Camelia, the Perl 6 bug

IRC log for #bioclipse, 2008-12-18

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

All times shown according to UTC.

Time Nick Message
07:24 egonw joined #bioclipse
08:12 Xavier joined #bioclipse
08:21 olass joined #bioclipse
08:21 olass hi
08:21 zarah hi olass
09:02 CIA-32 bioclipse: ospjuth * r8085 /bioclipse2/trunk/plugins/ne​t.bioclipse.cdk.debug.test/ (5 files in 3 dirs): Tests, pluginTests, and testdata for bug #187: sybyl atom typing on 232 mols SDF.
09:02 zarah bug #187 | http://tinyurl.com/4dy7v7
09:05 jonalv joined #bioclipse
09:10 jonalv hi
09:10 zarah hi jonalv
09:11 olass hi jonalv
09:13 jonalv olass: how are you? :)
09:14 olass ok I guess
09:14 olass really tired
09:14 olass hopefully not the bug :-)
09:14 olass but I feel ok
09:14 jonalv okey
09:14 olass and you?
09:14 olass not ill I hope?
09:14 egonw joined #bioclipse
09:15 jonalv yea I had a bit of whiplash (how do you spell that?) today since I jumped out of bed crazy early yesterday...
09:15 jonalv hi egonw
09:16 shk3 joined #bioclipse
09:17 egonw moin
09:17 zarah hi egonw
09:35 olass egonw: did you see bug 187?
09:35 olass I wrote a test for you
09:35 olass sybyl atom typing throws NPE
09:35 egonw yes, received it
09:35 olass good
09:35 olass was a bit of a showstopepr yesterday...
09:35 olass showstopper
09:35 egonw :(
09:36 egonw sorry about that
09:36 olass no worries
09:36 egonw how many molecules did it fail on?
09:36 olass glad we found it :-)
09:36 olass hmm
09:36 olass did not count
09:36 olass but was several in the file
09:36 olass I think it's the CDK atom typing that returns null
09:36 olass but not sure
09:36 egonw probably...
09:37 olass better we find the bugs now than later :-)
09:37 egonw if the CDK atom typing succeeds... it simply maps it to a Sybyl type
09:37 olass ok
09:37 egonw btw, it's not necessarily a bug...
09:37 egonw the SD file may have corrupt chemistry
09:37 olass oh?
09:37 olass true
09:37 egonw but will check that
09:37 olass how to deal with that case?
09:37 egonw anyway...
09:37 olass could it not set atom type to "undefined" in that way?
09:38 egonw file the structure as 'marked'
09:38 egonw and proceed with the rest
09:38 egonw olass: that's null...
09:38 olass right
09:38 egonw your code could set the atom type to 'X' then...
09:38 olass but the sybyl atom matcher throws NPE
09:38 olass this is not my code, it's in cdkdebug
09:38 olass and you wrote it :-)
09:38 egonw olass: right...
09:39 egonw the Sybyl atom typer should just return null too...
09:39 egonw that would actually be a bug...
09:39 olass is there no "undefined" sybyl atom type?
09:39 egonw there might be
09:39 olass how does openbabel treat these occurrencies?
09:39 egonw dunno
09:40 olass not that openbabel is perfect but anyway
09:40 olass ok, let me know what you find out
09:40 egonw maybe I should change that behavior...
09:40 egonw in CDK... 'no idea about that' == null
09:40 olass yeah, maybe
09:40 egonw so, when the typer has no clue, it returns null...
09:41 egonw but maybe 'X' is better, but that would sort of break CDK habits...
09:41 olass btter define "unknowns"
09:41 egonw well, I guess is why null is nicer...
09:41 egonw null != null
09:41 egonw whereas
09:41 egonw X == X
09:41 olass jonalv: never return null, right?
09:42 egonw yeah, heard that before...
09:42 olass well... there is a point with that
09:42 olass but you decide
09:42 egonw no, CDK Project decides...
09:42 egonw at least for CDK 1.2.x
09:42 olass right
09:42 olass but you have most influence
09:43 olass so your vote is heavy
09:43 egonw I'll propose a patch...
10:12 shk3 hi all
10:12 zarah hi shk3
10:12 zarah shk3: You have new messages. Write '@messages' to read them.
10:12 shk3 @messages
10:12 zarah shk3: 15 h 19 m 48 s ago, sneumann_ said to give http://msbi.ipb-halle.de/~sneumann/xset3.cmlspect a try
10:12 shk3 what are Bioclipse release plans?
10:14 shk3 will there be an alpha or beta?
10:16 egonw shk3: ask olass
10:16 egonw shk3: sneumann did not get speclipse to work with the current alpha on SF and your update site
10:17 shk3 he came across a bug which is fixed now
10:17 egonw ok, good
10:17 shk3 and it seems last time the export was with java 1.6
10:17 shk3 which is bad
10:18 egonw of the update site?
10:19 egonw main, speclipse, or experimental?
10:19 shk3 speclipse
10:21 egonw olass: ping
10:21 egonw shk3: would you like to put a new speclipse online?
10:22 egonw then we should just tag plugins... and release on the speclipse and experimental update site
10:22 shk3 well, it would need a new chemoinformatics featuer
10:22 egonw olass did not want to make a new alpha at this stage, I think, as the new JCP is not yet stable enough...
10:22 egonw shk3: yes, but you can include that this time on the speclipse update site too...
10:23 egonw or, otherwise, we
10:23 egonw or, otherwise, we'd use the experimental update site...
10:23 shk3 hm, that could be an option
10:23 egonw olass: ping
10:23 shk3 what do you mean with experimental update site?
