Camelia, the Perl 6 bug

IRC log for #cdk, 2010-08-30

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

All times shown according to UTC.

Time Nick Message
03:35 azeem_ joined #cdk
03:39 azeem left #cdk
05:05 egonw joined #cdk
07:02 Gpox joined #cdk
07:09 sneumann joined #cdk
07:33 s_wolf joined #cdk
07:38 s9asad joined #cdk
07:38 s9asad Hi
07:38 zarah hello s9asad, you fantastic person you
07:39 s9asad :-) wow what a stub
07:39 s9asad egonw: I tried your SMSD patch
07:40 s9asad but i am unable to apply it
07:40 s9asad maybe I am missing something
07:49 bag_ joined #cdk
07:52 egonw s9asad: what patch?
07:52 s9asad 0001-Latest-SMSD-code-1.2.0-Major-changes-are
07:54 egonw that's your patch...
07:54 egonw I just collapsed your patches, and removed the 0003
07:54 egonw this is the one that we applied last night
07:56 s9asad so you mean it already applied and pused?
07:56 s9asad or do you think I should generate complete patch and upload it
07:56 s9asad paches*
07:57 egonw no further work on that one needed...
07:57 egonw it should be OK...
07:57 s9asad oh thanks
07:57 egonw but please do feel free to glance over it and check
07:58 egonw btw, was I clear on why I felt I had to collapse the patch?
07:58 egonw that's more important now
07:58 s9asad yes I rerun it and cross check with my local
07:58 s9asad hmm very much!
07:59 egonw ok, good
07:59 s9asad so if syn with you upstream branch I will have these changes?
07:59 egonw what do you mean?
07:59 s9asad I mean the SMSD patch is already in the CDK
07:59 s9asad 1.2.0
07:59 egonw ah... "so if *I* sync with you*R* upstream branch..."
08:00 egonw yes, you will :)
08:00 s9asad :-) sorry still not awake ;-)
08:00 s9asad Thanks alot!
08:01 egonw do you have enough lead on the more efficient SMARTS parsing?
08:01 s9asad yes now I will work on SMART
08:01 egonw by using that existing 'noReinitialize' pattern Rajarshi described?
08:02 s9asad I have make some changes in the VFlib itself which might make it much better
08:03 s9asad right now a boolean check in the SMARTQueryTool .matches() has to generate all subgraphs
08:03 s9asad were as If I can check for one subgraph and return true or false.
08:03 s9asad This will make it faster
08:04 egonw OK, was wondering about that too...
08:04 s9asad and if someone wants all subgraph then we can have a different method for that
08:04 egonw yes, reply that to the list...
08:04 s9asad sure
08:04 egonw the need for a method: public boolean isSubgraph()
08:04 egonw the need for a method: public boolean isSubgraph(IMolecule)
08:05 s9asad yes
08:05 s9asad indeed
08:06 egonw that method can indeed take the shortcut, and return after finding the first hit
08:07 s9asad yes the VFlib is very good is doing that job
08:07 s9asad just a boolean query isSubgraph()
08:09 s9asad other algorithm could be ullmann but I don't have time for that but someone interested can code it
08:17 egonw s9asad: http://github.com/cdk/cdk/commits/master
08:17 zarah egonw's link is also http://tinyurl.com/2u6zetj
08:17 egonw that's the main repository now
08:19 s9asad oh
08:19 s9asad I was looking in your branch
08:20 egonw yeah, guess I forgot to sync that last night
08:20 s9asad no worries, thanks now I will update mine
08:29 s9asad left #cdk
08:36 s9asad joined #cdk
08:36 s9asad egonw: the SMSD patch looks one commit behind
08:37 s9asad I made some fixes which are missing
08:37 s9asad should I submit them now
08:37 egonw yes please
08:37 egonw write good commit messages, explaining what was wrong
08:40 egonw you also saw this:
08:40 egonw http://pele.farmbio.uu.se/ni​ghtly/test/result-smsd.html
08:40 zarah egonw's link is also http://tinyurl.com/36htaba
08:40 egonw ?
08:41 s9asad sure
08:42 s9asad so I would update the code and submit the pacthes
09:05 azeem_ is now known as azeem
09:20 jbrefort joined #cdk
09:31 bag_ left #cdk
11:09 s9asad hi egonw:
11:09 egonw hi s9asad
11:09 s9asad the missing SMSD patches should be uploaded as new
11:09 s9asad or appended to the old
11:09 egonw new
11:14 s9asad done
11:43 egonw s9asad: patch 0001: the commit message is incorrect
11:43 egonw it does not just add IMatcher, it also removes many files... ??
11:44 egonw e.g. it remov Isomorphism.java which is added by 0002 again ??
