Camelia, the Perl 6 bug

IRC log for #cdk, 2009-04-16

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

All times shown according to UTC.

Time Nick Message
04:59 egonw joined #cdk
05:22 jbrefort joined #cdk
06:27 Gpox joined #cdk
07:19 sneumann joined #cdk
07:21 egonw joined #cdk
08:09 shk3 joined #cdk
09:16 edrin joined #cdk
09:51 CIA-58 cdk: shk3 * r14554 /jchempaint/trunk/ (build.xml sign.sh): prepared stuff for applet signing
11:27 edrin1 joined #cdk
15:18 rajarshi joined #cdk
15:21 egonw hi rajarshi
15:22 rajarshi egonw, hi
15:36 rajarshi any interesting bugs to work on? in the office and trying to find stuff to do :)
15:37 egonw I would very much appreciate if you could make a patch to replace IAtom[] with List<IAtom> in IAtomContainer and run some bench makrs on it
15:38 egonw from your blog I expect a considerable speed up
15:38 egonw and same of IAtomContainerSet
15:38 rajarshi aah, i thought that it alredy had IAtom[].
15:38 rajarshi ok i'll go ahead and patch and run some benchmarks
15:38 egonw that does not address the question of that class should be replaced in itself
15:38 egonw but should be rather interesting for HTS apps
15:39 egonw it has IAtom[]
15:39 rajarshi AtomContainerSet already uses IAtomContainer[]
15:39 rajarshi that's why serial acccess is very fast
15:39 egonw but List<IAtom> would be faster not?
15:40 rajarshi i think for serial access direct indexing is faster
15:40 rajarshi i updated the graph on the post
15:40 egonw mom
15:41 egonw ah, ic
15:41 egonw what was wrong with the prev one?
15:41 rajarshi i didn't average over multiple runs
15:41 egonw ah, ok
15:44 rajarshi to be hinest I'm not sure why there is such a big difference since ArrayList uses Object[] internally
15:44 egonw well, the stats for remove etc were quite convincing
15:45 egonw JCP is using deleting
15:45 rajarshi indeed
15:45 rajarshi IMO, subclassing List is probably the best way to go for overall performance
15:48 egonw well, if you prefer to do something else...
15:48 egonw stereocenter detection?
15:49 egonw there are a few bug reports for 1.2.x
15:49 rajarshi i thought kieron was working on that - didn't want to duplicate efforts
15:49 egonw easy ones: 27192372768643
15:49 egonw ah, ic...
15:49 egonw I missed that then
15:50 egonw 2714283
15:50 rajarshi can you list the bug numbers afgain with spaces :)
15:50 egonw easy ones: 2719237 2768643
15:50 egonw :)
15:50 egonw well, you know where to find them :)
15:50 shk3 egonw: i changed tests in my git repository. How do I update the published repository?
15:50 rajarshi ok, i'll look at these
15:51 egonw much appreciated!
15:51 egonw I'll try to do some too, and release 1.2.1 tomorrow
15:51 rajarshi cool
16:08 * egonw is going home
16:17 jbrefort joined #cdk
16:22 maclean joined #cdk
16:37 shk3 joined #cdk
17:09 maclean hey rajarshi, how's it going?
17:09 rajarshi good
17:09 rajarshi how're you?
17:09 maclean recovering from a cold, but otherwise okay.
17:10 rajarshi are you doig anything at the cdkws?
17:10 maclean attending :)
17:10 maclean sorry to hear that you won't be there, US immigration sure does suck
17:11 rajarshi yeah - have to make a 3 week trip to India. Pieces of paper controlling our lives :(
17:11 maclean :(
17:11 rajarshi btw, nice discussion of PDB / BioPolymer redesign
17:12 maclean thanks.
17:12 maclean It's a tricky thing, though. Just because both CDK and Biojava have an 'Atom' class, doesn't mean they are the same thing.
17:12 rajarshi right
17:12 maclean They are different models of an atom, for different purposes.
17:12 maclean I guess.
17:13 rajarshi i looked at it some time ago, and proper bridign coed is non-trivial
17:13 maclean Very much so.
17:13 maclean But its frustrating not to be able to do chemistry analysis on biomolecules
17:13 rajarshi i agree
17:14 rajarshi have you had an interactions with th ebiojava people?
17:14 rajarshi do they have any ineterst in chemistry support?
17:14 maclean No, but Andreas Prlic (?) used to be at the EBI until quite recently. I think I just (a year ago) missed him.
17:14 rajarshi aah, right
17:15 maclean And they probably would have interest in the chemistry side, but it is a kind of cross-discipline barrier at work.
17:15 rajarshi true
17:16 maclean the CDK doesn't model the problems that protein crystallographers are concerned with, and vice versa.
17:16 rajarshi do you do much work on the protein/crystallography side?
