Camelia, the Perl 6 bug

IRC log for #cdk, 2011-09-12

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

All times shown according to UTC.

Time Nick Message
03:29 egonw joined #cdk
05:09 egonw joined #cdk
05:37 egonw_ joined #cdk
05:48 egonw__ joined #cdk
06:04 egonw joined #cdk
06:32 sneumann joined #cdk
06:35 egonw_ joined #cdk
06:45 sneumann joined #cdk
07:02 mgerlich joined #cdk
07:04 egonw_ joined #cdk
07:06 Gpox joined #cdk
07:31 s_wolf joined #cdk
08:03 jonalv joined #cdk
08:22 maclean joined #cdk
08:24 maclean morning
08:37 maclean lol. looking at sourceforge bug tracker rss feed (for all projects?) and one item is titled "4 < 3 is wrong" :)
08:37 jonalv +1
08:38 maclean Marked "Deleted : Invalid", though...
08:54 jonalv oh
09:07 egonw_ moin moin
09:08 maclean moin^2
09:08 CIA-38 cdk: Egon Willighagen cdk-1.4.x * rdf8e31e / src/main/org/openscience/cdk/de​bug/DebugMolecularFormula.java :
09:08 CIA-38 cdk: Fixed getting the isotope count for a particular isotope
09:08 CIA-38 cdk: Signed-off-by: maclean <gilleain.torrance@gmail.com> - http://git.io/Sdl98Q
09:08 CIA-38 cdk: Egon Willighagen cdk-1.4.x * rd81fc3d / src/main/org/openscience/cdk/de​bug/DebugMolecularFormula.java :
09:08 CIA-38 cdk: Added missing call to super
09:08 CIA-38 cdk: Signed-off-by: maclean <gilleain.torrance@gmail.com> - http://git.io/y24rGg
09:09 maclean NNChemObjectTestHelper.testNotify​Changed_SetFlag(newChemObject()); - no such method...
09:09 egonw_ ok, noted
09:09 egonw_ verifying...
09:15 egonw maclean: I cannot reproduce... it works here...
09:15 maclean hmmm.
09:15 maclean So your "NNChemObjectTestHelper" class has such a method?
09:16 egonw it seems so...
09:16 egonw double checking in the source code right now
09:16 egonw compiling from the command line worked...
09:16 egonw yeah, line 71 and 80
09:17 maclean Hmmmm.
09:18 maclean I see it there on github, too. Odd.
09:21 egonw indeed
09:21 maclean Ah, I hadn't pulled from upstream or whatever. Sometimes I look back longingly for the days of svn update.
09:22 egonw git fetch origin; git checkout cdk-1.4.x; git rebase origin/cdk-1.4.x ?
09:23 maclean Er, no none of that. I generally just want the code, not to remember twenty different commands to re-weave time and space :/
09:24 maclean Anyway, it now compiles, and I'll signoff on it.
09:26 maclean AARGH. NO.
09:26 maclean Not the smarts query tool error! :((
09:27 jonalv maclean: so you know about that? I had that in a branch and just made a quick fix for it
09:27 jonalv maclean: what is that all about?
09:27 maclean jonalv : no idea?
09:28 maclean it's very annoying though. Perhaps it's an eclipse bug, as Asad didn't seem to have it, and he uses netbeans.
09:28 jonalv I am just guessing "the smarts query tool error" is the same one I ahd
09:28 jonalv it just occured from nowhere it seemed...
09:28 jonalv but it was something about exceptions
09:29 maclean line 487 in SmartsQueryTool.
09:29 maclean Unhandled CDKException.
09:29 jonalv yea I just let the method rethrow CDKException in banrhc
09:29 jonalv *branch
09:29 maclean I think I mentioned it on the mailing list.
09:30 maclean Yes, I do that too.
09:30 jonalv sry I am drowning in mails from the cdk list can't keep up...
09:30 maclean fair enough.
09:32 jonalv okey read it
09:32 jonalv maclean: my branch is based on master though
09:32 maclean yes, I didn't have the problem in master before now, though.
09:32 maclean I only just updated.
09:32 jonalv okey
09:33 jonalv maclean: bu line 78 of the SMARTSParser clearly specifies a CDKException and a few lines further down one is thrown so I don't think it's a bug in Eclipse...