11:44 egonw 0003: new methods in this patch do not have JavaDoc
11:46 egonw 0004: you could split up that patch into: 0004a: remove outcommented code; 0004b: updated for change in index API: what used to be null is now reflected as -1 (which I guess what is going on)
11:47 egonw 0005: you import the wrong Assert class... use the org.junit.Assert instead
11:48 egonw "a)constructor call fixed" -> bad commit message... instead, describe what was wrong, and *how* you fixed it
11:49 egonw (that was 0006)
11:50 egonw 0007: there is no need to add empty methods for @BeforeClass, @AfterClass, and @After ... empty constructors are not needed either
11:51 egonw oh.... and 0001 does not even add an IMatcher interface as promised in the commit message
11:53 egonw please let me know where I need to be more verbose and give more detailed explanation
12:13 s9asad hi
12:13 zarah saluton s9asad
12:14 s9asad egonw: how can I update set of files without removing it?
12:15 egonw by modifying the content?
12:15 egonw I guess I am not sure what you mean...
12:15 egonw but a plain text editor comes to mind :)
12:15 s9asad if I am going to redo the whole excerise one by one then will be repeating the same ..
12:16 s9asad The brach with 1.2.0 code has files which was re-factored
12:16 egonw indeed
12:16 egonw but you can use git powers, not?
12:17 egonw you have to give me more detail on what the problem is
12:17 s9asad thats what I did in my netbeans!
12:17 egonw then netbeans sucks
12:17 s9asad copied the local code on top of the master
12:19 s9asad I think when  it deletes the code and adds it
12:19 s9asad if you think another way of doing it I can try
12:20 egonw try using the git from the command line
12:21 egonw git itself will never do something as stupid as deleting files then adding it again
12:21 egonw that is seriously damaging the history of files...
12:22 s9asad now I need your suggestion in this scenario
12:22 s9asad a) I have an updated code on my local branch
12:23 s9asad b) I would like to update the code in the master
12:23 s9asad how to proceed
12:23 egonw or both online on github?
12:24 s9asad yes both are online on the github
12:24 s9asad http://github.com/asad/cdk/co​mpare/0f64e4da78...50a070d46c
12:24 zarah s9asad's link is also http://tinyurl.com/33wh22r
12:24 s9asad my local commits
12:25 egonw I do not see the branch online
12:25 egonw only master and cdk-1.2.x
12:25 egonw oh...
12:25 egonw s9asad: always put your code in branches...
12:25 egonw much easier to work with then patches in your copy of master
12:26 egonw oh, and you do not need to sign-off your own patches...
12:26 egonw I'm sure you are happy with them :)
12:26 s9asad :-)
12:26 s9asad git remote add upstream git://github.com/
12:26 s9asad git fetch upstream
12:27 s9asad is what I am using
12:27 s9asad git fetch upstream
12:27 s9asad git merge upstream/master
12:27 s9asad thats why I didn't have branches
12:27 egonw ok...
12:27 egonw let's teach you some git magic
12:27 s9asad but yes if branches help I can do that
12:27 egonw got a command line?
12:27 egonw got some time?
12:27 s9asad yes
12:27 s9asad sure
12:27 s9asad always ready to learn
12:30 egonw ok, this is what you are going to do:
12:30 egonw merge some stuff, edit patches etc...
12:30 s9asad sure
12:30 egonw git rebase, to be precies
12:30 egonw git rebase, to be precise
12:31 s9asad hmm, I need to learn this
12:31 egonw yes, you need
12:31 egonw no more NetBeans
12:31 s9asad :-)
12:31 egonw NetBeans evil
12:31 egonw s9asad master
12:31 egonw s9asad full control
12:31 egonw s9asad work with command line
12:31 egonw (OK, now I'll cut with the crap talk :)
12:31 s9asad lets start
12:31 s9asad !
12:32 s9asad I am think others will also learn from it
12:33 egonw OK, let's see...
12:33 egonw first, I suggest to merge the first two patches
12:33 egonw that should fix the removal and adding of files...
12:33 s9asad okay
12:34 egonw ok, we 7 patches...
12:34 egonw so we type:
12:34 egonw git rebase -i HEAD~7
12:34 s9asad where in my local branch?
12:34 s9asad or cdk/cdk
12:34 egonw right... let's fix that first...
12:34 egonw your local git repos
12:34 egonw do this first:
12:34 egonw git checkout master
12:35 egonw git checkout -b 1-m-fixingMissingPatches master
12:35 s9asad Already on 'master'
12:35 egonw "1-" for your first development branch
12:35 s9asad Switched to a new branch '1-m-fixingMissingPatches'
12:35 egonw "-m-" indicating it's branched from master
12:35 s9asad k
12:35 egonw "-fixingMissingPatches" as a descriptive text so that you can later easily see what that branch is about
12:36 egonw git checkout master
12:36 egonw (return to master)
12:36 egonw and tell me what the alias for upstream cdk/cdk is...
12:36 egonw you use 'upstream', correct=
12:36 egonw you use 'upstream', correct?
