatve
11:25 olass this has to do with BC, not Eclipse
11:25 egonw so that it would still hit Planet Eclipse...
11:25 olass our design is not important for others
11:25 egonw huh? I'd say it has all to do with Eclipse...
11:25 olass so noone else would care
11:25 olass no
11:25 egonw I beg to differ...
11:25 olass nothing with Eclipse
11:25 olass ok
11:25 olass I donät mind
11:25 egonw I think that's the whole reason we are on that planet
11:25 olass I thin not
11:25 egonw because they like to hear what bioclipse is doing...
11:25 olass we can say what we do
11:26 egonw like overcoming limits of the current Eclipse design...
11:26 olass I don't like to blog about our internal design issues
11:26 olass it is our decision
11:26 olass to run our own target platform has this effect
11:26 olass we knew about it
11:26 olass so it is our decision
11:26 olass and our design
11:26 egonw define 'internal' in an open source project...
11:27 olass we deliberately are not supporting our plugins running in Eclipse
11:27 egonw yes, surely it was a deliberate choice...
11:27 olass that is not an Eclipse shortcoming
11:27 olass it is our design
11:27 egonw a blog won't bring it up for discussion...
11:27 olass egonw: feel free to forward my mail in your blog
11:27 egonw olass: if Eclipse would have allowed managers like we have now...
11:27 egonw you would still have used a private design??
11:28 egonw that does not make sense
11:28 olass managers are good for some things, not for others
11:28 stain what are these managers..?
11:28 egonw sure
11:28 stain (forgive me for asking as a non-internal)
11:28 olass our business logic
11:28 olass we write business logic in managers
11:28 olass and use Spring to inject this in our scripting language
11:28 olass hence we build our DSL dynamically
11:29 olass with AOP
11:29 maclean I think it is inevitable that Bioclipse will do things differently to Eclipse.
11:29 egonw which is cool...
11:29 olass maclean: I agree
11:29 maclean Eclipse is an IDE, Bioclipse is not.
11:29 olass will write an email now
11:29 egonw so later we'll have a taverna manager, and do from our command line:
11:29 olass please coment later on this
11:29 egonw workflow = myexperiment.downloadWorkflow(someID)
11:29 egonw taverna.run(workflow)
11:30 egonw taverna.run(workflow, outputFolder)
11:30 stain and there myexperiemnt and taverna are the injected managers?
11:30 egonw no, not yet...
11:30 egonw but that's the idea...
11:30 egonw there are several plugins now, which provide functionality in this way...
11:30 egonw from different third party libs
11:30 stain and they provide a "manager" like that which can be accessed from the scripts?
11:31 egonw the injecting ensures that the things stay proper plugins...
11:31 egonw stain: right
11:31 egonw much like Taverna plugins
11:31 stain sounds good
11:31 stain it would also be interesting to do something similar internally in the Taverna workbench so we can support scripting/macros
11:32 egonw Bioclipse is becoming like a R for life sciences
11:32 stain but again it would ahve to support plugins providing various.. stuff
11:32 egonw indeed...
11:32 egonw if Taverna would have been RCP based...
11:32 egonw we could even share those manager plugins...
11:33 maclean one thing that strikes me as I backlog this mornings argu ....er... discussion is this : is there an intention to allow multiple plugins that do the same thing
11:33 egonw and have a JavaScript Node/Action...
11:33 egonw what's the name again...
11:33 egonw like the BeanShell one, but then allowing calling these managers from the workflow
11:34 egonw (but I know that's not going to happen anytime soon...)
11:34 egonw I wish I had as much money as the  Bill Gates Foundation...
11:34 maclean We could contact him, but he might send us some mosquitoes in a jar instead.