09:34 jonalv or do my brnach differ from yours?
09:35 maclean No, I see that on 78.
09:38 maclean jonalv : Ahaha! I see something interesting in commit dfb3e8624323f015628cdbe7202530510d458056
09:38 jonalv maclean: okey, do tell
09:38 maclean one "jonathan alvarsson" is removing a try/catch block...
09:38 maclean :)
09:38 jonalv aha
09:39 jonalv what an evil person
09:39 maclean try "git show dfb3e8624323f015628cdbe7202530510d458056"
09:40 jonalv fatal: ambiguous argument 'git': unknown revision or path not in the working tree.
09:40 maclean Admittedly, the catch block did have the comment "bad practice, but this will never fail"
09:40 maclean Oh.
09:40 egonw I have it
09:40 egonw jonalv: git fetch $FOO
09:40 egonw where $FOO is the main CDK git
09:40 egonw (one of the two)
09:41 jonalv egonw: yea but what is $FOO?
09:41 jonalv oh any one?
09:41 jonalv I still don't have it
09:41 jonalv ah okey now I do :)
09:41 jonalv -me hates mondays
09:43 jonalv hm seems someone changed from CDKException to IllegelargumentException
09:43 * egonw is going for lunch
09:43 jonalv but why did this compile for the reviewer? What am I missing?
09:43 maclean no idea. maybe they did not review properly :)
09:44 jonalv also, why did it compile for me at that time?
09:44 jonalv ...probably same rason...
09:44 jonalv *reason
09:45 maclean Well, my copy is catching a TokenMgrException, and re-throwing an IllegalArgument - is that what you have?
09:45 maclean (in the SQT constructor, that is)
09:45 jonalv maclean: sry, where are we now?
09:46 jonalv yea correct
09:46 maclean line 135-142 smarts query tool constructor.
09:46 maclean ah, ok.
09:47 maclean So, should initializeQuery() throw the CDKException, and be caught in that constructor?
09:47 jonalv I suppose that works
09:47 jonalv not awesome though
09:47 jonalv I mean to throw and rethrow Exceptions should be avoided I think
09:48 jonalv maclean: oh scratch that
09:49 jonalv the actual exception is not thrown in that actual method
09:49 jonalv but by code called by that method
09:49 jonalv so not much we can do about it here
09:49 maclean I suppose.
09:50 maclean If the parser fails, it converts a parse exception to a cdk exception.
09:50 jonalv so yea I think the constructor can take care of it
09:50 jonalv Is there a general policy for how exceptions should be handled in CDK?
09:50 maclean Heh. Not that I know of.
09:51 maclean I think the general philosophy is of simplicity and laziness.
09:51 jonalv for example I think that using a checked CDKException instead of IllegalArgumentException is very bad
09:51 maclean (I didn't mean that in a bad way)
09:51 maclean jonalv : what would be better?
09:51 jonalv to throw the IllegalargumentException?
09:52 jonalv if you call a method with the wrong arguments you get an illegalargumentexception
09:52 jonalv that is sort of reasonable, no?
09:52 maclean I agree.
09:53 jonalv almost all exceptions can be replaced by IllegelargumentException or Illegelstateexception
09:53 jonalv also, I have strong feelings against checked exceptions
09:53 maclean :) This is true.
09:54 maclean There was a long discussion on this on the mailing list (again :) but I'm not sure where I stand on it.
09:54 jonalv and I suppose it didn't end up in a policy?
09:54 jonalv I think the main problem is that there is no real "CDK way" of doing this that we can refer to
09:54 jonalv or is there?
09:56 maclean Not that I know of. A consistent approach would be good, it's true.
09:56 maclean http://www.mail-archive.com/cdk-user​@lists.sourceforge.net/msg02116.html
09:56 zarah maclean's link is also http://tinyurl.com/6hlad62
09:56 jonalv well then you would have something to smash in the head of people trying to make this sort of changes...
09:57 maclean Heh.
09:59 jonalv this discussion is wrong
10:00 jonalv or rather I disagree stronlgy with people
10:00 jonalv :)
10:00 jonalv they seem to have forgotten that you can use the stacktrace to see where the exception occured
10:01 maclean Yes, I wondered that when I was reading it. I thought perhaps I was missing something...