12:36 s9asad yes
12:36 s9asad upstream
12:36 egonw ok, do this (after you are back on master)
12:37 egonw git reset --hard upstream/master
12:37 egonw git push origin 1-m-fixingMissingPatches
12:37 s9asad git reset --hard upstream/master
12:37 s9asad HEAD is now at 50a070d a)missing initialization errors fixed
12:37 egonw git push -f origin master
12:37 egonw mmm... that does not look good
12:38 egonw paste me the output of:
12:38 egonw git remote -v | grep upstream
12:38 egonw and
12:38 egonw git remote -v | grep origin
12:38 s9asad dksmsdgithub Asad$ git remote -v | grep upstream
12:38 s9asad upstreamgit@github.com:asad/cdk.git (fetch)
12:38 s9asad upstreamgit@github.com:asad/cdk.git (push)
12:38 s9asad git remote -v | grep origin
12:38 s9asad origingit@github.com:asad/cdk.git (fetch)
12:39 s9asad origingit@github.com:asad/cdk.git (push)
12:40 egonw so, your 'upstream' != cdk/cdk
12:41 egonw try:
12:41 egonw git remote rm upstream
12:41 egonw git remote add upstream git://github.com/cdk/cdk.git
12:41 s9asad did
12:41 egonw ok, then the output of this again: git remote -v | grep upstream
12:42 s9asad upstreamgit://github.com/cdk/cdk.git (fetch)
12:42 s9asad upstreamgit://github.com/cdk/cdk.git (push)
12:43 s9asad now
12:44 egonw ok, better :)
12:44 egonw git reset --hard upstream/master
12:44 s9asad ok
12:44 s9asad fatal: ambiguous argument 'upstream/master': unknown revision or path not in the working tree.
12:44 s9asad Use '--' to separate paths from revisions
12:46 egonw sorry... do first:
12:46 egonw git fetch upstream
12:46 egonw then repeat:
12:46 egonw git reset --hard upstream/master
12:47 s9asad HEAD is now at 0f64e4d Updates for the getImplicitHydrogenCount() renaming
12:47 egonw excellent
12:47 egonw ok, next step: back to your branch
12:47 egonw git checkout  1-m-fixingMissingPatches
12:47 egonw do you have 'tig' installed?
12:47 s9asad ?
12:48 s9asad no
12:48 egonw or gitk or gitx ?
12:48 s9asad gitk yes
12:48 egonw tig -> git.wiki.kernel.org/index.php/Tig
12:48 egonw ok, thatäs fine
12:49 egonw you're in the 1- branch now?
12:49 egonw if so, fire up gitk
12:49 s9asad yes
12:50 s9asad done
12:50 egonw ok, verify that 1-m-bla is 7 patches ahead of master
12:50 egonw confirmed?
12:50 s9asad yes
12:51 egonw perfect
12:51 egonw so, what we did so far was:
12:51 egonw 1. make a development branch
12:51 s9asad yes
12:51 egonw 2. clean master, to be a clean copy of upstream/master
12:51 egonw next: we'll make a backup of the 7 patches
12:51 s9asad yes very important
12:52 egonw git checkout -b 2-m-backupOfSevenOriginalPatches
12:52 egonw git checkout 1-m-fixingMissingPatches
12:52 egonw done
12:52 s9asad done
12:52 egonw now we get to the 'git rebase' work
12:52 s9asad ok
12:52 egonw ok, first we'll merge the first two patches:
12:52 s9asad yes
12:53 egonw git rebase -i HEAD~7
12:53 egonw that should open up an editor... please confirm
12:53 s9asad gitk?
12:54 s9asad should I close my gitk
12:54 s9asad then fire this command and see what happens
12:55 s9asad I think its a vim on commandline
12:56 s9asad asad:cdksmsdgithub Asad$ git rebase -i HEAD~7
12:56 s9asad pick 029c09f IMatcher added Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
12:56 s9asad pick 23095a2 a)Refactored matchers, atom matcher and bond matcher b)getTanimoto score fixed for single atoms
12:56 s9asad pick 52e9b7c a) Default constructor supported as per changes in the CDK b) VFMCS index error should point to -1 not null Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
12:56 s9asad pick 88b68dc a) if index -1 throw exception Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
12:56 s9asad pick a2ae076 a)Global Assert defined Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
12:56 s9asad pick 4a7a44d a)constructor call fixed Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
12:56 s9asad pick 50a070d a)missing initialization errors fixed Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
12:56 s9asad # Rebase 0f64e4d..50a070d onto 0f64e4d
12:56 s9asad #
12:56 s9asad # Commands:
12:56 s9asad #  p, pick = use commit
12:56 s9asad #  r, reword = use commit, but edit the commit message
12:56 s9asad #  e, edit = use commit, but stop for amending
12:56 egonw ok, that looks good
12:56 egonw so, we see here a lot of comment lines starting with #
12:56 egonw that contains some help
12:56 s9asad ok
12:57 egonw we also see seven 'pick' lines, one for each of the 7 patches
12:57 egonw OK?
12:57 s9asad yes
12:57 s9asad P-R-E-S-F
12:58 s9asad so whats next
13:00 egonw sorry, had to do something
13:00 egonw was briefly afk
13:00 egonw so, we are going to merge patch one and two
13:01 egonw the list of seven patches is oldest in the top line, newest in the bottom line...
13:01 egonw now, you want to merge patch 0001 and 0002, so the top two lines...
13:01 s9asad okay
13:01 s9asad how
13:01 egonw that means you will say that the *second* patch is to be squashed with the first
13:01 s9asad by test edit?