11:34 egonw :)
11:36 maclean oh, egonw - I'm getting something weird with the command-line jcp properties
11:36 egonw please file a bug report with the script
11:36 maclean ok
11:36 egonw almost done with the refactoring of saving chem files in BC2
11:56 CIA-55 bioclipse: egonw * r8937 /bioclipse2/trunk/plugins/net.bioclipse.cdk.busine​ss/src/net/bioclipse/cdk/business/CDKManager.java: Actually throw an exception when no suitable writer is found; set the writer, fixing the no output problem
11:57 CIA-55 bioclipse: egonw * r8938 /bioclipse2/trunk/cdk-externals​/trunk/org.openscience.cdk.io/ (2 files in 2 dirs): Updated to CDK 1.1.5.20090220, commit 14211: fixes instantiation of writers not in the org.openscience.cdk.io plugin, by allowing them to be registered
11:58 CIA-55 bioclipse: egonw * r8939 /bioclipse2/trunk/plugins/net.bioclipse.cdk.busine​ss/src/net/bioclipse/cdk/business/CDKManager.java: Register the CMLWriter, so that the writer factory knows about it
11:59 CIA-55 bioclipse: egonw * r8940 /bioclipse2/trunk/plugins/net.bioclips​e.cdk.jchempaint/META-INF/MANIFEST.MF: Added missing dependencies: fixes writing of CML and MDL V2000 molfiles
12:00 maclean heh. bug fixed through the process of writing a bug report.
12:01 CIA-55 bioclipse: gilleain * r8941 /bioclipse2/trunk/plugins/net.bioclipse.​cdk.jchempaint/src/net/bioclipse/cdk/jch​empaint/business/JChemPaintManager.java: Method was always returning 0.
12:06 egonw there are a few more...
12:07 maclean few more what?
12:09 egonw bugs
12:09 maclean what!? Nobody told me this! More bugs!
12:09 maclean :)
12:11 edrin joined #bioclipse
12:11 olass edirn: I am writing an email to bioclipse-devel
12:12 olass I think I have at least an attack plan for the previous discussions
12:12 olass edrin: please read and respond to this (not sent yet)
12:12 edrin ok
12:12 edrin sure
12:14 CIA-55 bioclipse: egonw * r8942 /bioclipse2/trunk/plugins/ (2 files in 2 dirs): Fixed writing of SDF files
12:14 edrin diagnostics now
12:14 edrin must go
12:14 edrin there
12:17 olass ok
12:18 CIA-55 bioclipse: jonalv * r8943 /bioclipse2/trunk/plugins/net.b​ioclipse.data/data/Javascripts/ (extractsmiles.js extractsmilesAndMass.js): fixed the two example js-scripts in example data
12:45 egonw all: ping
12:45 egonw anyone with experience with this exception:
12:45 egonw sealing exception: nu.xom is sealed...
12:45 jonalv pong :)
12:45 * jonalv has no idea
12:49 egonw java.lang.SecurityException: sealing violation: package nu.xom is sealed
12:49 jonalv egonw: what are you truing to do? :)
12:50 egonw run the CDKManagerTest
12:50 jonalv huh
12:50 jonalv strange...
12:53 jonalv @pz 170
12:53 zarah http://pele.farmbio.uu.se/cgi-b​in/bugzilla/show_bug.cgi?id=170
12:54 jonalv http://pele.farmbio.uu.se/cgi-bin/bugzilla/bugl​ist.cgi?query_format=advanced&amp;short_desc_ty​pe=allwordssubstr&amp;short_desc=&amp;long_desc​_type=substring&amp;long_desc=&amp;bug_file_loc​_type=allwordssubstr&amp;bug_file_loc=&amp;dead​linefrom=&amp;deadlineto=&amp;bug_status=NEW&am​p;bug_status=ASSIGNED&amp;bug_status=REOPENED&a​mp;emailassigned_to1=1&amp;emailtype1=substring​&amp;email1=&amp;emailassigned_to2=1&amp;emailr​eporter2=1&amp;emailcc2=1&amp;emailtype2=substr​ing&amp;email2=&amp;bugidtype=include&amp;bug_i​d=&amp;votes=1&amp;chfieldfrom=&amp;chf
12:54 zarah jonalv's link is also http://tinyurl.com/c4lopm
12:56 egonw jonalv: what's with that list?
12:57 jonalv egonw: that's all bugs that have at least one vote
12:57 jonalv intresting reading I think
12:57 egonw ah, ic...
12:57 egonw yes, indeed interesting...
12:59 egonw http://lists.ibiblio.org/pipermail/xo​m-interest/2007-February/003390.html
12:59 zarah egonw's link is also http://tinyurl.com/cvhvb6
12:59 egonw that's the root of the sealing problem :(
13:01 egonw blah...
13:10 CIA-55 bioclipse: egonw * r8944 /bioclipse2/trunk/ (4 files in 4 dirs): Placed CustomSerializer in a non-'nu.xom' package, to solve a access problem, caused by xom-1.1.jar being 'sealed'
13:18 egonw OK, saving properties in SD files is broken...