10:01 jonalv Egon is talking about Waves, this feels very old :)
10:03 maclean Right. I'm adding a patch to the tracker to fix this : I may or may not have made the right choice.
10:03 maclean jonalv : could you review it?
10:04 jonalv maclean: yea, how do I do that again? :)
10:04 maclean It's laborious, but you download it, apply it to a branch, check it compiles & makes sense, then git --amend --signoff
10:05 jonalv and pu it back on the patch site?
10:05 jonalv will do
10:05 maclean then git format-patch -1 and re-upload to s/f.
10:05 jonalv just send a link
10:05 maclean Ok, got to make the s/f item first.
10:05 jonalv I like this guy: http://www.mail-archive.com/cdk-user​@lists.sourceforge.net/msg02136.html
10:05 zarah jonalv's link is also http://tinyurl.com/6csxc3e
10:05 jonalv he speaks the truth
10:10 maclean https://sourceforge.net/tracker/?func=detail&amp​;aid=3408000&amp;group_id=20024&amp;atid=320024
10:10 maclean runtime exceptions, eh? If only I could remember what those are.
10:10 jonalv the ones where you don't need to write all the boilerplate
10:11 maclean Yeah, but.
10:11 maclean That means you are using the CDK more like a scripting language.
10:11 maclean Which is fine : just use python (jython).
10:12 jonalv hm
10:13 jonalv I am beginning to see your point as well
10:13 maclean It's all about what kind of language people prefer. The larger the library, the more of a 'dialect' it becomes in a language.
10:13 jonalv we have a checked Bioclipse Exception in Bioclipse as well. Which I also stronlgy hate
10:14 jonalv but I still think to just throw a CDKException is wrong
10:14 maclean It should be finer grained?
10:14 jonalv maybe if you have a ChemistrySpecifikExceptionN externds CDKExcpetion then it makes sense
10:14 jonalv ye
10:14 jonalv a
10:15 jonalv either that or just use one of the java ones which are created for just that purpose
10:17 jonalv maclean: um how do I aply the patch. git apply seems to do the wrong thing
10:17 maclean Perhaps some kind of review needs to be done, to see if there already exist 'natural' subclasses of CDKException, or if some of them can be replaced.
10:18 maclean jonalv : Oh. "git am -3 $patch"
10:18 maclean ?
10:18 jonalv -3?
10:18 maclean That falls back to a 3-way merge, so there may be ">>>>" lines.
10:18 jonalv uhm, okey :)
10:19 maclean As in, it will complain about stuff. Then you have to fix the stuff in the file, then "git add", then "git am --continue". Sigh. So many random things to remember :/
10:23 jonalv maclean: do I change to accepted ones I ahve looked at it?
10:23 maclean Er, I don't actually know. I wonder that myself : perhaps egonw can tell?
10:25 jonalv hm he is on lunch
10:25 jonalv I should probably eat as well :)
10:25 maclean okey.
10:26 jonalv seems impossible to close it
10:26 jonalv so I don't know if only the person who applies should mark as Accepted
10:27 jonalv oh found the closed option
10:27 jonalv I will change to Accepted
10:27 jonalv and hope for fogiveness if it was wrong
10:28 jonalv I would love to see something like this: http://www.bugzilla.org/doc​s/2.18/html/lifecycle.html
10:29 maclean Heartily agree. I very much dislike s/f bug tracking system.
10:31 * jonalv afk for lunch
10:31 egonw back
10:32 egonw yeah, you change to accepted
10:32 maclean ah, ok.
10:32 egonw if there has been sufficiently reviewing
10:32 egonw which is one review for bug fixes with unit tests in cdk-1.4.x
10:32 egonw and *two* for API or functional changes
10:33 maclean Hmmm. Not sure which category the SQT one falls into.
10:35 egonw the patch by jonalv?
10:35 egonw yeah, two reviews
10:35 maclean Grmph.
10:35 egonw but Rajarshi also looked at it already, not?
10:36 maclean ?
10:36 maclean Not that I know of.