13:01 s9asad text*
13:02 egonw so, you replace the 'pick' text on the *second* line with 's' (short for squash), so that you get:
13:02 egonw pick e4d99cd IMatcher added Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:02 egonw s 0ecceee a)Refactored matchers, atom matcher and bond matcher b)getTanimoto score fixed for single atoms
13:02 egonw then, save the file and exit
13:03 egonw this will then open a new editor which starts with this text:
13:03 egonw # This is a combination of 2 commits.
13:03 s9asad This is a combination of 2 commits.
13:03 s9asad # The first commit's message is:
13:03 s9asad IMatcher added
13:03 s9asad Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:03 s9asad # This is the 2nd commit message:
13:03 s9asad a)Refactored matchers, atom matcher and bond matcher
13:03 s9asad b)getTanimoto score fixed for single atoms
13:03 s9asad Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:03 s9asad # Please enter the commit message for your changes. Lines starting
13:03 s9asad # with '#' will be ignored, and an empty message aborts the commit.
13:03 s9asad # Not currently on any branch.
13:03 s9asad # Changes to be committed:
13:03 s9asad #   (use "git reset HEAD <file>..." to unstage)
13:03 s9asad #
13:04 egonw correct
13:04 egonw now, every line that start with a # is a comment line, and not part of the commit message, and only for guidance in the editor
13:04 egonw now, you should see three lines which do not start with a '#'
13:05 egonw those are the commit message...
13:05 egonw I see:
13:05 egonw IMatcher added Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:05 egonw a)Refactored matchers, atom matcher and bond matcher b)getTanimoto score fixed for single atoms
13:05 egonw Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:05 s9asad yes
13:05 egonw now, the first line is obsolete and comes from a very old patch...
13:06 s9asad aha
13:06 egonw the patch does not actually refer to a IMatcher.java at all
13:06 s9asad yes indeed
13:06 egonw so, the second line is what is interesting instead
13:06 s9asad yes
13:06 egonw so, remove the first, and the third line
13:06 egonw (no need to sign-off your own commits; leave that to the reviewers)
13:06 egonw and do any further editing of the commit message as you like
13:07 s9asad a)Refactored matchers, atom matcher and bond matcher
13:07 s9asad b)getTanimoto score fixed for single atoms
13:07 egonw e.g. put a) and b) on separate lines
13:07 egonw indeed :)
13:07 s9asad :-)
13:07 egonw OK, then, save and exit
13:07 s9asad Total 136 (delta 68), reused 0 (delta 0)
13:07 s9asad warning: There are too many unreachable loose objects; run 'git prune' to remove them.
13:07 s9asad Successfully rebased and updated refs/heads/1-m-fixingMissingPatches.
13:08 egonw perfect
13:08 egonw OK, now my IRC client needs rebooting... brb
13:08 s9asad sure
13:08 egonw left #cdk
13:09 egonw joined #cdk
13:09 egonw s9asad: OK, fire up a new gitk (kill any old running versions)
13:10 egonw and notice that there are now only 6 patches left
13:10 s9asad yes
13:11 egonw mom, looking up my other comments
13:11 s9asad :-)
13:11 egonw ok, comment on 0003:
13:12 egonw new methods in this patch do not have
13:12 egonw JavaDoc
13:12 egonw this means, you will have to 'edit' this patch...
13:12 egonw here goes:
13:12 s9asad hmm
13:12 egonw git rebase -i HEAD~6
13:12 s9asad okay
13:12 s9asad and editor
13:13 egonw right
13:13 egonw now, the 0003 patch is now the second
13:13 egonw now, the 0003 patch is now the cccebff29e2f4c9d7fb69f8ada9d0aba928571a1second
13:13 egonw now, the 0003 patch is now the cccebff29e2f4c9d7fb69f8ada9d0aba928571a1
13:13 egonw or cccebff2 is it will likely appear in the editor
13:13 egonw replace 'pick' by 'e', short for edit
13:13 egonw save and exit
13:13 egonw you will see this message:
13:13 s9asad wait
13:14 egonw Stopped at cccebff...
13:14 s9asad I can't find
13:14 egonw You can amend the commit now, with ...
13:14 s9asad pick 0c65030 a)Refactored matchers, atom matcher and bond matcher b)getTanimoto score fixed for single atoms
13:14 s9asad pick 19f04a4 a) Default constructor supported as per changes in the CDK b) VFMCS index error should point to -1 not null Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:14 s9asad pick 6de62c6 a) if index -1 throw exception Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:14 s9asad pick 94a45b9 a)Global Assert defined Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:14 s9asad pick 65db93d a)constructor call fixed Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:14 s9asad pick c874604 a)missing initialization errors fixed Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:14 egonw oh, right...
13:14 egonw you're in a different git universe and have different hashes...
13:14 egonw sorry, my mistake
13:14 egonw 19f04a4
13:14 egonw that's the one
13:14 egonw replace 'pick' by 'e', short for edit
13:14 egonw save and exit
13:15 egonw you will see this message:
13:15 egonw Stopped at cccebff...
13:15 egonw You can amend the commit now, with ...
