Camelia, the Perl 6 bug

IRC log for #cdk, 2009-03-16

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

All times shown according to UTC.

Time Nick Message
05:48 jbrefort joined #cdk
07:24 egonw joined #cdk
07:33 egonw joined #cdk
07:44 edrin joined #cdk
07:49 egonw joined #cdk
08:00 Gpox joined #cdk
10:02 s-wolf joined #cdk
10:03 s-wolf hi there
10:04 s-wolf i have a question regarding the 2D picture generation in cdk 1.2
10:06 s-wolf i saw that the class 2DRenderer was renamed to Java2DRenderer
10:08 s-wolf so i changed the references and some small things but when I try to write an image file (png) to disk the picture is just blank
12:05 carsten joined #cdk
12:27 shk3 joined #cdk
13:07 rajarshi joined #cdk
13:12 egonw hi rajarshi
13:18 rajarshi hi egonw
13:21 rajarshi spring break here
13:25 egonw nice
13:25 egonw still winter here :)
13:26 CIA-62 cdk: gilleain * r14495 /cdk/branches/jchempaint-primary/src/main​/org/openscience/cdk/renderer/selection/ (5 files): Reverting the removal of the interface method 'public void select(IChemModel)' as it is needed for selecting outside the hub.
13:27 CIA-62 cdk: gilleain * r14496 /jchempaint/trunk/src/main/org/openscienc​e/jchempaint/action/CopyPasteAction.java: Minor change to allow for changes to the selection package.
14:09 egonw rajarshi: Bioclipse just gained a R console
14:09 egonw I have not tried it yet
14:09 egonw but it's more or less what is used in the QSAR paper
14:10 egonw as the R is called there via the regular JS console, AFAIK
15:18 rajarshi cool
15:18 CIA-62 cdk: gilleain * r14497 /cdk/branches/jchempaint-primary/src/​main/org/openscience/cdk/controller/ (5 files):
15:18 CIA-62 cdk: - Altered makeNewStereoBond to return a reference to the new bond.
15:18 CIA-62 cdk: - Fixed AlterBondStereoModule so that it does what it did before.
15:39 CIA-62 cdk: gilleain * r14498 /jchempaint/trunk/src/main/org/openscienc​e/jchempaint/action/CopyPasteAction.java: Fixed bug to select all introduced by previous commit.
15:53 rajarshi egon, do you know when cdk nightly will be back up?
15:57 egonw on pele
15:57 egonw yes, more todo
15:57 egonw ummm... the script is supposed to work?
15:57 egonw I just update from SVN?
15:57 egonw same call from the command line
15:57 egonw and just update of config file?
15:57 rajarshi identical
15:57 egonw OK
15:57 rajarshi just update the config file with a new variable
16:00 egonw ok, will try tonite
16:00 rajarshi thnaks
16:02 egonw I will also try to put the vflib branch online
16:03 egonw at pele
16:03 egonw and this overview page...
16:03 egonw bla...
16:31 CIA-62 cdk: gilleain * r14499 / (7 files in 3 dirs):
16:31 CIA-62 cdk: - Added a parameter called 'selectionRadius' for the size of the selection element.
16:31 CIA-62 cdk: - Divided selection generation to make use of BasicBondGenerator methods.
16:35 CIA-62 cdk: gilleain * r14500 /cdk/branches/jchempaint-primary/src/main/org/open​science/cdk/controller/AlterBondStereoModule.java: Uncommented the selecting of altered stereo bonds.
16:39 jbrefort joined #cdk
16:48 carsten joined #cdk
16:48 CIA-62 cdk: gilleain * r14501 /cdk/branches/jchempaint-primary/src/main/org/​openscience/cdk/controller/AddBondModule.java: Make behaviour of adding bonds consistent with adding wedge bonds, namely that the sprouted atom is highlighted after adding the bond, rather than the root atom.