13:18 egonw the CDK API sucks
13:19 maclean :(
13:19 egonw another quick hack to clean up...
13:20 egonw anyway... most of CDKManagerTest is working again...
13:20 egonw test results are actually better then beta2
13:20 egonw ah, no
13:20 egonw on par...
13:20 maclean good
13:21 egonw yes
13:21 egonw as I did fix some important issues
13:21 egonw like editing a .mdl file
13:21 egonw which is a MDL molfile...
13:21 egonw but which would have been saved as CML...
13:21 olass great
13:22 egonw CDKManagerPluingTest has 5 extra fails, compared to beta2
13:23 egonw 17 fails in total right now
13:23 egonw quite a few are methods that are not supposed to be called...
13:24 egonw UnsupportedOperation exceptions
13:24 maclean oh
13:24 egonw and one IllegalArgument
13:24 egonw the 5 new ones seem to be like:
13:24 egonw Resource "/Virtual/bla" does not exist
13:25 maclean virtual is a bit of a headache sometimes
13:25 maclean @pele Virtual
13:25 zarah [498] Null pointer exception creating... http://tinyurl.com/com2zu
13:25 zarah [268] Failing test for spectrum          http://tinyurl.com/ccczo6
13:25 zarah [522] findUnusedFileName in WizardHel... http://tinyurl.com/bprltr
13:25 zarah [417] ui.open(ICDKMolecule mol): Open... http://tinyurl.com/cx3xaj
13:25 zarah [510] New CML file is not recognized ... http://tinyurl.com/b4vaso
13:25 zarah Entire list (7) at http://tinyurl.com/cect36
13:26 maclean 498, especially.
13:26 maclean although sam seems to have assigned that bug to himself...
13:27 egonw I reassigned to him for further details...
13:28 maclean I can reproduce it, btw.
13:30 egonw OK, good
13:30 egonw that gets us somewhere...
13:30 egonw so, OS/X related...
13:30 egonw well, who knew
13:30 maclean Sam is OSX, too?
13:31 egonw well, am I not the last of the Mohicans?
13:31 maclean Stefan is linux, too
13:32 egonw ah, ok
13:33 jonalv so who is Sam? And will he fix this bug? :)
13:33 olass Sam wil not
13:33 maclean He's the inchi guy, no?
13:33 olass he is not a bc developer
13:33 olass yup
13:34 CIA-55 bioclipse: egonw * r8945 /bioclipse2/trunk/plugins/net.bioclipse​.cdk.jchempaint/src/net/bioclipse/cdk/j​chempaint/editor/JChemPaintEditor.java: Only save() when the input format is supported, otherwise autoinitiate a saveAs(): data read from PubChem can now be opened in JCP, and saved with saveAs
13:34 egonw this is a rather cool patch
13:35 jonalv so perhaps someone should educate him and then reassign the bug perhaps?
13:35 egonw BC2 can now basically open any 2D coord file the CDK can read (it doesn't, but it can)...
13:35 masak egonw++
13:35 egonw as it will automatically see it cannot save, and ask for another file with saveAs()
13:35 jonalv egonw++ //a rather cool patch
13:36 maclean "it doesn't, but it can"?
13:36 maclean It doesn't but it /could/
13:36 maclean :)
13:36 egonw maclean: yes, that's what I wanted to type...
13:36 egonw and had in mind...
13:37 egonw my hands decides for the shorted 'can'
13:37 maclean heh
13:37 egonw that happens when you are typing on IRC and closing bugs in Mylyn at the same time :)
13:39 egonw OK, /me will go offline for the afternoon... and catch some sun...
13:39 egonw will be back tonite
13:39 maclean happy sunning
13:39 egonw olass: I'm good with things for the beta3
13:39 olass egonw: will not have time for beta3 today
13:39 egonw olass: OK, good
13:40 olass since target platfom changed, we need full rebuild of all
13:40 olass and new SF
13:40 olass just a bit too much to have time for
13:40 egonw ah, ic
13:40 egonw indeed
13:40 egonw well, it's good news for me...
13:40 egonw now I can implement the pubchem.search() too
13:40 olass how come?