13:15 egonw git commit --amend
13:15 s9asad Stopped at 19f04a4... a) Default constructor supported as per changes in the CDK b) VFMCS index error should point to -1 not null Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:15 s9asad You can amend the commit now, with
13:15 s9asad git commit --amend
13:15 s9asad Once you are satisfied with your changes, run
13:15 s9asad git rebase --continue
13:15 egonw Once you are satisfied with your changes, run
13:15 egonw right
13:15 egonw good
13:15 egonw now, this is what happened:
13:15 egonw it applied the first patch (as you selected 'pick')
13:15 s9asad aha
13:16 egonw it applied the second patch (as you selected 'edit') but stopped *after* applying it...
13:16 s9asad hmm
13:16 egonw now you can add the missing JavaDoc to the source files
13:16 egonw and run git add on those changed .java files
13:16 egonw and then use (IMPORTANT!) git commit --amend
13:16 s9asad how will I know about those files
13:17 s9asad via gitk?
13:17 egonw until the full JavaDoc is in order and 'git commit --amend'-ed... and you type git rebase --continue (to apply all other 'pick'-ed patches
13:17 egonw or just: git show
13:17 egonw which will show the more recent patch
13:17 egonw which is the one you marked with 'edit'
13:17 egonw in this case
13:18 s9asad now #       modified:   src/main/org/openscience/cdk/smsd/algorith​m/mcsplus/GenerateCompatibilityGraph.java
13:18 s9asad #       modified:   src/main/org/openscience/cdk/sms​d/algorithm/mcsplus/MCSPlus.java
13:18 s9asad #       modified:   src/main/org/openscience/cdk/smsd/a​lgorithm/rgraph/CDKRMapHandler.java
13:18 s9asad #       modified:   src/main/org/openscience/cdk/smsd/a​lgorithm/single/SingleMapping.java
13:18 s9asad #       modified:   src/main/org/openscience/cdk/smsd/a​lgorithm/vflib/VFlibMCSHandler.java
13:18 s9asad #       modified:   src/main/org/openscience/cdk/smsd/algori​thm/vflib/builder/TargetProperties.java
13:18 s9asad #       modified:   src/main/org/openscience/cdk/smsd/a​lgorithm/vflib/map/VFMCSMapper.java
13:19 s9asad so I will open then in command line and add docs
13:19 egonw how did you get that?
13:19 s9asad git show
13:19 egonw oh... I guess you have some default options then :)
13:19 egonw or maybe I do :)
13:19 s9asad :-)
13:20 egonw but looks OK
13:20 egonw so, there you have the changed files
13:21 s9asad I will add comments, 1 min
13:24 s9asad doen
13:24 s9asad now what
13:25 s9asad I added a comment in all the files
13:25 egonw 'git add' the changed files
13:25 s9asad as they require default constructor
13:25 egonw and then do: 'git commit --amed'
13:25 egonw well, as long as you add the JavaDoc too, I am happy with comments
13:26 s9asad now an editor
13:28 egonw sorry, I do not know what you mean with that...
13:28 s9asad git commit --amed
13:28 s9asad has lead to an vim editor
13:28 s9asad should I save and quit
13:28 s9asad amend*
13:29 egonw right, it will likely ask you to (optionally) update the commit message...
13:29 egonw if no changes are needed, you save and quit
13:29 s9asad git commit --amend
13:29 s9asad [detached HEAD 9fb2ca1] a) Default constructor supported as per changes in the CDK b) VFMCS index error should point to -1 not null Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:29 s9asad 7 files changed, 31 insertions(+), 8 deletions(-)
13:29 egonw ok, run: git show
13:29 egonw and check that the patch is now in order
13:30 egonw if not, repeat the process
13:30 egonw of editing files, git add, git commit --amend
13:30 egonw until satisfied
13:31 s9asad looks okay
13:31 egonw ok, then do:
13:31 egonw git rebase --continue
13:31 s9asad warning: There are too many unreachable loose objects; run 'git prune' to remove them.
13:31 s9asad Successfully rebased and updated refs/heads/1-m-fixingMissingPatches.
13:31 egonw perfect
13:32 egonw nw, fixes for 0005 and 0007 can be done with the 'edit' command too...
13:32 s9asad so
13:32 egonw fix the patch, git add, git commit --amend, and git rebase --continue when happy about the changed patch
13:32 egonw now, 0004 and 0006 are different
13:33 egonw 0006 -> 'reword' is the replacement here... which will allow you to write a better commit message...
13:33 egonw (you may want to check what the patch is exactly about...)
13:33 egonw patch 0004 you need to do the following:
13:33 egonw 'edit' the patch...
13:33 egonw like you just did with 0003
13:33 egonw but...
13:33 s9asad so how to follow that
13:34 egonw to split that patch up, you will not 'amend' the patch, but do something fancy...
13:34 egonw (but you may first like to make a new backup of the current state)
13:34 egonw git checkout -b 3-m-secondBackup
13:34 egonw git checkout 1-m-foo
13:35 egonw here is a good explanation to 'split' patches:
13:35 egonw http://www.jukie.net/bart/blog/20081112150409
13:35 zarah egonw's link is also http://tinyurl.com/36tkmdp
13:35 egonw step 1 you have already done before
13:35 egonw HEAD~10 is the last ten patches...
