Camelia, the Perl 6 bug

IRC log for #cdk, 2010-09-26

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

All times shown according to UTC.

Time Nick Message
03:04 azeem_ joined #cdk
03:07 azeem left #cdk
06:46 jbrefort joined #cdk
08:25 egonw joined #cdk
09:06 egonw one day left...
09:06 egonw big virtual party tomorrow!
09:27 egonw alchimiste: please update the FR wikipedia page for the cdk wrt the version
09:31 bag_ joined #cdk
10:39 egonw left #cdk
11:21 egonw joined #cdk
12:28 maclean joined #cdk
12:28 maclean noon
12:28 egonw hej maclean
12:28 maclean hej
12:28 egonw party tomorrow?
12:29 maclean heh. I saw the mail. I'll be online, yes.
12:29 maclean but today, I have a single goal : unit tests and javadocs for the rest of the jcp patch.
12:30 egonw oh, wow...
12:30 egonw the full patch?
12:30 egonw I'd be happy with renderbasic to start with :)
12:30 egonw I saw your comments...
12:30 egonw on render-api
12:32 maclean yeah, pretty minor ones.
12:33 maclean the LeafElement thing is just sugar.
12:35 maclean egonw : Oh, and so with renderbasic, all the unit tests should be in test-renderbasic, and so on for the rest of 3 to 14?
12:35 egonw correct
12:35 maclean Ok.
12:46 egonw got an interesting patch lined up...
12:46 egonw you know the CDKSourceCodeWriter?
12:46 egonw I have just updated the unit test to actually test of the output compiles :)
12:46 egonw I have just updated the unit test to actually test if the output compiles :)
12:51 maclean Hah. Cool.
12:53 maclean I think I have a way to test the AWTDrawVisitor, btw - draw individual elements on a small image, then test for the (presence?) of pixels.
12:53 egonw indeed
12:59 bag_ left #cdk
13:09 rajarshi joined #cdk
13:10 rguha joined #cdk
13:10 egonw hi rajarshi, rguha
13:10 maclean :)
13:10 rguha hi everybody
13:10 rguha 10 years!
13:10 rajarshi left #cdk
13:10 egonw yeah, time goes fast
13:10 egonw i'm getting old :(
13:14 maclean bah. we have an annoying show here in the uk called "Are you smarter than a 10 year old?" - the obvious answer is "yes". 30yr olds are probably smarter than 20yr olds, too :)
13:15 egonw and then it is downhill
13:15 maclean Heh. I was assuming a generalisation to 40, 50, etc...
13:17 maclean Just think what you can do (code) now that the egonw of 10 yrs ago couldn't.
13:17 egonw indeed
13:24 egonw rguha: regarding that SMARTS
13:24 egonw I thought it was supposed to match the tetrazole ring...
13:24 egonw and just that...
13:24 egonw but not really into the $() syntax...
13:25 egonw OK, so the JOELib actually had:
13:26 egonw [$(),$()]
13:26 egonw with all SMARTS comma separated
13:26 rguha @egonw actually in the recursive SMARTS, the part in the $() is the atom environment so my understanding is that it matches everything *except* the atom environment
13:26 egonw does the CDK do recursive SMARTS?
13:27 rguha yes, but I think it's buggy
13:27 egonw :(
13:27 rguha but apparently my understanding is wrong since n(CCC)1nncn1 gives the 2 N's as the matching atoms
13:28 sneumann joined #cdk
13:28 rguha hi sneumann
13:29 sneumann hi there, thanks for the rcdk update
13:29 rguha you're welcome
13:29 sneumann I *may* have a working prototype or the molecule clustering + cluster representative molecule
13:29 sneumann I was looking for, I just have to verify my script
13:29 egonw rguha: yeah, [$(n1nnnc1)] seems to match the whole molecule
13:29 rguha cool
13:30 rguha egonw: but then why does it just match the 2 N's and not the whole ring?
13:30 egonw yeah, I don't get the color coding iether
13:30 egonw if I ensure it does not match, it's all white...
13:32 egonw rguha: I think it only shows the first atom...
13:32 rguha hmm, yes, that's possible and explains the two unique matches
13:32 egonw indeed
13:33 maclean sneumann : you were after the SMSD in the CDK the other day, were you? I'm guessing you got at it through rCDK?
13:34 sneumann exactly.
13:35 sneumann I still have to figure whether it is using the default MCS or the SMSD,
13:35 sneumann the manpage says it is the former
13:35 sneumann in rcdk
13:35 maclean the SMSD has an Algorithm enum, with one value being default MCS, I think.
13:35 rguha IIRC the Default algorithm automatically switches between the various MCS algorithms in the SMSD. That's why I didn't specify the MCSPlus explcitily
13:36 sneumann I'll play around more in the evening, bye
13:36 maclean bye
13:36 rguha BYE
13:36 rguha egonw: interesting - rearrange the SMARTS string and two otehr N's get matched: [$(n1nnnc1)]
13:37 rguha so yes, it is matching the first atom in theenvironment - which I don't understand why
13:41 sneumann left #cdk