17:17 egonw joined #cdk
17:17 maclean not on the experimental end, no but my PhD was (er is..) in protein structural stuff
17:17 maclean I know much more about protein structure than chemistry :)
17:17 rajarshi aah. You're a PhD student at EBI?
17:18 maclean no. I was a PhD student in Glasgow, supervised by Prof. Milner-White.
17:19 rajarshi aah
17:19 maclean On protein backbone carbonyl-cation interactions.
17:20 rajarshi aah OK. Then the CDK is pretty limiting from your side of the fence :)
17:21 maclean well if I had tried to use it in my phd, then yes. I mostly used biopython.
17:22 maclean actually, I mostly wrote my own code. I have a terrible habit of wheel-reinventing.
17:22 rajarshi :)
17:22 rajarshi I think everybody who writes code in chem or bio, ends up writing PDB parser at one point in their life
17:23 maclean oh absolutely
17:30 egonw bbl
17:30 maclean k
17:48 egonw back
17:54 rajarshi egon, 4 bugs cleaned up, patches sent to list
17:55 egonw good, will review them tonite
18:58 egonw cool
18:58 egonw you can fork Gists
18:58 egonw http://gist.github.com/96594
19:00 egonw I wonder if one can pull from forks too...
19:10 maclean egonw : I think that the list on the right hand side shows the urls of the revisions
19:11 maclean looks like it follow the pattern http://gist.github.com/%gist_number%/%git_hash%
19:11 maclean er... that's not a real url...
19:21 egonw ah, but that's not what I meant :)
19:21 egonw I was wondering if Ola could pull patches from my fork back into his Gist
19:22 maclean that sounds like git magic that I don't understand :)
19:23 egonw hehehe
19:23 egonw you'll learn soon enough
19:23 egonw jchempaint-primary will be converted into a set of git branches
19:23 egonw after BC2.0.0
19:29 maclean a SET of branches?
19:29 egonw yes
19:29 maclean A forest?
19:29 egonw no, no ...
19:29 egonw no bootstrapping involved either
19:30 egonw basically just a series of patches
19:30 maclean An undirected acyclic bipartite graph?
19:30 egonw that can be applied in sequence
19:30 egonw but easily reviewed independently
19:30 egonw ummm
19:30 * maclean was browsing graph theory books today :)
19:30 egonw directed acyclic graph
19:31 * egonw heads to course material hacking
19:31 maclean good good.
19:32 egonw plenty of stuff to do there
19:34 egonw maclean: if you want a git magic crash course
19:34 maclean hmmm?
19:35 egonw yes?
19:35 maclean what? I thought there was a second part to that sentence.
19:36 rajarshi :)
19:36 maclean if I want a git magic crash course...
19:36 egonw ... I can give one
19:36 maclean ohh.
19:36 maclean at the workshop?
19:36 egonw now?
19:37 maclean ohh.
19:37 maclean well, I'm deep in fixing the bioclipse JCP centering/drawing bug...so, no
19:37 maclean thankyou though.
19:37 egonw ah, ok
19:39 egonw rajarshi: can you please set your IDE to not use import foo.*; ?
19:40 rajarshi oh it did?
19:41 rajarshi it should only do it for awt and swing classes. did it do it for CDK classes?
19:42 egonw yes
19:42 egonw git diff --cached
19:42 egonw also good to find redundant whitespace
19:42 rajarshi ok, fixed that
19:42 egonw which tends to show up as red
19:42 egonw new patch?
19:43 egonw or fixed in IDE?
19:43 rajarshi no, my idea settings
19:43 maclean weird. I thought that there was an explicit package-private modifier in java. wishful thinking, I guess.
19:43 egonw rajarshi:
19:43 egonw +     * Places the atoms in a linear chain.
19:43 egonw +     *
19:43 egonw +     * Expects the first atom to be placed and
19:43 egonw that new line has no meaning
19:44 egonw if you do not add <p> too
19:47 egonw rednael?
19:48 rajarshi i thoguht it didn't matter for the first line since it gets taken as the summary for the javdoc
19:48 egonw that is true
19:49 rajarshi can you give an example of a * import with cdk packages? It shouldn't have done that
19:50 egonw one of the 4 patches you sent tonite
19:50 egonw 0001-Partial-...
19:51 rajarshi aah, ok, it always does it for classes in the interfaces package
19:51 egonw rajarshi: can you explain the testBadRingClosure1() method?
19:51 egonw why is that smiles wrong:
19:51 egonw p.parseSmiles("c1ccccc1Cc1ccccc1");
19:51 egonw reuse of ring numbers is allowed
19:51 rajarshi ??
19:51 rajarshi the last one was meant to be missing
19:52 egonw ok, looking forward to a revised patch :)
19:52 rajarshi aargh, my mistake. Can you just drop that last 1. much simpler :)
19:52 egonw ok
19:52 egonw I'll fix the patch, ok?