13:36 egonw HEAD~6 is what you have left in your branch on top of master
13:36 egonw best to not rebase patches from master itself
13:36 s9asad okay
13:36 egonw now, the splitting happens in this blog in step 2-7
13:36 egonw step 2) is crucial
13:37 egonw it 'undoes' the previous patch
13:37 s9asad hmm
13:37 egonw allowing you to then commit the right bits in the right order
13:37 egonw to commit only part of the changes in *1* file...
13:38 egonw use: git add -p
13:38 egonw that will allow you to interactively add the bits you want in one patch
13:38 s9asad so
13:38 egonw it will iteratively walk you through the patch blocks in the file...
13:38 s9asad so I should first do what
13:39 egonw for each block it will ask you to "yes, stage this block", or "no, do not stage this block"...
13:39 egonw but you can also sometimes split blocks into smaller blocks with the 's' command
13:39 egonw git rebase -i HEAD~6
13:39 s9asad oh gosh now I am confused :-(
13:39 egonw mark the fourth patch with 'edit'
13:39 egonw and then step 2) in that blog post
13:40 s9asad you mean git reset HEAD~1
13:40 egonw that's step 2), yes
13:40 s9asad Unstaged changes after reset:
13:40 s9asad Msrc/test/org/openscience/cdk/smsd/alg​orithm/rgraph/CDKRMapHandlerTest.java
13:41 s9asad now
13:41 s9asad what should I do
13:41 egonw I think you got the wrong patch...
13:42 s9asad oh
13:42 egonw git rebase --abort
13:42 s9asad done
13:42 s9asad then
13:42 egonw and start again with: git rebase -i HEAD~6
13:43 egonw "a) if index -1 throw exception"
13:43 egonw that's patch 0004
13:43 s9asad pick 0c65030 a)Refactored matchers, atom matcher and bond matcher b)getTanimoto score fixed for single atoms
13:43 s9asad pick 9fb2ca1 a) Default constructor supported as per changes in the CDK b) VFMCS index error should point to -1 not null Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:43 s9asad pick d5d6674 a) if index -1 throw exception Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:43 s9asad pick ca3fb3c a)Global Assert defined Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:43 s9asad pick 18dc1b2 a)constructor call fixed Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:43 s9asad pick 77f49d6 a)missing initialization errors fixed Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:43 egonw remember that you merged two patches...
13:43 egonw so, 0004 is now the third patch
13:43 egonw with 000x I follow the original numbers
13:43 s9asad d5d6674
13:45 s9asad if index -1 throw exception?
13:45 s9asad sorry did u mean pick d5d6674 a) if index -1 throw exception Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
13:46 egonw yes
13:46 egonw yes, 'edit' that patch
13:46 egonw then split following step 2) from the blog
13:46 s9asad Unstaged changes after reset:
13:46 s9asad Msrc/main/org/openscience/cdk/smsd/algor​ithm/vflib/VFlibSubStructureHandler.java
13:46 s9asad Msrc/main/org/openscience/cdk/smsd/al​gorithm/vflib/VFlibTurboHandler.java
13:47 egonw correct
13:48 s9asad now what should i do
13:48 egonw now, you want to make *two* patches
13:48 egonw one with removal of the outcommented code
13:48 egonw the other with the index change
13:48 egonw type: git add -p
13:48 s9asad ic
13:48 egonw the first block has only the index check change for VFlibSubStructureHandler.java
13:48 egonw accept that by typing 'y'
13:48 egonw reject the next one by typing 'n'
13:49 egonw (that's the removal of the outcommented code in the same file)
13:49 egonw then accept (y) the index in the second file,
13:49 egonw and reject (n) with removal of the outcommented code in the second file
13:49 s9asad done
13:49 egonw verify that your staged changes are in order with:
13:49 egonw git diff --cached
13:49 egonw or
13:50 egonw git diff --staged
13:50 egonw which should only show the index blocks
13:50 egonw now, commit these changes with:
13:50 s9asad yes
13:50 s9asad git add?
13:50 s9asad then git commit?
13:50 egonw git commit -m "Updated checking of indices which now are -1 if unset, instead of null"
13:50 egonw no 'git add'
13:51 egonw you already did that with: git add -p
13:51 s9asad pdated checking of indices which now are -1 if unset, instead of null
13:51 s9asad 2 files changed, 2 insertions(+), 2 deletions(-)
13:51 egonw perfect
13:51 egonw now see that you have left unchanged:
13:51 egonw git diff
13:52 egonw that should show the removal of the outdated code
13:52 s9asad yes
13:52 egonw stage those changes
13:52 egonw with (as they are the only changes left): git add -u
13:52 egonw again, you can verify with: git diff --staged
13:52 s9asad yes
13:53 s9asad so commit
13:53 egonw then: git commit -m "Removed unused code"
13:53 s9asad 2 files changed, 1 insertions(+), 301 deletions(-)
13:53 egonw now you have splitted one dirty patch, into two clean ones
13:53 s9asad wow
13:53 s9asad complex but effective
13:53 egonw git rebase --continue to finish this round of changes
13:54 egonw s9asad: well done!