18:23 rajarshi put 4GB in my mac. It's blazing :)
18:51 egonw :)
18:52 egonw does it see all 4GB?
18:52 egonw my linux only sees 3.5 of the 4...
18:52 egonw there is a kernel patch, but not applied by ubuntu by default
18:52 egonw and have not backed my own tuned kernel yet...
18:54 rajarshi all 4GB
18:54 egonw OS/X++
18:54 rajarshi IDEA is now smooth (unfortunate that I need 4GB toget it so smooth )
18:55 egonw :)
18:55 egonw Eclipse does better
18:55 egonw I can run three of them smoothly
18:55 rajarshi nice
18:55 egonw actually, have not tried 4
20:03 bag joined #cdk
20:24 * egonw is looking into updating his Nightly's on Pele
20:25 rajarshi nice - I didn't realize how useful it is till it's not there :)
20:25 egonw yeah... btw, next on my shopping list... RDF interface... (or just XML)
20:26 egonw will be handy for the table overview for different Nightlies
20:26 egonw maybe some Ajax on that page...
20:26 rajarshi yes, that'd be nice
20:26 egonw or maybe static JSON?
20:26 egonw nightly.py should be easy enough to export some key statistics as RDF/XML/JSON
20:27 egonw as well as the regular HTML
20:27 rajarshi well, it can always be rewritten as pure XML
20:27 rajarshi actually that would clean up the code by using ElementTree
20:27 rajarshi then you can do XSLT to transform to whatever you want
20:27 egonw sounds good
20:27 rajarshi don't know XSLT myself though :)
20:28 * egonw is good with XSLT
20:28 rajarshi one worry of my plate :)
20:29 egonw yes, the front page as XML and use a bit of XSLT to wrap it with some HTML
20:29 egonw OK, let's go for RDF?
20:29 egonw there is a RDF onto for projects...
20:29 egonw mom
20:29 rajarshi couldn't you convert the XML to RDF directly?
20:30 rajarshi it'd be nice for nightly,py to generate the lowest common denominator
20:30 rajarshi and then do whatever we want with it
20:30 egonw RDF in the XML syntax?
20:30 egonw anyone, is fine...
20:31 rajarshi ok
20:31 egonw XML, but then we can create static HTML too
20:31 rajarshi yes, that's what I was thinking
20:31 egonw which is likely fine too :)
20:31 rajarshi the same way we do XSLT on PMD output to get HTML
20:31 rajarshi we'd do the same thing to the nightly output
20:31 egonw right
20:32 egonw and then we can have a daily which aggregates the XML sources, and summarize it as HTML
20:32 rajarshi nice
20:32 rajarshi is there a good way to get nightly into git?
20:32 rajarshi or are we sticking with svn for that?
20:32 egonw the problem here is that we only have one repository on SF
20:33 egonw one in the git world...
20:33 egonw one project, is one repository...
20:33 rajarshi oh right
20:33 egonw so Nightly needs its own repository...
20:33 rajarshi what about moving nightly back into the cdk repo?
20:33 rajarshi it's pretty closely tied to th ebuild system
20:34 egonw but it is a rather separate project nevertheless...
20:34 egonw with a different release schedule...
20:34 egonw well, it's not even released
20:34 egonw ummm... how does the config say whether to use git or svn?
20:34 rajarshi i didn't think it was mean to be released
20:34 egonw or does the new nightly no longer do svn at all?
20:34 rajarshi oh, I switched it completely to git
20:35 rajarshi doesn't do svn anymore
20:35 egonw ah, ok, fine
20:35 rajarshi but it's easy to specify which branch
20:35 rajarshi actually, nightly should utput somewhere on the web page which branch is worked on
20:36 egonw so, just one path, for all branches?
20:36 egonw as in, nightly_repo ?
20:37 rajarshi yes
20:37 rajarshi aren't all your branches in one dir?
20:37 egonw they are now...
