Camelia, the Perl 6 bug

IRC log for #cdk, 2011-03-18

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

All times shown according to UTC.

Time Nick Message
04:47 egonw joined #cdk
05:40 egonw left #cdk
06:14 egonw joined #cdk
06:44 egonw_ joined #cdk
06:47 egonw left #cdk
06:49 egonw__ joined #cdk
06:52 egonw_ left #cdk
07:04 egonw joined #cdk
07:08 egonw__ left #cdk
07:18 egonw left #cdk
07:22 egonw joined #cdk
07:27 egonw left #cdk
07:30 egonw joined #cdk
07:57 Gpox joined #cdk
08:33 egonw left #cdk
09:02 egonw joined #cdk
09:18 jbrefort joined #cdk
09:27 maclean joined #cdk
09:28 maclean morning
09:54 egonw hi maclean
09:54 maclean hi egonw
09:55 maclean how are the grant proposals?
09:56 egonw in a too large quantity, with too soon deadlines :)
09:56 egonw doing 4 things at the same time now...
09:57 egonw send out emails, chat here online, listening to a rather nice talk about risk assessment, *and* writing a grant app :)
09:57 egonw who says men can't multitask :(
09:57 maclean Heh : so my suggestion might be to log off? :)
09:58 maclean Multitasking is a bit of an illusion : better to do one thing well, than many things less well...
09:58 egonw log off... nice idea, but doesn't work anyway... need to pick up info onine
09:58 egonw ask things, align things...
09:59 egonw logging off is not an option in writing anything nowadays
09:59 egonw but the idea is right'
10:00 maclean :) well, good luck with it anyway.
10:02 jkerssem joined #cdk
10:04 maclean Hi, Jules?
10:05 jkerssem hello :-)
10:05 jkerssem thought I'd drop by for a bit, (with some GIT questions)
10:05 egonw hi jkerssem
10:05 jkerssem Hi egon!
10:06 jkerssem I'm trying to set up a proper development environment today
10:06 jkerssem Eclipse and everything
10:06 egonw OK
10:06 maclean Ok.
10:07 jkerssem so now I was wondering: how to best get the CDK in there for maximum patch-making-ness
10:07 jkerssem I already found the Eclipse GIT plugin
10:07 jkerssem so next up: which repo/branch to get?
10:07 egonw jkerssem: I will be in Eindhoven early April
10:07 jkerssem egonw: nice!
10:07 egonw jkerssem: excellent!
10:07 maclean Uhm, I think cdk/cdk on github?
10:07 egonw got a GitHub account already?
10:08 egonw that's a easy way to share larger commits...
10:08 egonw small hacks you can do without
10:08 jkerssem no github account yet
10:08 jkerssem do I have to? o:-)
10:08 egonw no, start without
10:08 jkerssem oki
10:09 egonw jkerssem: what OS agani?
10:09 egonw gonna do git on the command line, or in Eclipse?
10:09 jkerssem linux, ubuntu 10.04 x64
10:09 * egonw does it on the command line
10:09 jkerssem and preferably all in eclipse
10:09 egonw ok, install egit
10:09 egonw if not done already
10:09 jkerssem Egit: done
10:09 egonw ok
10:09 jkerssem then file > import > project from GIT
10:09 jkerssem done too
10:09 egonw now, I don't know how to get egit check out a git repos :)
10:09 egonw cool!
10:10 jkerssem then it asks for the URI
10:10 jkerssem which I also have now
10:10 jkerssem (https://github.com/cdk/cdk.git)
10:10 egonw excellent
10:10 egonw now, pick a random typo in the CDK source
10:11 egonw makes a good learning exercise
10:11 egonw many of them, so second of third class should have one
10:11 jkerssem a, before we go there
10:11 jkerssem which branches do what?
10:11 jkerssem I have 1.2.x | 1.4.x | 1.4.x-padel and master
10:12 jkerssem Master had the API changes
10:12 jkerssem if i recall correctly
10:12 jkerssem but where would we fix the PDBReader things we discussed? 1.4?
10:12 jkerssem and who or what is Padel :-S
10:12 jkerssem (I'll do the typo first, I promise)
10:13 egonw I will blog about padel this weekend
10:13 jkerssem I just want to know which branches are what
10:13 egonw it was recently published
10:13 egonw qsar descriptor calculation tool
10:13 egonw and uses and extends the cdk
10:13 jkerssem ooh, sounds fancy :-)
10:13 egonw maclean: what branch?
10:13 egonw master or cdk-1.4.x ?
10:14 maclean Uhm. I was wondering that too. Unfortunately I get fuzzy about their different purposes.