13:54 s9asad Successfully rebased and updated refs/heads/3-m-secondBackup.
13:54 egonw ok, that's bad
13:54 egonw s9asad: not so well done
13:54 s9asad :-(
13:54 egonw no worries... :)
13:54 egonw this will fix that then:
13:54 egonw git checkout master
13:55 egonw then give me the output of: git branch
13:55 s9asad Your branch is behind 'origin/master' by 7 commits, and can be fast-forwarded.
13:55 s9asad 1-m-fixingMissingPatches
13:55 s9asad 2-m-backupOfSevenOriginalPatches
13:55 s9asad 3-m-secondBackup
13:55 s9asad * master
13:55 egonw ok, we'll fix that too
13:55 egonw ok, first we'll fix origin/master
13:55 egonw git push -f origin master
13:55 s9asad sure
13:56 egonw the '-f' will force an overwrite...
13:56 s9asad Total 0 (delta 0), reused 0 (delta 0)
13:56 s9asad To git@github.com:asad/cdk.git
13:56 s9asad + 50a070d...0f64e4d master -> master (forced update)
13:56 egonw resetting the origin/master to your local master hash
13:56 egonw yes, perfect
13:56 egonw ok, now we fix the 3-m-backup
13:56 s9asad ok
13:57 egonw git branch rename 3-m-secondBackup 4-temp
13:57 egonw git branch rename 1-m-fixingMissingPatches 3-m-secondBackup
13:57 egonw git branch rename 4-temp 1-m-fixingMissingPatches
13:57 s9asad I have this error
13:57 s9asad git branch rename 3-m-secondBackup 4-temp
13:57 s9asad usage: git branch [options] [-r | -a] [--merged | --no-merged]
13:57 s9asad or: git branch [options] [-l] [-f] <branchname> [<start-point>]
13:57 s9asad or: git branch [options] [-r] (-d | -D) <branchname>
13:57 s9asad or: git branch [options] (-m | -M) [<oldbranch>] <newbranch>
13:58 s9asad Generic options
13:58 s9asad -v, --verbose         be verbose
13:58 s9asad -t, --track
13:58 egonw OK... my bad
13:58 egonw do this instead then:
13:58 egonw git branch -m 3-m-secondBackup 4-temp
13:58 s9asad k
13:58 egonw git branch -m 1-m-fixingMissingPatches 3-m-secondBackup
13:58 egonw git branch -m 4-temp 1-m-fixingMissingPatches
13:59 s9asad done
13:59 egonw that's a classic switch(1-m-fixingMissingPatches,3-m-secondBackup)
13:59 egonw so
13:59 egonw now the 3-m-secondBackup is the actual backup
13:59 s9asad okay
13:59 egonw and the latest version is again 1-m-fixingMissingPatches
13:59 egonw ok, make a third backup:
13:59 s9asad aha
13:59 egonw git checkout -b 4-m-thirdBackup 1-m-fixingMissingPatches
14:00 egonw and do not forget this one just after that:
14:00 egonw git checkout 1-m-fixingMissingPatches
14:00 egonw :)
14:00 s9asad :-)
14:00 s9asad so where are we now
14:00 egonw ok, I think you now need to apply these new skills to fix the original 0005 patch
14:01 egonw org,junit.Assert instead of junit.framework.Assert (or so)
14:01 egonw and reword the commit message of the old patch 0006
14:01 egonw and finally, remove those empty method in patch 0007
14:02 egonw I recommend making a backup after each successful fix
14:02 s9asad git rebase -i HEAD~6?
14:03 egonw yes, that's fine
14:03 s9asad so that should be my first step
14:04 egonw yes, always
14:04 egonw (well, only always when rebasing)
14:04 egonw and always check how far you really must go back
14:05 s9asad k
14:05 egonw (*never* rebase patches from the master tree)
14:05 s9asad aha
14:06 s9asad pick 411f2bf a)Global Assert defined Signed-off-by: Syed Asad Rahman <s9asad@gmail.com>
14:06 s9asad is a reword?
14:06 s9asad or you want me to edit that?
14:08 egonw edit
14:08 egonw see patch report
14:08 egonw this is the old 0005 patch: "
14:08 egonw you import the wrong Assert class... use
14:08 egonw the org.junit.Assert instead
14:08 egonw "
14:09 s9asad okay
14:13 s9asad so after edit
14:13 s9asad git commit --amend
14:13 s9asad I need to do git add -p
14:13 s9asad and then git commit --continue
14:16 s9asad sorry I mean git rebase --continue
14:16 s9asad I got this message git rebase --continue
14:16 s9asad You have uncommitted changes in your working tree. Please, commit them
14:16 s9asad first and then run 'git rebase --continue' again.
14:21 egonw meanwhile...
14:21 egonw make another back up
14:21 egonw and push the current 1-m-foo to github
14:21 egonw git push -f origin 1-m-fixingMissingPatches
14:22 s9asad did
14:22 s9asad so now I have a back up and can continue with rebase?