19:53 rajarshi yes, thanks
19:53 egonw (instead of making a fix patch)
19:54 egonw git++
19:54 egonw me loves editing patches :)
19:54 rajarshi :)
19:56 egonw OK, all 4 patches applied and push to SF
19:57 egonw rajarshi: btw... check the SF git help pages
19:57 egonw I think there is a reason why you should use your SF email when doing commits
19:58 egonw you can overwrite your machine global git.email just for SF gits
19:58 egonw by editing cdk/.git/config or so
20:00 rajarshi ok
20:02 rajarshi hmm, don't see anything saying I should use a SF email
20:03 egonw mmm, I read it somewhere...
20:03 rajarshi well, they say you "should" but not exactly why
20:04 egonw http://twittersheep.com/res​ults.php?u=egonwillighagen
20:04 zarah egonw's link is also http://tinyurl.com/clwxso
20:04 egonw rajarshi: is there something you want to learn?
20:04 egonw XMPP cloud services maybe?
20:05 rajarshi hmm, seems interesting
20:05 egonw yes, it is
20:05 rajarshi i've seen your blog posts on it
20:05 egonw and the lib developer is around
20:05 egonw it's fairly easy
20:05 timvdm joined #cdk
20:05 egonw the code for the CDK services is in SVN
20:05 egonw hi timvdm
20:05 rajarshi still not clear why it's better than SOAP. Can it be present REST interfaces?
20:06 egonw it's very good at async things
20:06 egonw and has strong data typing
20:06 timvdm hi
20:06 zarah hi timvdm
20:06 rajarshi oh right, I remember you telling me about the async featurs
20:06 egonw you can create data bindings on the fly
20:06 egonw even for the CML schema
20:06 rajarshi aah
20:06 egonw so no need for WSDL
20:07 rajarshi ok, thats a big plus :)
20:07 egonw yes, we think so too :)
20:07 egonw do you have PSI on your machine?
20:08 rajarshi noi don't think so
20:08 rajarshi is tat the jabber server?
20:08 egonw no, client software
20:08 rajarshi any jabber client should do? or just PSI?
20:09 egonw no, any client will do
20:09 egonw but PSI has nice service discovery
20:10 egonw creating the input data, however, does not work so well with a regular jabber client
20:10 rajarshi aah, ok
20:10 egonw you need one that allows you to craft a bit of XML
20:10 egonw Bioclipse works though
20:10 rajarshi aah, the emacs of life sciences :)
20:10 egonw :)
20:11 egonw I hope the R of life sciences
20:12 egonw rajarshi: do you happen to know how I can draw a color box behind some text?
20:12 rajarshi in Java2D?
20:12 egonw in latex
20:12 egonw sorry
20:12 rajarshi colorbox
20:13 rajarshi \colorbox{red}{foo}
20:13 egonw ok, sounds good
20:16 egonw OK, good enough for me
20:17 rajarshi these materials are oging to be on the CDK wiki? Or the EBI page?
20:17 egonw EBI page, I guess
20:18 rajarshi oh ok
20:18 egonw I'll make sure link things in the wiki
20:20 maclean gah! no wonder I was getting confused. some of the .cml/.mol files I was testing with (in Sample Data) are off-center. So they draw (correctly) off center...
20:20 egonw maclean: what happened to center-of-mass?
20:21 maclean what about center of mass?
20:21 egonw never mind
20:21 egonw in retrospect...
20:21 egonw don't want to interfere with your thinking...
20:21 maclean hmm. no, no its okay.
20:22 egonw œħ nœ
20:22 maclean I'm wondering if JCP shouldn't re-center when drawing.
20:22 egonw łħªþ ħªß µ← ĸ€←”œª®ð ðœn€¿¿¿
20:22 maclean HE COMES!
20:22 egonw ªnð ħœł þħ€← ħ€€łł ðœ ı ŋ€þ →þ ”ª©ĸ þœ nœ®µ€ł¿¿¿
20:22 maclean :) (internet joke...)
20:23 egonw œħ nœ
20:23 maclean what the hell? :D
20:23 egonw ©ªn ←œ↓ ®€ªð µ€ nœ®µªłł←
20:23 rajarshi ???
20:23 maclean ZALGO!
20:23 egonw helP
20:23 rajarshi is on its way
20:24 egonw phew
20:24 egonw my keyboard went nuts
20:24 maclean awww. I was hoping that you were fighting off creatures from between the walls of reality.
20:24 egonw it seems the right Alt was stuck
20:25 maclean dread horrors from beyond the stars.
20:25 rajarshi ha ha
20:25 egonw the helP was actually copy pasted :)
20:25 egonw nice laptop
20:25 egonw crappy keyboard
20:26 egonw mmmm...