10:14 egonw I think 1.4.x would be a good target
10:14 egonw renderbasic needs a tid bit more tuning
10:14 maclean If that's the one for less stable stuff, then yes.
10:14 jkerssem but the new error reporting would definitely be master.. no?
10:14 egonw jkerssem: no, that would be perfect for 1.4
10:14 jkerssem since its API-changing, or not?
10:15 egonw it's not introducing new API
10:15 egonw it's merely using API that is already in 1.4
10:15 jkerssem the entire Error-handling infrastructure you described is not new API?
10:15 * jkerssem is now confused...
10:16 maclean There is an existing STRICT error-reporting mechanism in cdk.io
10:16 egonw no, STRICT and RELAXED is already part of 1.4
10:16 egonw the MDL reader is using it...
10:16 maclean PDBReader just doesn't use it yet
10:16 egonw the PDBReader is not using it yet
10:16 egonw right :)
10:16 jkerssem ah, okay then
10:16 jkerssem so, 1.4 it is
10:16 jkerssem should I direct my Typo-tutorial there too?
10:16 egonw yes
10:17 jkerssem okay then
10:17 * jkerssem keeps his fingers crossed
10:17 jkerssem :)
10:17 jkerssem (thar she clones!)
10:18 egonw ok, now the first thing to realize is the fact that you can best use branches for separate projects you work one
10:18 egonw jkerssem: you already had a clone in your Eclipse workspace, not?
10:19 jkerssem I decided to do a completely clean slate today
10:19 jkerssem I think I messed up that clone by fiddling in the files directly with VI etc.
10:19 jkerssem and trying all these command-line patch formattings
10:19 jkerssem one of the reasons I sometimes prefer GUI over cmdline
10:20 jkerssem less change of screwups if you don't know (yet) what you're doing
10:20 egonw hahahaha
10:23 jkerssem my usual M.O. is GUI first > learn how it works > fiddle some on the CMDline > fiddle some CMDline some more > become the master!
10:23 jkerssem I usually stop at step 2 or 3 :-P
10:23 maclean since it is git, you missed out a lot of swearing. In my experience, anyway.
10:24 maclean Then you get to the stage where you do "git diff master..otherbranch" and it works, then the swearing dies down a bit.
10:25 jkerssem well, for work, we still use SVN
10:25 * egonw mumbles git-svn
10:26 jkerssem or rather, everybody else uses SVN and I finally decided to buckle up and figure it out..
10:26 maclean Why, what were you using?
10:26 jkerssem we _apparently_ have an internal SVN server somewhere...
10:26 egonw let me know once you made a single char typo fix stage...
10:26 egonw git svn clone svn+ssh://foo/bar
10:28 jkerssem victory part 1: Eclipse now has an org.openscience.cdk project :-D
10:30 maclean If it doesn't compile, you may have to run the ant target "javacc" from build.xml
10:50 egonw jkerssem: making progress, or just a long fika?
10:50 jkerssem thanks maclean
10:50 maclean n/p
10:50 jkerssem egonw, yes I was making progress, but now I'm confused again
10:50 jkerssem I found a single, one char typo in the PDBReader
10:51 jkerssem fixed it, diff in history looks ok
10:51 jkerssem but when I select "create patch", the result is ~500K....
10:51 jkerssem (KB)
10:51 jkerssem and includes things like the AUTHORS file :-S
10:51 egonw :)
10:52 maclean Hmmm. I had that the other day when I hadn't pulled, I think.
10:52 egonw OK, something has gone bad with your clone, I'd say
10:52 egonw jkerssem: try this:
10:52 jkerssem hmm, that may explain the red exclamation mark on the project...
10:52 egonw right click org.openscience.cdk
10:52 egonw choose: Reset ...
10:53 egonw select cdk-1.4.x and hit the Reset button
10:53 jkerssem Do you mean r-click > Team > Reset?
10:53 jkerssem and if so
10:53 egonw um... yes, right
10:53 jkerssem should it be soft, mixed or hard?
10:53 egonw sorry, missed to report the Team bit :)
10:53 egonw oh...
10:54 egonw dunno...
10:54 jkerssem that's okay :)
10:54 egonw that it did not ask here...
10:54 jkerssem default is mixed
10:54 egonw umm... hard
10:54 egonw hard means := ignore whatever I did myself
10:54 jkerssem okay, hard it is :)
10:55 jkerssem Aha, that un-fixed the typo :)
10:55 egonw yes
10:55 egonw that's why a small change is a good learning excerise :)
10:56 jkerssem definately
10:56 jkerssem so now what..