13:43 sneumann joined #cdk
13:55 rguha ttyl
13:55 rguha left #cdk
14:26 maclean egonw: patch 3075839 ('small typo fix') has no attached files.
14:26 maclean https://sourceforge.net/tracker/?func=detail&amp​;atid=320024&aid=3075839&group_id=20024
14:26 egonw argh... stupid SF
14:27 egonw it does now
14:29 maclean I wish sf's patch tracker was more like github.
14:31 egonw yes
14:31 egonw same here
14:31 egonw maybe I should start using the pull request thingy, now the github allows automatically filing as issue
14:31 egonw but that would split up where to file patches...
14:32 maclean well, I like the list view that sf provides, just not the 'attached files' aspect.
14:33 maclean oh well, as they say : "if wishes were horses, beggars would ride..."
14:53 egonw true
14:53 maclean Ok so that's AbstractRenderer fully javadocced and whitespace-fixed. That only took...an hour.
14:53 maclean http://github.com/gilleain/cd​k/commits/tests-2-renderbasic
14:53 egonw but then there is open source, and wishes become itches, become patches...
14:54 egonw cool
14:58 maclean only another 27 or so to go :)
15:01 egonw indeed a lot of work...
15:06 maclean right, first should go get some food.
15:50 maclean uh-oh. If I do this:
15:50 maclean ant -f pmd.xml -Dmodule=renderbasic
15:50 maclean it seems to run many times.
15:50 maclean possibly forever.
15:51 egonw confirmed...
15:52 egonw you need to do instead
15:52 egonw ant -f pmd.xml -Dmodule=renderbasic test-module
15:52 maclean ohhhh...
15:53 egonw what your call will do is run the pmd testing on renderbasic
15:53 egonw as many times
15:53 egonw as there are modules being tested
15:53 maclean ah! :)
15:53 egonw which is 20 or so
15:53 maclean BUILD FAILED
15:53 maclean Can't find resource pmd/${pmd.test}.xml.
15:54 egonw ant -f pmd.xml -Dmodule=renderbasic -Dpmd.test=custom-test test-module
15:55 maclean well that runs... but the output file?
15:56 maclean reports/pmd/renderbasic.xml?
15:59 maclean never mind, I found this helpful blogh post from some Dutch guy:
15:59 maclean http://chem-bla-ics.blogspot.com/2010/0​3/cleaner-cdk-code-3-run-pmd-tests.html
16:00 egonw :)
16:25 maclean egonw : should I file a feature request to have xslt post-processing of the pmd?
16:25 maclean I tried it, and it was quite easy.
16:28 egonw you can also change the formatter
16:28 egonw in pmd.xml
16:28 egonw which is now set to xml
16:28 maclean I added <xslt in="${reports.tests}/pmd/${module}.xml" style="pmd/wz-pmd-report.xslt" out="${module}.html"/>
16:33 egonw sounds good to me
16:33 maclean also, this page was interesting:
16:33 maclean http://tomcopeland.blogs.com/juniorde​veloper/2005/12/demo_of_some_ni.html
16:33 maclean (the "demo of it here" link)
16:36 egonw yeah, I am using that approach on rdf.openmolecules.net too
16:40 egonw exceprt for the JS bits
16:41 maclean well, I think the js bits are the point :)
16:42 maclean Hmmm. So I think my feature request would actually be about a "module-level" nightly.
16:43 maclean So that you could do something like "ant -f nightly_or_something.xml -Dmodule=MyModule report-module"
16:43 maclean to get ojdcheck, pmd, junit, etc.
16:43 egonw check the build.xml
16:43 egonw qa-module or so
16:44 maclean ahh, right qa, yes.
16:48 maclean but it doesn't produce, eg:
16:48 maclean http://pele.farmbio.uu.se/nightly​-jcp/test/result-renderbasic.html
16:51 maclean arrgh. no, that's reports/result-renderbasic.txt
16:51 maclean except that module-level reports never run right the first time.
17:17 egonw wow... http://www.openscience.org/~egonw/
17:19 maclean JCFL : "There is no tutorial yet....you can request a tutorial. If need is urgent enough or enough request, i'll make time to write it." :)
17:20 egonw hahaha
17:20 egonw well, it's all described in the paper...
17:20 maclean heh.
18:08 maclean hmmm. ojdcheck doesn't deal with enums very well. It complains about 'missing docs for valueOf' which is indeed a method of class Enumeration....
18:09 egonw yes: http://github.com/egonw/ojdcheck/issues#issue/28
18:10 maclean and I'm guessing that public inner classes don't need @cdk.module tags?
18:13 egonw correct
18:13 maclean great :)
18:14 egonw https://sourceforge.net/tracker/?func=detail&amp​;aid=3043084&amp;group_id=20024&amp;atid=120024
18:14 egonw but then again... that bug should no longer be there...
18:32 jbrefort left #cdk
18:35 maclean Hmmm. RectangleElements might not be centered on thier x,y point....
19:06 maclean left #cdk
20:00 jbrefort joined #cdk
20:44 sneumann left #cdk
20:51 jbrefort left #cdk
22:01 egonw left #cdk

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