10:36 maclean https://sourceforge.net/tracker/?func=detail&amp​;atid=320024&amp;aid=3408000&amp;group_id=20024
10:36 maclean (that one)
10:37 maclean Ah, I think you were thinking of 3313187
10:37 egonw dunno
10:37 egonw brain does not function well today
10:42 CIA-38 cdk: Egon Willighagen cdk-1.4.x * r294cecf / src/test/org/openscience/cdk​/nonotify/NNMappingTest.java :
10:42 CIA-38 cdk: Added two missing tests to overwrite the notifying default implementation
10:42 CIA-38 cdk: Signed-off-by: maclean <gilleain.torrance@gmail.com> - http://git.io/B55AGw
10:49 maclean egonw : so how do I annotate the SQT patch as needing pushing to cdk/cdk? I made it, jonathan reviewed it, now...
11:17 egonw sign-off, mark as accepted, reassign to me and/or Rajarshi
11:18 maclean Ah, steps 1-2 done. Doing step 3.
11:19 maclean It's quite important, as it means I can't (easily) work on cdk without it :)
12:56 jbrefort joined #cdk
13:02 CIA-38 cdk: maclean master * rface5c5 / src/main/org/openscience/cdk/smi​les/smarts/SMARTSQueryTool.java :
13:02 CIA-38 cdk: Handle a CDKException thrown by the Parser by throwing an IllegalArgumentException in the constructor
13:02 CIA-38 cdk: Signed-off-by: jonalv <jonathan.alvarsson@gmail.com> (+33 more commits...) - http://git.io/D3jlZw
13:02 maclean \o/
13:02 jonalv uhm
13:02 jonalv +33 more commits?
13:03 egonw jonalv: you were happy with all of it, not?
13:03 egonw why did you sign off that last patch otherwise?
13:03 jonalv I only checked the changes in the last patch
13:03 jonalv I signed of a patch
13:04 maclean I think the other 33 commits are merging of cdk-1.4.x into master
13:04 egonw maclean: right
13:04 egonw jonalv: it's just the CIA being cryptic
13:04 maclean Don't worry, jonalv : I wasn't sneaking past a whole load of code on your say-so :)
13:04 jonalv okey, you don't do that very often do you?
13:04 egonw last time was Aug 29
13:05 egonw so, a whopping 2 weeks ago
13:05 jonalv wow 33 commits in 2 weeks
13:05 jonalv nice!
13:05 jonalv well 34
13:05 egonw no need to be sarcastic :)
13:05 jonalv I thin it was much
13:05 egonw oh carp.... it feels like Fri
13:06 egonw I know, I know...
13:06 jonalv I mean check how many Bioclipse has in that time
13:06 jonalv considering gpox was on vacation I would guess very few...
13:06 jonalv :)
13:06 egonw well, it took 3 months to develop those patches
13:07 jonalv egonw: btw, if you feel like generating more you could have another look at my klekotha root (spelling) branch
13:07 jonalv I think it should be ready to be applied soon
13:07 jonalv soonish (TM)
13:07 jonalv :)
13:07 egonw send me an email
13:07 jonalv oki
13:07 maclean git log --oneline --grep="^.. final"
13:08 jonalv huh?
13:08 maclean That's one reason for so many patches ^_^
13:09 jonalv what smiley is that?
13:09 jonalv my IRC client drew a cute girl
13:09 jonalv lol
13:10 maclean It was one that looked a little like my regex. It is indeed non-Western, like =^_^=
13:10 maclean (cat?)
13:10 jonalv hm maybe
13:34 maclean g2g
13:35 jonalv oh egonw you updated master, maybe I should rebase my branch then :)
14:41 egonw joined #cdk
14:46 egonw joined #cdk
14:51 maclean joined #cdk
14:53 egonw_ joined #cdk
14:56 * maclean is finally writing test cases for a patch that is 2 years old...
14:57 jonalv maclean++
14:57 maclean well, it shouldn't have taken me so long, really.
15:25 egonw__ joined #cdk
15:46 jbrefort joined #cdk
15:48 egonw joined #cdk
16:11 jbrefort joined #cdk
16:13 maclean joined #cdk
17:32 egonw joined #cdk
17:40 maclean joined #cdk
18:13 sneumann joined #cdk
19:34 sneumann joined #cdk

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