20:38 egonw but since I may want to put up a URL for branches of others
20:38 egonw in addition to my branches...
20:38 egonw well, dunno yet...
20:38 rajarshi you could just make a copy of the config file and change nightly_repo
20:38 egonw right now I have a git on pele which mirrors, but will put up the vflib there too...
20:38 rajarshi ok
20:38 egonw and likely other branches which are still in the review process
20:38 rajarshi right
20:39 rajarshi multiple config files will solve that cleanly
20:39 egonw but I would use a different URL, I hitnk, for my personal branches for people to test
20:39 egonw rajarshi: yes, indeed
20:39 egonw that's how I have the 6/7 nightlies running now :)
20:39 rajarshi :) you must have a nice machine
20:39 egonw 4 core machine
20:40 rajarshi yummy
20:40 egonw weirdly enough with only 2GB memory
20:40 egonw I need to get that upgraded
20:43 egonw rajarshi: you use anynomous clones for nightly?
20:43 egonw or your developers checkout?
20:43 rajarshi anon clone
20:44 rajarshi shouldn't matter - nightly doesn't push
20:47 egonw rajarshi: btw, you are aware of the #bioclipse channel, right?
20:47 egonw with some time left until you moving to NIH, maybe you want to checkout bioclipse...
20:47 egonw did I tell you about the R integration?
20:47 rajarshi yes, saw that
20:48 rajarshi i'm planning to download and play with Bioclipse this week as I write the paper
20:48 egonw excellent
20:48 egonw happy to hear that
20:48 egonw all feed back more than welcome...
20:48 rajarshi sure
20:48 egonw just file bug reports on pelezilla
20:48 rajarshi :)
20:49 egonw Loading data from persistent storage
20:49 egonw Traceback (most recent call last):
20:49 egonw File "/home/egonw/nightly/nightly/trunk/nightly.py", line 1038, in <module>
20:49 egonw oldDate = data[2]
20:49 egonw IndexError: list index out of range
20:49 egonw what's that?
20:50 rajarshi is this the fist time you're running it?
20:50 egonw do I need to issue 'first' again?
20:50 rajarshi remove any nightly.pickle files
20:50 egonw oh... new folder because of git checkout
20:50 egonw right
20:50 rajarshi and then do nightly.py configile firsttime
20:50 egonw sorry
20:50 rajarshi i'll add some more checks tonight to be a bit more robust
20:51 egonw where can I find those .pickle files?
20:51 rajarshi in the same dir you are running nightly.py
20:52 egonw mmm
20:52 egonw no such files
20:53 egonw "firsttime"
20:53 rajarshi yes
20:54 egonw ah, ic
20:54 egonw well... not running nightly inside the nightly_dir
20:55 rajarshi hmm
20:56 CIA-62 cdk: rajarshi * r14502 /nightly/trunk/nightly.py: Robustified the code to catch cases where the older version of the pickle file is being used
20:59 egonw Traceback (most recent call last):
20:59 egonw File "/home/egonw/nightly/nightly​/trunk/cdk1.2.x/nightly.py", line 1048, in <module>
20:59 egonw oldRevision, oldDate = getGitInfo()
20:59 rajarshi aargh, works here
20:59 rajarshi can you psate the whole error
20:59 rajarshi does git log --pretty=medium --date=iso work on your command line?
21:00 egonw nightly_repo = /home/egonw/nightly/git/cdk
21:00 egonw nightly_dir  = /home/egonw/nightly/nightly/trunk/cdk1.2.x
21:00 egonw nightly_web  = /var/www/nightly-1.2.x/
21:00 egonw doe sthat make sense?