13:40 olass oh, good
13:40 olass egonw: I'd like to request some nice scripts from you
13:41 egonw BTW, try pubchem.download(176, "/Virtual/aceticAcid.xml")
13:41 egonw olass: http://delicious.com/tag/bioclipse+gist ?
13:41 zarah egonw's link is also http://tinyurl.com/5gqjet
13:41 olass I know
13:41 olass more scripts anyway
13:41 olass like this:
13:41 jonalv olass: so new beta release on monday? :)
13:41 olass maybe
13:42 olass 9-15 I have class
13:42 olass ethics
13:42 jonalv oh okey
13:42 olass se cannot promise
13:42 olass but will try
13:42 olass will bring my laptop to class ;)
13:42 maclean Sounds unethical
13:43 olass right
13:43 olass might fail that class...
13:45 egonw bbl
13:47 maclean hmm the problem with Sam's bug seems to be that there is no URL/URI handler for the protocol memory: when using IResource.getLocationURI()
13:59 olass Gpox: any idea about that?
14:00 maclean presumably, other parts of bioclipse that handle paths to Virtual use some other mechanism...
14:05 jonalv zarah: tell edrin I once upon a time spent a very long time on getting JS-scripts to run as jobs. They would update the progressmonitor of that as they where running so any progress information (as text) from a long running method on a manager would be given to the user waiting for the method to end. I am a bit sad about the fact that this no longer works now that we are running your new editor. Could you have a look at that code and
14:05 jonalv try to incorproate it in your solution? I guess this is another reason to not reinvent the wheel...
14:05 zarah Consider it noted.
14:05 jonalv hm...
14:05 jonalv zarah: tell edrin try to incorproate it in your solution? I guess this is another reason to not
14:05 jonalv reinvent the wheel...
14:05 zarah Consider it noted.
14:05 jonalv gahg
14:05 Gpox the WizardHelper dose not support EFS it uses java.io.File to check for files
14:06 jonalv zarah: tell edrin reinvent the wheel...
14:06 zarah Consider it noted.
14:06 * jonalv is beginning to realise that this might not be the best way to do that...
14:10 olass maclean: currently, opening a molecule makes it a bit too small IMO (it does not fil the editor)
14:10 maclean grrr
14:10 olass is this a bug?
14:10 olass hmm
14:10 olass clicking the rake (recalculate)
14:11 olass makes the mol bigger but placed incorrectly outside editor, with no scrollbar
14:11 jonalv olass: I have filed that as a bug... :)
14:11 olass good
14:11 olass that's why I asked :-)
14:11 jonalv oh not the recalculate but the size
14:11 maclean for the moment : jcp.setBondWidth(...)
14:11 maclean ack, no, not that
14:11 olass maclean: just wondered if you were aware of this
14:12 maclean jcp.setBondLength()
14:12 maclean olass: very, very aware.
14:12 olass maclean++
14:12 maclean please to be reading jonalv's bug flood.
14:12 olass jonalv is a good tester
14:12 jonalv olass:  thanks :)
14:13 * jonalv is trying out the javascript stuff now
14:13 * jonalv is trying to write javascripts
14:13 olass jonalv++ //for being a good a tester and writing extensive bug reports
14:13 maclean @pele 436
14:13 zarah http://pele.farmbio.uu.se/cgi-b​in/bugzilla/show_bug.cgi?id=436
14:13 maclean (for olass)
14:13 olass oki
14:13 olass thx
14:38 edrin jonalv: "is beginning to realise that this might not be the best way to do that" -> focused on the job called methods?
14:38 zarah edrin: You have new messages. Write '/msg zarah @messages' to read them.
14:39 jonalv edrin: I was referring to the way of telling you about the problem...
14:40 jonalv zarah: tell egonw Regarding bug number 564. I guess I could do it sometime next week if you don't wanna do it? (It would be nice to have it in the next release...)
14:40 zarah Consider it noted.
14:40 edrin @msg
14:40 zarah edrin: 21 h 13 m 48 s ago, jonalv said I will use it...
14:40 zarah edrin: 35 m 10 s ago, jonalv said I once upon a time spent a very long time on getting JS-scripts to run as jobs. They would update the progressmonitor of that as they where running so any progress information (as text) from a long running method on a manager would be given to the user waiting for the method to end. I am a bit sad about the fact that this no longer works now that we are running your new editor. Could you
14:40 zarah edrin: 34 m 42 s ago, jonalv said try to incorproate it in your solution? I guess this is another reason to not
14:40 zarah edrin: 34 m 14 s ago, jonalv said reinvent the wheel...