10:56 jkerssem "fetch from upstream" reports "no changes, everything up to date"
10:56 jkerssem just retry?
10:56 egonw that sounds good
10:56 egonw when did it happen?
10:57 egonw upstream is github
10:57 egonw and between now and an hour ago, no new patches reached upstream
10:57 egonw so I'd expect your checkout to still be up to date
10:57 jkerssem I'd expect the same
10:57 jkerssem configure fetch indeed reports upstream to be: "git://github.com/cdk/cdk.git"
10:58 jkerssem and from the looks of the project explorer I'm on branch [CDK cdk-1.4.x]
10:58 egonw good
10:58 egonw ok, wait with that typo fix
10:58 egonw as I said, best to do things in branches first
10:59 jkerssem alright
10:59 egonw then you can never get something conflictnig local, and simplify keeping your local cdk-1.4.x /master /etc in sync with upstream
11:00 jkerssem sounds smarts :)
11:00 egonw o.o.cdk -> Team -> Branch
11:00 jkerssem check
11:00 egonw and click new branch, and make sure cdk-1.4.x is selected to branch *from*
11:00 egonw I can recommend a to use a namnig scheme
11:01 egonw I use: branchNumber-branchedFrom-someCode
11:01 egonw like: 267-14x-typoFix
11:01 egonw the number is subsequential, to reflect about when I created that branch
11:02 egonw 14x to understand what I branched from
11:02 jkerssem that sounds better than my plannen "jules" :-)
11:02 egonw "typoFix" -> to reflect what I am trying to do in that branch
11:02 jkerssem so that would be: 1-14x-typofix
11:02 egonw sounds good :)
11:03 jkerssem yay, my very own, very first git branch :-)
11:03 egonw ok, lunch time here now :(
11:03 * jkerssem pops the champagne
11:03 egonw jkerssem++
11:03 jkerssem enjoy :)
11:03 egonw can we continue in about 2 hours?
11:03 jkerssem I should eat something too
11:03 jkerssem np
11:08 egonw left #cdk
11:47 mgerlich left #cdk
12:08 egonw joined #cdk
12:16 egonw jkerssem: noted the patch
12:16 egonw maclean: want to get some patch sign off fame?
12:16 maclean :)
12:17 maclean Will there be marching bands, and ticker tape parades?
12:17 egonw yes, in my blog
12:17 maclean Heh. Anyway, the type patch, yes?
12:17 egonw I always add a 'who reviewed' top X ranking based on this
12:17 egonw typo patch...
12:17 egonw nice recursion!
12:18 maclean unintentional
12:18 egonw maclean++ (anyway)
12:34 maclean Hmm. Anyway, done.
12:38 jkerssem yay :)
12:42 jkerssem it's even signed off already :-D
12:45 maclean Yup. And with that, I go to lunch.
12:47 jkerssem enjoy :)
13:34 egonw left #cdk
13:34 egonw joined #cdk
14:19 maclean Hmmm. By some coincidence I get R/S right... only a 50% chance after all.
14:38 egonw :)
14:51 CIA-57 org.openscience.cdk: Egon Willighagen bioclipse2.6 * r87d24b6 / (89 files in 23 dirs): Added SMSD - http://bit.ly/hIGeWq
14:52 egonw jkerssem: mmm... egit seems bad at branching etc...
14:53 egonw got tons of changes because apparently a patch has been undone or so...
14:53 egonw sounds pretty bad :/
14:53 * egonw returns to git from the command line
14:53 jkerssem Well, it is still in incubation...
14:53 egonw so that something goes wrong, I know it's myself to blame
14:53 jkerssem I guess some things are best done on command
14:53 jkerssem for now at least
14:54 jkerssem btw: is there a way to remove branches?
14:54 sneumann left #cdk
14:54 jkerssem I created one, but am not sure where it based from
14:54 jkerssem so I redid it
14:54 jkerssem but now the old one is obsolete
14:55 egonw git branch -D name
14:57 maclean also git branch -l will list the branches
14:59 egonw "Your branch is behind 'github/master' by 41 commits, and can be fast-forwarded."
14:59 egonw yeah, something fishy with egit
15:01 egonw maclean: why is the author line saying you are the author?
15:01 egonw jkerssem: I thought you made the patch?
15:01 jkerssem I did...
15:01 egonw naughty maclean :)
15:01 jkerssem I then posted it to the patch tracker
15:02 egonw maclean: how did you sign the patch then?
15:02 jkerssem I guess maclean missed an entry somewhere
15:02 egonw "git commit --amend --signoff" should have done it? what did you use that made it overwrite the author line?