20:26 egonw not entirely what I wanted
20:26 egonw now all text is orange :)
20:27 maclean very dutch....
20:27 egonw yes, I like orange
20:28 egonw rajarshi: so, you like a crash course xmpp services?
20:28 egonw writing one, that is?
20:28 rajarshi sure
20:28 rajarshi i have time to kill
20:29 rajarshi where is the example code?
20:29 egonw http://cdk.svn.sourceforge.net/viewvc​/cdk/cdk-xws/trunk/xws-qsar-services/
20:29 zarah egonw's link is also http://tinyurl.com/c25dyk
20:30 egonw mmm... I used Eclipse projects...
20:30 egonw but the jars are all there...
20:30 egonw so you can just use javac to compile
20:31 egonw implement the IFunction interface
20:32 egonw and link to it via the net/bioclipse/xws/componen​t/functions/Functions.java
20:32 egonw that class name is still hard coded
20:32 egonw edrin: ping
20:32 edrin yes
20:32 egonw edrin: will rajarshi need to set up a local ejabberd server?
20:32 egonw to try the services?
20:33 edrin just to use them or run the services himself?
20:33 egonw can components also hook up with other XMPP servers?
20:33 egonw run services
20:33 egonw I mean, I can host them at BMC but that's not really nice for testing
20:34 egonw and another server with XMPP services would be nice anyway :)
20:34 rajarshi i won't have a server for long though
20:34 edrin in this case he must run a server himself obviously
20:34 egonw e.g. on his laptop?
20:34 edrin this is possible
20:34 edrin indeed
20:34 egonw does the default ubuntu ejabberd exist?
20:35 rajarshi any servers for OS X?
20:35 edrin i once run an ejabberd on my windows laptop but it was more userfriendly to just use a test server online
20:35 maclean_ joined #cdk
20:36 edrin rajarshi: interestingly OS X has one built in
20:36 egonw any free servers online he can connect too, perhaps?
20:36 egonw welcome back maclean_
20:36 maclean_ hello.
20:36 egonw OS/X has one built in?
20:36 maclean_ ^this
20:36 egonw to which he can hook up io-data components?
20:36 rajarshi ok, i'm getting confused. if I wwrite a service don't I have to host it:? how does an online server help me? I need to get it there right?
20:36 edrin egonw: in this case i would recommend the ws1.bmc.uu.se or the ayena.de
20:37 egonw ws1 ports are not open to outside :(
20:37 edrin rajarshi: a service is like a client, it just connects to the server...
20:37 egonw rajarshi: your component can run anywhere
20:38 edrin rajarshi: I will give you the account data for a test server a friend of mine hosts...
20:38 rajarshi ok, thanks
20:38 edrin rajarshi: what is your email address ?
20:38 rajarshi rajarshi.guha@gmail.com
20:39 edrin (sorry, I should have it ;)
20:41 egonw I wonder why rednael is not posting commit messages...
20:44 rajarshi hmm, looking at the code
20:44 rajarshi what is the end goal of this? generate a set of XMPP services, right?
20:44 egonw one service, set of functions, yes
20:45 rajarshi ok
20:46 rajarshi isn't it a lot of grunt work - bnasically wrapping CDK classes as XMPP functions?
20:46 egonw yes
20:46 egonw well, this one is
20:46 egonw you can do whatever service you like
20:47 rajarshi right
20:47 edrin ok, mail sent
20:47 edrin cc egon
20:47 rajarshi thanks
20:48 edrin rajarshi: don't wonder about this host/port/ip forwaring thing... the ayena server is just a private server for him (and for me for testing)
20:48 rajarshi ok, if I get stuck i'll bug you :)
20:48 edrin fine :)
20:49 egonw rajarshi: theoretically, with JavaScript XMPP libraries around, one can call these services using GreaseMonkey userscripts too
20:50 egonw but have not worked out how exactly
20:50 rajarshi aah, that'd be useful
20:50 egonw yes, think so too :)
20:53 edrin http://www.apple.com/server​/macosx/features/ichat.html
20:53 zarah edrin's link is also http://tinyurl.com/3qno8c
20:54 edrin Seems to be jabberd1.4 http://en.wikipedia.org/wiki/IChat_Server
20:54 zarah edrin's link is also http://tinyurl.com/c9xfxh
20:58 egonw cool
20:58 egonw the PDF now lists the file with source code that backs up the code snippets
21:06 rajarshi off for the day, goodnight
21:06 egonw bye
21:06 egonw thanx for the fixes!
21:23 egonw joined #cdk
23:20 timvdm joined #cdk
23:21 IAmLegendToo joined #cdk
23:22 IAmLegendToo Any willing to help me confirm I am doing a problem right? Topic is nuclear transmutation.
23:24 IAmLegendToo left #cdk

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