10:23 egonw I'm still a bit tight up in the next CDK 1.2.x release...
10:24 egonw http://pele.farmbio.uu.se/experimental/update/
10:24 zarah egonw's link is also http://tinyurl.com/42bg7n
10:24 egonw where we can publish experimental features
10:24 shk3 ah
10:27 shk3 no idea how to export the chemoinformatics feature
10:27 edrin joined #bioclipse
10:27 shk3 are there any difficulties with that?
10:28 egonw no, I think it is just adding that feature to the update site
10:28 egonw make sure to also add the bioclipse_feature
10:28 egonw or maybe the update feature...
10:28 egonw let me check
10:28 egonw so that updates in the core plugins hit your update site too
10:28 edrin hi
10:28 zarah hi edrin
10:28 egonw hi edrin
10:28 edrin hi ege
10:29 edrin egonw:
10:30 edrin how is bioclipse proceeding ?
10:31 egonw steadily :)
10:32 edrin when do you release?
10:33 egonw ask olass
10:33 egonw no 2.0.0 release this week
10:35 egonw JCP is close, but too unstable at this moment...
10:39 edrin what is the state of the scripting plugin? is it working?
10:45 egonw masak was working on it...
10:45 egonw but his laptop was stolen...
10:45 egonw so, some delay there...
10:55 masak joined #bioclipse
11:06 egonw joined #bioclipse
11:14 shk3 btw, is it really Bioclipse code style not to have any lines between methods?
11:23 egonw one line, not?
11:34 shk3 well, jonathan refactored some of my classes to have none.
11:43 shk3 egonw: another question
11:43 shk3 just had a look at the expandable content for the sdf
11:43 shk3 it uses file extension to determine when it's used.
11:43 shk3 shouldn't it use the sdf content type?
11:44 shk3 or is that not possible with the expandable content?
11:59 olass shk3: I think Gpox wanted to get rid of the children in SDFiles
11:59 olass but we'll have to ask him
11:59 olass after christmas when he gets back
11:59 shk3 ok
12:00 olass there was problem syncing this with cached content in large files
12:09 olass shk3: next release of bc will be in january when JCP and new SDF is in better shape.
12:09 shk3 ok
12:10 shk3 then that will be date for next speclipse as well
12:10 olass sounds logical
12:10 olass let's coordinate this in january
12:10 shk3 I have some things i want to do in it anyway
12:10 olass ok, good
13:14 jonalv joined #bioclipse
13:14 jonalv oh hai
13:15 zarah hi jonalv
13:18 masak hi zarah
13:24 jonalv zarah: pele
13:24 egonw brb
13:24 zarah jonalv: Pelezilla is at http://pele.farmbio.uu.se/​cgi-bin/bugzilla/index.cgi
13:24 masak botsnack
13:25 zarah ;)
13:33 egonw joined #bioclipse
14:25 sneumann hi shk3,
14:26 sneumann got your mail about multiple spectra in cmlspect
14:26 sneumann For MS runs with up to 1800 spectra, the interface will be a challenge ;-)
14:26 sneumann But then "normal" use case will be
14:26 sneumann in analogy to annotating NMR spectra
14:27 sneumann to have single (or few) MS2 spectra for one compound
14:32 shk3 then you would need a molecule in the file and use specmol editor
14:32 shk3 not sure how this behaves with 1800 spectra, though
14:37 olass it is a challenge to construct GUIs for large number of entities
14:53 egonw joined #bioclipse
15:11 egonw olass: updated, and now running the plugin test for cdkx
15:11 olass great
15:12 jonalv egonw: you mean xkcd?
15:12 jonalv @gg xkcd
15:12 zarah jonalv: http://xkcd.com/
15:13 masak jonalv: man, that joke never gets old, does it? :P
15:13 masak (@gg)++
15:32 shk3 egonw: will you come to hinxton in january?
15:34 egonw very possibly
15:35 shk3 ah
15:35 shk3 should the copy&paste stuff also be part of the chemmodelrelay?
15:35 shk3 that relates to jcp...
15:35 egonw if we can generalize it enough, yes
15:44 shk3 well, it looks like the java copy&paste implementations are quite general
15:51 shk3 olass: thanks for the link to the contenttype doc
15:51 shk3 didn't find that myself :-(
15:58 olass shk3: hth
15:59 shk3 sounds like
15:59 jonalv @pele services
15:59 zarah [191] Create new sdf file + open said... http://tinyurl.com/4dshlc
15:59 zarah [192] Unable to create menu item "net... http://tinyurl.com/4sxcns
15:59 zarah [20] Services are not recorded           http://tinyurl.com/5b4ydb
15:59 zarah Entire list at http://tinyurl.com/4c35ta
16:16 shk3 jonalv: bug 191 is a known problem. It seems that sdfs don't work in virtual, see bug 180
16:16 jonalv shk3: oh intresting
16:16 jonalv @pz 180
16:17 zarah http://tinyurl.com/3w2r8n
16:24 edrin left #bioclipse
16:51 CIA-32 bioclipse: shk3 * r8086 /bioclipse2/trunk/plugins/net.bioclipse.chemoinfo​rmatics/src/net/bioclipse/chemoinformatics/conten​tlabelproviders/MoleculeFileContentProvider.java: using content types
16:57 sneumann bye
17:18 masak left #bioclipse
17:43 CIA-32 bioclipse: shk3 * r8087 /bioclipse2/trunk/plugins/net.bioclipse.cdk.u​i.sdfeditor/src/net/bioclipse/cdk/ui/wizards/ (3 files): if started from navigator, directory is preselected.
20:00 ilogger2 joined #bioclipse

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