15:03 jkerssem ah, my fault
15:03 jkerssem there was no author-line in my original patch...
15:03 jkerssem and I think I know why...
15:03 jkerssem lemme check, just a moment
15:09 jkerssem Yup, I think it was my fault
15:09 jkerssem EGit has this nasty hidden checkbox in the "create patch" dialog
15:09 jkerssem that says "format as GIT patch"
15:09 jkerssem it's all alone on the second tab
15:09 jkerssem and unchecked by default
15:10 jkerssem which means there's no author information included
15:10 egonw maclean: happy if I update the author info here locally?
15:12 jkerssem I'd be happy ;-)
15:14 jkerssem okay, branching is indeed a pain in EGit
15:14 jkerssem and hard resets mess everything up..
15:14 jkerssem I just did a hard Reset back to 1.4.x
15:14 jkerssem and there's still things commitable...
15:15 jkerssem seems to be leftovers from ant-build javacc
15:18 egonw from the command line:
15:18 egonw git checkout -- .
15:18 jkerssem ?
15:18 jkerssem what does that do?
15:18 egonw that will undo any chance uncommited
15:18 egonw because of reset gone bad
15:19 CIA-57 org.openscience.cdk: Egon Willighagen bioclipse2.6 * r8e8a19c / plugins/org.openscience.cdk​.smsd/META-INF/MANIFEST.MF : Added a missing dependency on the signatures bundle - http://bit.ly/fHk7Bs
15:19 egonw maclean: went ahead
15:20 CIA-57 cdk: Jules Kerssemakers cdk-1.4.x * r92609d2 / src/main/org/openscience/cdk/io/PDBReader.java :
15:20 CIA-57 cdk: Fixed typo in PDBReader constructor
15:20 CIA-57 cdk: Signed-off-by: maclean <gilleain.torrance@gmail.com> - http://bit.ly/gpHh8d
15:20 egonw not entirely according to the book...
15:20 egonw but it's just a bloody typo fix
15:20 egonw ok, brb
15:21 maclean Sorry, was AFK.
15:22 maclean It was a 'bare' patch, so I applied, committed the change, then signoff-ed.
15:22 jkerssem it's not "just" a typo fix :P
15:23 jkerssem it's my very first branch-ed patch with Eclipse EGit...
15:23 egonw brb
15:23 maclean ok
15:23 jkerssem which immediately explains alllllll the trouble :P
15:23 egonw left #cdk
15:24 maclean it's odd that there was stuff left over from ant javacc
15:24 maclean I would have thought it only touched stuff in the /bin dir, which should be ignored by the .gitignore
15:25 maclean In fact, a lot of stuff is in the .gitignore, including the rather mysterious "*~"
15:28 jkerssem I really don't know how it happened
15:29 jkerssem I did some branch pruning, hard resetting etc. and it all went away
15:29 jkerssem it were 3 .html's with "editor" and "applet" in their names
15:30 jkerssem and joy o joy, I just noticed there's a second typo I could have also fixed :)
15:30 jkerssem exactly the same, but in the other constructor :P
15:32 maclean Heh. Well, at least the author was consistent. Like having "pathes" everywhere in the graph code.
15:32 maclean Or, "higlight".
15:48 egonw joined #cdk
15:54 maclean Does anything bad happen if you have static inner classes created in a static method?
15:55 maclean Like: "public class MyClass { private static class Inner {...} public static boolean myStaticMethod() { Inner inner = new Inner(); } "
15:55 maclean Hmmm. That's not very readable.
15:56 maclean https://github.com/gilleain/stereo/blob/mas​ter/src/stereo/AtomContainerProjector.java
15:57 maclean maclean's link is also http://bit.ly/hqkQ4e
16:13 jbrefort left #cdk
16:18 egonw_ joined #cdk
16:21 egonw__ joined #cdk
16:21 egonw left #cdk
16:22 egonw__ is now known as egonw
16:24 egonw_ left #cdk
16:35 egonw_ joined #cdk
16:36 egonw left #cdk
16:49 egonw_ left #cdk
16:58 maclean left #cdk
18:20 egonw_ joined #cdk
18:21 egonw_ is now known as egonw
18:22 egonw mmm... a shame maclean is gone :(
18:23 egonw jkerssem: you might find it interesting too: BIOINFORMATICS 27 (5): 619-625 MAR 1 2011
18:42 egonw left #cdk
19:05 jbrefort joined #cdk
19:48 egonw joined #cdk
19:52 sneumann joined #cdk
19:54 egonw left #cdk
19:57 sneumann left #cdk
21:33 Gpox left #cdk
22:04 jbrefort left #cdk

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