14:40 edrin jonalv: arent scripts run as jobs?
14:41 jonalv edrin: I don't think so no
14:41 edrin at least in the "run in separate thread" ?
14:41 jonalv edrin: oh perhaps if I use the run in seperate thread way
14:41 edrin jonalv: the other "GUI thread way" will be removed
14:41 jonalv but they are not being given that jobs progressmonitor are they? :)
14:42 edrin they are
14:42 * jonalv realises that he just clicked the play button when he tried
14:42 jonalv it probably ran it in the main thread...
14:42 jonalv hang on...
14:44 jonalv edrin: no it does not
14:44 jonalv that is it's run as a job
14:44 jonalv but the progressmonitor is not updated
14:45 jonalv example is cdk.loadMolecules
14:45 jonalv that method does a call with text information for each new molecule it reads in
14:45 jonalv so you we see something like 5/500 molecules read
14:46 jonalv if it worked that is...
14:46 jonalv edrin: do you understand?
14:46 edrin yes
14:46 edrin who is doing that call?
14:46 edrin the manager functions?
14:46 jonalv edrin: which call?
14:46 edrin 5/500 molecules read
14:47 jonalv yea the manager method
14:47 edrin how can I hook this calls?
14:47 edrin these
14:47 edrin jonalv: i am using the monitor for having a the possibility to kill a script
14:48 jonalv CDKManager row 432
14:48 edrin you can check from within the script if the user canceled it
14:48 jonalv edrin: yea that is another thing. I am checking that it each manager call
14:49 jonalv so if it's about to do a manager call and the script is canceled it throws one of those canceled exceptions
14:49 jonalv or rather I did that...
14:49 jonalv but you are not uysing that code...
14:49 edrin no
14:49 edrin how can i actively hook these events?
14:49 jonalv AOP
14:50 edrin is there a function like
14:50 jonalv edrin: do you have the code infront of you?
14:50 edrin net.bioclipse.managerprovider.Ma​nagerTools.hookMonitorEvents();
14:50 edrin jonalv: no
14:51 jonalv oki I will find the relevant code and send you some svn links...
14:51 edrin ok
14:53 jonalv http://bioclipse.svn.sourceforge.net/viewvc/​bioclipse/bioclipse2/trunk/plugins/net.biocl​ipse.scripting/src/net/bioclipse/scripting/J​sThread.java?revision=8879&amp;view=markup
14:53 zarah jonalv's link is also http://tinyurl.com/cyw53n
14:54 jonalv that is the thread that ran the javascripts
14:54 jonalv it keeps a monitor
14:55 jonalv http://bioclipse.svn.sourceforge.net/viewvc/biocl​ipse/bioclipse2/trunk/plugins/net.bioclipse.scrip​ting/src/net/bioclipse/scripting/business/JSJobCr​eatorAdvice.java?revision=8718&amp;view=markup
14:55 zarah jonalv's link is also http://tinyurl.com/cemjgp
14:55 CIA-55 bioclipse: gilleain * r8946 /bioclipse2/trunk/plugins/net.bioclip​se.nmrshiftdb/src/net/bioclipse/nmrsh​iftdb/wizards/ElucidateWizard.java: Very, very (unreadable) long lines.
14:55 jonalv that class get's that monitor from the thread and sends it in to the method when it is called
14:56 jonalv edrin: this is yet another reason to way we don't like that you are redoing everything that we have already done...
15:01 jonalv edrin: do you understand that code?
15:01 jonalv edrin: are yoiu there? :)
15:01 edrin i am trying to read it
15:01 edrin unfortunately i am busy with others things, too
15:01 jonalv aha :)
15:03 jonalv @pz jmol cdk
15:04 zarah [532] Autoupdated fails with NullPoin... http://tinyurl.com/d34kco
15:04 zarah [563] jmol.load dies with NullPointer... http://tinyurl.com/byv2he
15:04 zarah [564] Make jmol manager use the new w... http://tinyurl.com/d4tw4q
15:04 zarah [420] Open ICDKMolecules in Jmol         http://tinyurl.com/c62zfn
15:04 zarah [450] Implement a JmolMolecule and im... http://tinyurl.com/dmjhod
15:04 zarah Entire list at http://tinyurl.com/c56cjm
15:07 edrin jonalv: that is indeed a nice feature, is every manager method hooked with such events? i guess this is some spring feature you are using to hook and distribute this information, right?
