Camelia, the Perl 6 bug

IRC log for #cdk, 2012-05-06

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

All times shown according to UTC.

Time Nick Message
06:09 egonw joined #cdk
06:29 jbrefort joined #cdk
06:46 jbrefort joined #cdk
06:47 jbrefort left #cdk
06:56 egonw joined #cdk
08:48 danielszisz joined #cdk
08:50 egonw yes, I am here :)
08:50 egonw I have currently three patches left
08:50 egonw for the MMFF test, and one for some additional comment
08:50 egonw mom, let me push it...
08:50 danielszisz ok
08:52 danielszisz by the way i reported that bug about C-N .prm problem.
08:52 egonw I particularly like to suggest using 'git rebase'
08:52 egonw I will push those three patches I have left...
08:52 egonw oh, one important thing:
08:52 egonw please develop in branches, not in master
08:52 danielszisz so cdk-1.4.x
08:52 egonw because some patches will go in earlier than other
08:53 egonw no, not that one either :)
08:53 egonw no, local, custom development branches
08:53 egonw like all those xxx-ff-someDescription on my GitHub
08:53 egonw I use that particular scheme...
08:53 egonw xxx is a counter
08:53 egonw (so reflecting some time aspect)
08:54 egonw ff is the branch I work against
08:54 egonw e.g. 397-14x-patchesFromYap
08:54 egonw or 398-m-danielMaster
08:54 egonw that way, my 'master' and 'cdk-1.4.x' are always exactly the same as upstream
08:55 danielszisz i see. so if i make a patch which branch should i assign to?
08:56 egonw depends from which you forked
08:56 egonw so normally I do on the command line:
08:56 egonw git checkout -b 399-m-somePatch master
08:56 egonw or
08:56 egonw git checkout -b 399-14x-somePatch cdk-1.4.x
08:57 danielszisz that creates a custom "work" branch you can make the changes to
08:57 egonw exactly
08:57 egonw and those you share, and others can pull from
08:58 egonw and when the upstream cdk branch updates, you can simply do:
08:58 egonw git fetch origin
08:58 egonw git checkout master
08:58 egonw git rebase origin/master
08:58 egonw git checkout cdk-1.4.x
08:58 egonw git rebase origin/cdk-1.4.x
08:58 egonw ...
08:58 egonw and because you now do not have private patches there, these will always be an exact copy
09:01 danielszisz so i make commits to the forked working dir. then push it to github and from there you can pull my changes . am i right?
09:02 egonw perfectly
09:02 danielszisz right. ok.
09:02 egonw now, the other thing I like to recommend...
09:02 egonw is 'git rebase'...
09:03 egonw it is very well explained in this (free) book: http://newartisans.com/2008​/04/git-from-the-bottom-up/
09:03 danielszisz ok
09:04 egonw this makes it easy to merge patches...
09:04 egonw so, consider this branch with three of your patches:
09:04 egonw (btw, here's an example, though this has too many branches :) https://github.com/yapchunwei/cdk/branches
09:04 egonw so, consider this branch with three of your patches:
09:05 egonw https://github.com/egonw/cdk​/commits/398-m-danielMaster
09:05 egonw three patches
09:05 egonw the first is separate, a simple // comment added
09:05 egonw the second and third can be merged, for example
09:05 egonw because the second (Header and mod...) is a fix for the third (New MMFF...)
09:05 egonw that you could do with:
09:05 egonw git rebase -i HEAD~3
09:06 egonw (see that free book)
09:06 egonw -i is for interactive
09:06 egonw HEAD~3 means the last three patches
09:06 egonw this is also explained in this post:
09:06 egonw http://gitready.com/advanced/2009/02/​10/squashing-commits-with-rebase.html
09:06 zarah egonw's link is also http://tinyurl.com/cmqj32
09:06 egonw this merging of patches is called 'squashin'
09:07 egonw that allows you to make the patches as clean as possible...
09:07 egonw it's the editing after the peer review, so to say :)
09:07 egonw making the whole process go a lot smoother...
09:07 egonw making it much easier for us to pull in your work