21:00 egonw Processing old data
21:00 egonw Traceback (most recent call last):
21:00 egonw File "/home/egonw/nightly/nightly​/trunk/cdk1.2.x/nightly.py", line 1048, in <module>
21:00 egonw oldRevision, oldDate = getGitInfo()
21:00 egonw File "/home/egonw/nightly/nightly​/trunk/cdk1.2.x/nightly.py", line 898, in getGitInfo
21:00 egonw commitDate = datetime.datetime.strptime(commitDate, "%Y-%m-%d %H:%M:%S")
21:00 egonw File "/usr/lib/python2.5/_strptime.py", line 330, in strptime
21:00 egonw (data_string, format))
21:00 egonw ValueError: time data did not match format:  data=  fmt=%Y-%m-%d %H:%M:%S
21:00 rajarshi hmm, does that git command work for you?
21:01 rajarshi paths look fine
21:01 egonw yes
21:01 egonw commit d2084f0fadae611f59270884b58c4db05dacbc90
21:01 egonw Author: Egon Willighagen <egonw@users.sourceforge.net>
21:01 egonw Date:   2009-03-16 10:30:08 +0100
21:02 rajarshi yeah, iso date parsing is a pain
21:02 rajarshi i've just put in a new version of nightly
21:02 CIA-62 cdk: rajarshi * r14503 /nightly/trunk/nightly.py: Updated code to not parse ISO dates for now, ince we dont actually use it yet
21:02 egonw ack
21:03 egonw ok, it proceeds now
21:04 rajarshi good
21:05 rajarshi btw, do you run the build using the parallel directive? does it speed up significantly?
21:06 egonw mom
21:06 egonw I think I blogged about that
21:06 egonw hang on
21:06 egonw http://chem-bla-ics.blogspot.com/2​008/11/parallel-building-cdk.html
21:06 zarah egonw's link is also http://tinyurl.com/5qbqxj
21:07 egonw so, not as much as I hoped
21:07 egonw 25-30% I guess
21:07 rajarshi aah, thanks. not bad
21:08 rajarshi i've also wondered about the issue of dependencies
21:08 rajarshi it seems really crude that we have to do it ourselve by hand
21:08 egonw yes... it's because we want to keep the source files together
21:08 egonw otherwise, we could have done with maven or so (not that I particularly like the build system)
21:09 rajarshi hmm
21:12 rajarshi brb
21:28 egonw rajarshi: http://pele.farmbio.uu.se/nightly-1.2.x/
21:28 zarah egonw's link is also http://tinyurl.com/5nf6h5
21:29 egonw 89 new fails...
21:29 egonw that scared me
21:29 egonw but then I saw it was just all the current onces...
21:29 egonw being the first time it ran Nightly again :)
21:29 rajarshi :)
21:30 rajarshi nie
21:30 rajarshi are you going to running one for master?
21:30 egonw yes, that one I'll run now
21:30 rajarshi ok
21:32 egonw is there a new link for DTP atom type QA?
21:33 rajarshi no, that ws run separately, not from nightly
21:33 egonw but the link at Indiana will be OK?
21:33 rajarshi no, all ngithly links from IU redirect to pele
21:34 egonw ok, so DTP atom type QA is no longer running at this moment?
21:34 rajarshi correct
21:35 egonw ok
21:35 egonw nightly4trunk is now running
21:36 rajarshi cool thanks
21:38 egonw btw...
21:38 egonw I think for much content we need to move to XML
21:38 egonw I like to summarize things on 'module' pages too...
21:39 egonw so that we can properly have people manage module...
21:39 egonw and see on one page what is wrong with there module
21:39 rajarshi yes, sounds good
21:40 rajarshi i'll see about generating voluminous xml output :)
21:40 egonw :)
21:40 rajarshi we can then pick and choose what to display and how to display it
21:40 egonw thanx
21:40 egonw indeed
21:40 rajarshi need a break from writing anyway
21:53 egonw ok, going offline now...
21:53 egonw nigthly is still running on master...
21:53 egonw but almost done it seems...
21:54 egonw the crontab should still be operational...
21:54 egonw so, that's set
21:54 egonw bye for now!
21:54 rajarshi bye

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