15:07 jonalv edrin: no not all managers yet. As for now only the cdk manager and the structuredb manager but I am redoing them one at a time...
15:08 jonalv edrin: yes Spring AOP is the magic bullet here :)
15:09 edrin jonalv: so if you want to cancel a script the manager call must throw an exceptions
15:10 CIA-55 bioclipse: goglepox * r8947 /bioclipse2/trunk/plugins/net.bioclipse.​chemoinformatics/src/net/bioclipse/chemo​informatics/wizards/WizardHelper.java: Changed fileExist to use eclipse files and not java.io.File fixes bug #498
15:10 zarah bug #498 | http://tinyurl.com/com2zu
15:10 jonalv edrin: yea that is the easiest way. Otherwise rhino will go on with the next statement...
15:10 edrin i am currently looking in the cdk manager, where can I see such a method?
15:10 jonalv edrin: the cool thing is that the script writer never sees thes Eclipse things...
15:11 jonalv I am pretty sure loadMolecuels should have it
15:11 jonalv edrin: row 435
15:12 jonalv edrin: but the Spring advice also checks the monitor before every manager call
15:12 jonalv edrin: so if the script uses a manager it will check the cancel status before running the manager method
15:13 edrin jonalv: yes, sure
15:14 jonalv But if we did this the way olass++ suggested than we would be fine I think. That is: your editor does not run the script but a Bioclispe plugin runs the actual script
15:14 edrin jonalv: ok, so how do you pass the monitor to spring?
15:14 jonalv the JSThread has the monitor remember?
15:14 olass edrin: can you please reply to the mail with your suggestions as well?
15:14 jonalv The advice just get's the monitor from the thread
15:15 jonalv edrin: row 46 in JSJobCreatorAdvice
15:16 jonalv that only works when we have one unique JS_Thread though
15:16 jonalv we need to redesign this when your great plan with unlimited nukmbers of JS-threads comes along :)
15:17 edrin jonalv: oh
15:17 edrin jonalv: ic
15:17 edrin jonalv: right now you can only run one script at a time?
15:17 jonalv edrin: but I just realised that it actually is Thread.currentThread so that shouldn't be too hard...
15:17 edrin or use the jsconsole
15:18 jonalv edrin: yes it all used the same JS-thread and the same JS-enviornment
15:18 edrin jonalv: ic
15:18 jonalv both the js-console and the editor used the same rhino instance back then...
15:18 edrin olass: ok. got no mail yet
15:19 edrin jonalv: they use the same?
15:19 olass mailed to bioclipse-devel
15:19 edrin jonalv: even the same context?
15:19 olass it IS here
15:19 jonalv edrin: yea they did. Now you are creating one of your own in you new editor...
15:19 jonalv so now you can't look at the variables created by a JS-script after you have run it using the js-console any longer...
15:20 edrin jonalv: independent from all other things we discussed, each script must have its own context. we cannot reuse a context for all scripts the are run (except one could reste it?)
15:20 edrin jonalv: of course not
15:21 edrin jonalv: if you do it like you said
15:21 edrin then
15:21 edrin example:
15:21 edrin script1.js: i = 5;
15:21 edrin script2.js: i = i + i;
15:22 edrin if you first run script1 and then script2 this should work and produce i = 10
15:22 edrin if you run script2.js without calling script1.js first
15:22 edrin this will fail because i was not defined
15:22 jonalv edrin: correct
15:22 edrin this behavior is not logical
15:22 jonalv edrin: yes it is
15:23 jonalv edrin: It might not be optimal though
15:23 jonalv edrin: but that is a completely different thing
15:24 jonalv anyways, we have already decided to be able to run multiple JS-scripts at the same time so that behcaior is not the one we are going to have in the future. I just explained how things used to work...
15:24 edrin yes
15:24 jonalv I still would like to get back the status feedback to the user when running JS-scripts containing manager methods that can support this...
15:26 edrin jonalv: yes, that's really a good feature I agree
15:26 jonalv good :)
15:27 jonalv I filed a bug about it for you...
15:29 olass edrin: did you agree on what I wrote in the email?