09:08 danielszisz i see
09:08 egonw git rebase you can also use to rebase your patch branch on the latest master
09:08 egonw e.g.
09:08 egonw git checkout -b 1-m-moreMMFFFixes master
09:08 egonw # do stuff
09:08 egonw git checkout master
09:08 egonw git fetch
09:08 egonw git rebase origin/master
09:08 egonw / which pulled in 5 new patches
09:09 egonw / then you can rebase your patches on the latest master with:
09:09 egonw git checkout 1-m-moreMMFFFixes
09:09 egonw git rebase master
09:09 egonw # which will go fine unless your patches conflict with the 5 new patches
09:09 egonw this is very well explained in that book
09:09 danielszisz ok
09:09 egonw the reason I say this...
09:09 danielszisz i will print it and read it
09:09 egonw I encounterd such conflicts for your patches :)
09:10 egonw but could resolve them myself
09:10 egonw but often the original author is the better person to do that, which is you in this case
09:11 egonw ok, I got to go away-from-keyboard for a while, to keep the kids happy :)
09:11 danielszisz ok
09:11 egonw if you have questions, please post them when I'm back
09:11 danielszisz ok, thanks
09:11 egonw one last thing...
09:11 egonw about those two MMFF...Test patches...
09:11 egonw the Copyright (C) line should list the actual authors...
09:12 egonw which is you, not Chris and I, I guess?
09:12 danielszisz yes
09:12 danielszisz ok
09:12 egonw and also only they years when the code was written...
09:13 egonw ok, ttl!
09:13 egonw and, many, many thanx for working on the builder!
09:13 egonw that is very, very much appreciated!
10:19 danielszisz joined #cdk
10:40 danielszisz joined #cdk
10:45 egonw joined #cdk
11:07 danielszisz hello Egon
11:07 danielszisz are you there?
11:08 danielszisz can i have some questions for you?
11:09 egonw yes, I'm here
11:09 egonw sure, please go ahead
11:10 danielszisz i made a new branch on github 300-master-modelbuilder3dPatches that i will use for the modelbuilder3d debugging.
11:11 egonw btw, in case you want to in the future, you can delete branches on github like this (assuming github=origin)
11:11 egonw git push origin :gh_pages
11:11 egonw git push origin :dev-1.4.x-padel
11:12 egonw is that 300 branch rebased on master?
11:13 danielszisz it says cannot rebase you have unstaged changes on master
11:13 danielszisz how can i discard the changes on master ?
11:13 egonw if you do 'git status' then you see uncommited changes
11:13 danielszisz yes did it
11:13 egonw git reset --hard HEAD~1
11:13 egonw that will erase them
11:14 egonw to make sure there are no important changes :)
11:14 egonw oh, another comment...
11:14 egonw if you add a new test class for a module 'x', you should add that to the test suite class too, MxTests ...
11:14 egonw in this case:
11:14 egonw Mbuilder3dTests.java ro so
11:14 egonw or so
11:15 danielszisz ok
11:16 danielszisz so if i reset the changes in master will it keep them on the new working branch, bacause i dont want to lose them
11:16 danielszisz ?
11:18 egonw yes, you can simply do that by making a branch first...
11:18 egonw e.g.:
11:18 egonw git checkout -b 1-savedPatchesFromOldMaster master
11:18 egonw basically, any branch or tag is just a label for a particular commit
11:19 egonw each commit is the start of a history
11:19 egonw that is, each commit has a parent commit
11:19 egonw (directed graph)
11:19 egonw with the first commit ever as root
11:19 egonw so, by tracking the last commit at any stage, will preserve the full history
12:13 danielszisz joined #cdk
12:19 Conrad joined #cdk
14:18 jbrefort joined #cdk
14:35 egonw joined #cdk
14:54 Conrad joined #cdk
14:59 Conrad joined #cdk
15:17 jbrefort joined #cdk
17:57 egonw joined #cdk
18:06 sneumann joined #cdk
18:40 Conrad joined #cdk
18:45 Conrad joined #cdk
19:19 jbrefort joined #cdk
20:47 jbrefort joined #cdk

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