Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2015-02-24

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

All times shown according to UTC.

Time Nick Message
00:43 opkick [pdl] wchristian pushed 1 new commit to master: http://git.io/AXk5
00:43 opkick pdl/master 2753c35 Craig DeForest: fix a bug with empty 'ref' sections leading to error messages; support...
02:59 mohawk Mithaldu, you remember you were expressing some strong emotions about longlong-double-fix and the twisty turny merging etc?
02:59 mohawk i'm fighting that now :-)
03:43 mohawk Your branch and 'upstream/longlong-double-fix' have diverged,
03:43 mohawk and have 243 and 74 different commits each, respectively.
07:20 ribasushi joined #pdl
07:38 ribasushi joined #pdl
08:33 Mithaldu mohawk: yeah, it's not very fun
14:08 mohawk i'm getting there though
14:25 mohawk top tips for developers:
14:25 mohawk don't commit your debugging work
14:26 mohawk don't commit your disabling of chunks of code even for dev purposes because you store up trouble for down the line
14:26 mohawk don't commit your stuff when it still has merge conflicts in it because you didn't even bother to test it
14:32 jberger can I use "if(0){ ... }" to comment out large chunks of code?
14:32 jberger srsly you found committed merge conflict marks?
14:33 jberger that is grounds for removing commit bits IMO
14:39 mohawk let me find the commits
14:42 mohawk before i paste that, this recent one features a mistake of trying to remedy a build failure in the tests: https://sourceforge.net/p/pdl/code/ci/1e3c95f85f75129b7c9a3a4c4923492e0c6deaf7/
14:42 mohawk merge conflicts pushed:
14:42 mohawk https://sourceforge.net/p/pdl/code/ci/aebe0dc33a0c8ffdcce17c951b23e7591fbad956/
14:43 mohawk hm, that's from 2002 and apparently just the once
14:43 mohawk i'm sure i saw more
14:44 jberger merge conflict in the changes file? strange
14:44 mohawk https://sourceforge.net/p/pdl/code/ci/c8e4e8f00e9bcc92f11f2a5399b03e84cc427093/
14:45 mohawk this one is from failing to rebase, and failing to test
14:45 mohawk https://sourceforge.net/p/pdl/code/ci/b48059ae5189d3663f5a082c5aaf4bba621f244c/
14:47 mohawk caused by doing a merge instead of rebasing, and not bothering to resolve the conflicts: https://sourceforge.net/p/pdl/code/ci/8376d3bb02f979d7472a24d614fb1fc30390a81c/
14:47 jberger oh dear god
14:47 jberger is that stuff still in live code?
14:47 mohawk this is on LLDF
14:48 mohawk which is certainly why it got abandoned last year
14:48 mohawk it's substantively in the 60-ish commit range
14:49 mohawk i rebased it onto current master, and about halfway through there's a point where there are improvements and it passes all the tests
14:49 mohawk if/when chm merges my current MRs, i'll MR at least that part to reduce the overhang, if i haven't finished fixing up the rest
14:50 mohawk the second half features lots of committing intermediate debugging stuff
14:50 mohawk note to those following at home: this is not how software engineers do their thing
14:50 mohawk good ones self-code-review by doing "git diff" before committing
14:51 mohawk and they work on topic branches which might get rebased / reworked to get the commits correct, before merging
14:52 mohawk god knows how many times i rebased and reworked my XSMULTI changes before they were accepted into EUMM
14:52 mohawk ballpark, 20, maybe 30 or 40
14:52 mohawk and i made damn sure each commit looked good, and still passed all the tests
14:55 jberger mohawk++ # indeed
14:59 mohawk further top tip: (run on a branch) git rebase -i -x "make realclean ; perl Makefile.PL && make && make test" master
15:01 mohawk will generate a git rebase -i that steps through each of the commits on that branch, building and testing, and it will stop and leave you checked out at that commit on any that fail
15:01 mohawk obviously it will also stop at any that fail to apply cleanly
15:02 mohawk when you've got it going again, just git rebase --continue
15:02 mohawk and it will continue on its way
15:06 mohawk commit 01e4a5ce44634f1a4984d8c2c1a5409429ef10e7
15:06 mohawk Author: Chris Marshall <devel.chm.01@gmail.com>
15:06 mohawk Date:   Mon Apr 28 11:33:01 2014 -0400
15:06 mohawk Finish rough support for PDL_Anyval as badval, undefval, and values
15:06 mohawk
15:06 mohawk Most tests in the test suite pass.  Now need to track down remaining
15:06 mohawk problems test by test.
15:06 mohawk we don't commit things like that
15:06 mohawk we actually fix them, THEN commit
15:10 jberger that is a good tip and I didn't know that one
15:11 jberger it work sometimes I need to commit broken on my branch, so I can push to another machine for testing (don't ask) but then I typically rebase squash those out
15:11 mohawk as a temporary measure, that's great
15:11 mohawk but leaving that shit in there, not so much
15:12 mohawk the "fixup" (or "f") thing in git rebase -i is great too
15:15 jberger I use squash as a general term, often it is fixup
16:32 Mithaldu chm: i just discovered something weird
16:32 Mithaldu chm: apparently OpenGL.pm constants are not exported as perl constant functions
16:43 vicash mohawk: that is just a difference between svn and git users. svn users mostly commit their work-in-progress stuff to the server, git users do that locally and push to server later... git branching is much better than svn for this reason. with git, you can commit without internet access & push when you have access
16:44 sivoais mohawk: your MRs are empty now because of how SF deals with moving HEADs
18:43 opkick [pdl] wchristian pushed 24 new commits to master: http://git.io/ASMm
18:43 opkick pdl/master fcf9823 Ed J: Flatten Libtmp/Opt
18:43 opkick pdl/master 3e8d43e Ed J: Core.pm no need to be .PL
18:43 opkick pdl/master 552a32e Ed J: Add "use warnings", remove undef dims in reshape
21:43 opkick [pdl] wchristian pushed 1 new commit to master: http://git.io/AHZM
21:43 opkick pdl/master 4780cc8 Chris Marshall: Update Known_problems and Release_Notes for CPAN devel release 2.007_11
22:13 opkick [pdl] wchristian pushed 1 new commit to master: http://git.io/AHge
22:13 opkick pdl/master e93b5a9 Chris Marshall: Update VERSION to 2.007_12 for more development
22:13 opkick [pdl] wchristian tagged v2.007_11 at 0937cca: http://git.io/AHgv

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