15:29 edrin currently reading
15:33 edrin olass: yes, that's exactly what I mean
15:34 olass good
15:34 olass do you see any problem with that design?
15:35 edrin one would have to think about if the JSExecution plugin could optionally import the bioclipse functionality or if it is required to create two of them, one for bc and one general
15:36 edrin because the general plugin's functionality would completely be part of the bioclipse additional functionality
15:36 olass edrin: I agree, this deserves some thoughts
15:37 olass edrin: please think about this and we'll continue the discussion next week
15:37 edrin yes
15:37 edrin but i really think this is the way to go
15:38 edrin olass: the JSExecution plugin
15:38 olass ok
15:38 olass edrin: from our point of view, this is post 2.0
15:39 edrin sure
15:39 olass just wanted to clarify :-)
15:39 edrin i can imagine... it is some work to do this
15:39 olass yes, and we will focus on Bioclipse-related stuff first
15:40 edrin jonalv: what do you suggest regarding the run script thing now? do you think all scripts should be run in the one jseditor/global context?
15:40 edrin olass: yes
15:40 jonalv olass:  I don't think this is post 2.0
15:40 jonalv I think this needs to be fixed asap
15:40 olass ok...
15:41 olass jonalv: what needs fixing?
15:41 jonalv The running of js files
15:41 * olass was refferring to the refactoring
15:41 olass ok
15:41 edrin especially we should get clear about the js context thing
15:41 edrin imo
15:41 edrin jonalv: yes
15:42 edrin olass: currently a problem would be that two independent scripts would share their context
15:42 jonalv @pz
15:42 zarah jonalv: Pelezilla is at http://bugs.bioclipse.net
15:42 olass jonalv: could you think about this over the weekend?
15:42 olass masak talked about one context for each scripts...
15:43 CIA-55 bioclipse: gilleain * r8948 /bioclipse2/trunk/plugins/net.bioclipse.cdk.jchem​paint/src/net/bioclipse/cdk/jchempaint/business/ (IJChemPaintManager.java JChemPaintManager.java): Get/Set zoom in console.
15:43 jonalv it's primarilty bugs:
15:43 edrin olass: currently there is one context per script, but only because currently jseditor runs rhino itself
15:43 jonalv @pz 562
15:43 zarah http://pele.farmbio.uu.se/cgi-b​in/bugzilla/show_bug.cgi?id=562
15:43 jonalv @pz 561
15:43 zarah http://pele.farmbio.uu.se/cgi-b​in/bugzilla/show_bug.cgi?id=561
15:45 jonalv I would actually prefer if the new editor just simply called the old code for running JS-scripts for now. And then this can be refactored in chunks when people have the time...
15:45 edrin "or call the old JavaScript running code" <- if i do this the scripts will share the context, right? this is not an option in my opinion... i would rather live without status msgs compared tp having non-logic scipts
15:46 jonalv edrin: Please don't call that non logical
15:46 jonalv it is very logical
15:46 jonalv it jsut means that you have one JS-environment. There is nothing unlogical with that
15:46 edrin jonalv: not really, because you can never know what a script does when you call it... because there might still something in the context
15:47 olass edrin: you mean unintuitive, not non-logical
15:47 olass edrin: something can be very logical, even if it is unintuitive
15:47 jonalv edrin: if you write a script that would fail because the variables are not initialzed in your variant then yes you can get strange behavior but don't write that kind of scritps
15:47 jonalv besides it's only for a short time period
15:47 olass yes
15:47 olass I agree
15:47 jonalv I really think we should try to fix this asap
15:47 edrin olass: in my opinion it is not logical, because in fact this would mean that each .js file must be considered a part of a big mega script.
15:48 olass then we close this whole as soon as possible
15:48 olass edrin: I agree that this in unintuitive
15:48 edrin olass: another disadvantage of using the existing method (shared context) is that you would not get the line number where an error is in the script
15:49 olass agree
15:49 olass unintuitive
15:49 olass but with one context it is perfectly logical
15:49 jonalv I am not saying that it is a good solution I am just saying that it is what we can get quickly
15:49 edrin olass: for example if you do:
15:49 olass edrin: I agree
15:49 olass edrin: but that is not the point
15:49 olass the point is that this is what we have time to do quickly
15:49 olass and THEN we fix it
15:49 jonalv edrin: but we must be able to get the line numbers the same way that you are getting them...