14:22 egonw you already rebased, I think
14:23 egonw that was at least in the output you just gave me
14:23 s9asad yes
14:23 s9asad I mean git rebase -i HEAD~6
14:23 egonw yes
14:23 egonw btw, why not merge these tow patches:
14:23 egonw a)Global Assert is fixed and assigned to org.junit.Assert
14:23 egonw correct Assert class fixed to import org.junit.Assert;
14:24 egonw into one single, cleaner patch
14:24 s9asad okay
14:24 s9asad now can we do it togethr lz
14:24 egonw use 'squash'
14:24 egonw nope
14:24 egonw I need to finish book chapters...
14:24 egonw you can do it...
14:24 egonw nothing new from here
14:24 s9asad just a q
14:25 egonw no problemo
14:25 s9asad til now what ever I have done is this
14:25 s9asad on my master?
14:25 egonw no, you have done nothing on your master
14:25 s9asad or how to make a back up so that I can't loose it
14:25 egonw all was done on the 1-m-foo branch
14:26 s9asad ic
14:26 egonw push this 1-m-foo branch to github now and then
14:26 egonw as you just did
14:26 egonw and you have the other backup branches
14:27 egonw and if you like, you can even do 'git format-patch -6' now and then...
14:27 egonw or -7
14:27 egonw better:
14:27 egonw git format-patch master
14:27 egonw that's makes patches for all patches since 'master'
14:27 egonw so, all in your 1-m-foo branch
14:27 egonw and then moves those 000*.patch onto a USB stick or so...
14:28 egonw burn them on CD/DVD
14:28 egonw something like that
14:28 egonw email them to your own gmail account is a good one too :)
14:28 s9asad :-) email is better
14:31 s9asad how to finish with6 and 7
14:31 s9asad then go for merger
14:32 egonw for 0006 you replace 'pick' by 'reword' and fix the commit message
14:33 egonw for 0007 you use 'edit', remove the empty methods, git add -u, git commit --amend, git rebase --continue
14:33 s9asad ok
14:33 s9asad thanks
14:33 egonw for the merge of those two 'assert' patches, you use 'squash'
14:52 s9asad done
14:52 s9asad now to submit the changes?
15:11 s9asad left #cdk
15:37 CIA-50 cdk: Syed Asad Rahman master * r911af3f / src/test/org/openscience/cdk/smsd/​algorithm/vflib/VFMapperTest.java : a)removed blank/unused methods and fixed imports Signed-off-by: Syed Asad Rahman <s9asad@gmail.com> (+6 more commits...) - http://bit.ly/9CeM1B
15:44 sneumann left #cdk
15:51 s9asad joined #cdk
16:04 s9asad left #cdk
16:13 slyrus wow. lots of scrollback to read this morning
16:13 egonw bbl
16:13 egonw yeah, good git tutorial session
16:13 slyrus MORE GIT
16:14 slyrus I suppose git branches and pull requests on github instead of sf.net patchsets would be too much to ask :)
17:17 CIA-50 cdk: conni75 * r15574 /jchempaint/trunk/src/main/​org/openscience/jchempaint/ (11 files in 5 dirs): Added a chain drawing tool (#40)
17:17 Gpox left #cdk
17:27 slyrus left #cdk
17:30 slyrus joined #cdk
17:50 CIA-50 cdk: Saravanaraj cdk-1.2.x * r6d9b3d2 / src/test/org/openscience/cdk/s​miles/SmilesGeneratorTest.java :
17:50 CIA-50 cdk: Unit test for bug #3040273.
17:50 CIA-50 cdk: This unit test works and shows the error on the SMILES Generator code.
17:50 CIA-50 cdk: Signed-off-by: Egon Willighagen <egonw@users.sourceforge.net> - http://bit.ly/aKOae4
17:55 CIA-50 cdk: Saravanaraj cdk-1.2.x * rb9b2272 / src/main/org/openscience/cdk​/smiles/SmilesGenerator.java :
17:55 CIA-50 cdk: Fix for branching bracket issue when generating SMILES for BrC1C(Br)C(Br)C(Br)C(Br)C1Br
17:55 CIA-50 cdk: This fix check for the position of the new chain with respect to its
17:55 CIA-50 cdk: parent. It also checks the position of the parent to see if it is at the begining
17:55 CIA-50 cdk: of the SMILES string.
17:55 CIA-50 cdk: Signed-off-by: Egon Willighagen <egonw@users.sourceforge.net> - http://bit.ly/9kvRUU
17:55 CIA-50 cdk: Egon Willighagen cdk-1.2.x * r223fc9a / src/test/org/openscience/cdk/s​miles/SmilesGeneratorTest.java : Removed output to STDOUT - http://bit.ly/98BNG1
17:56 sneumann joined #cdk
18:11 sneumann left #cdk
18:16 sneumann joined #cdk
19:46 sneumann_ joined #cdk
19:46 sneumann left #cdk
19:58 egonw left #cdk
20:50 jbrefort left #cdk
20:52 sneumann_ left #cdk
21:07 egonw joined #cdk
21:13 egonw left #cdk
21:53 bag_ joined #cdk
23:03 bag_ left #cdk

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