15:49 olass and we work out a better solution
15:50 olass edrin: ok?
15:50 edrin olass: if a reviewer calls our scripts in bioclipse one after the other... this is almost not predictable what will come out. they could fail, result in something strange, or evene open a worm hole...
15:50 olass edrin: ok?
15:50 olass edrin: do you read what I write?
15:50 jonalv edrin: lol what are you meddling with in your scripts? :)
15:51 maclean universe.open("wormhole")
15:51 olass edrin?
15:51 edrin jonalv: if you want to get the line numer this way it would maybe say: in line 16213 of script "neverendingstory.js"
15:51 jonalv edrin: no it most certailny would not. That is not how it works and you know that
15:51 olass edrin: can you maybe be more productive now?
15:52 edrin olass: i try
15:52 olass so far I have only heard complaints, not solutions
15:52 olass what do you propose?
15:52 olass that can be done over the weekend?
15:52 olass by you?
15:52 edrin one sec
15:52 edrin student is here
15:52 olass ok
15:54 edrin olass: i can add the top level print() method, all bc scripts should run then, right?
15:55 olass no idea
15:55 olass jonalv?
15:55 edrin jonalv: ?
15:55 jonalv edrin: those scritps are already fixed
15:56 jonalv I am just a bit dissapointed becasue I spent a huge amount of time getting that feedback to work and now it doesn't work again...
15:56 edrin about the Spring integration and hooking of manager calls/monitors: i would really discourage to enforce using them now, because the error line number + a context per script are more important than status reporting in the progress view IMO.
15:56 jonalv edrin: so it's not that important...
15:57 edrin jonalv: i like this thing
15:57 olass I like it too
15:57 olass :-)
15:57 jonalv edrin: I could probably hack that stuff into your code but I really want to know if everything will change again and my work become undone yet another time...
15:57 edrin jonalv: but can you also add a .executeWithItsOwnContext(String the_script); ?
15:58 jonalv edrin: I guess we could but let's not go there shall we? :)
15:58 jonalv that would just complicate things for our poor users...
15:59 edrin i am pretty sure that JSExecution plugin will exactly have this method ...
15:59 jonalv The problem is the following as I see it
15:59 jonalv I could add this code to your "JS-thread" but that would probably mean adding some Bioclispe deps which you don't want
15:59 jonalv so we need to create the run-js-file plugin and putthe code there
16:00 jonalv that might be a bit more work...
16:00 edrin jonalv: if the problem must be solved now you can add as many deps to jseditor as you want
16:00 jonalv edrin: perhaps you could create that plugin during the weekend?
16:00 edrin jonalv: run-js-file plugin is the best solution IMO
16:00 jonalv edrin: then you can make that plugin in two flavours. One with Bioclispe deps and one without...
16:00 edrin is not much work, and it can be a hack right now
16:01 edrin jonalv: ok, I can do this
16:01 jonalv if you write that plugin during the weekend then I could have a look at it next week
16:01 edrin jonalv: that's a good idea
16:01 * jonalv is tryhing to figure out whether we have forgotten something now...
16:02 edrin i dont know another prob right now
16:02 jonalv good, let's try this then. I will inform masak as well
16:02 edrin i can tell you about my experiment problems if you want
16:03 edrin jonalv: ok
16:03 jonalv :)
16:03 edrin jonalv: that's really a good solution i think
16:03 jonalv yes
16:04 jonalv olass++
16:04 olass for what?
16:05 maclean ...and peace returned to the land.
16:05 edrin I really must do some lab stuff now...
16:06 edrin olass: did you have a look on the current manuscript ?
16:06 olass started to read it
16:06 edrin egon wasnt so happy with it
16:06 edrin so I moved stuff a lot
16:06 olass but been so busy today
16:06 edrin into must be shortened
16:07 jonalv olass: you suggested this solution I think. Or did I do it? (me++) :)
16:07 olass haha
16:08 * jonalv is gonna head to swim now. cya
16:08 olass bye!
16:10 CIA-55 bioclipse: jonalv * r8949 /bioclipse2/trunk/plugins/net.bioc​lipse.cdk.business/src/net/bioclip​se/cdk/business/ICDKManager.java: added some more descriptions to man for cdk.saveCML
16:26 edrin left #bioclipse

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