Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2014-06-07

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

All times shown according to UTC.

Time Nick Message
00:40 haarg it's not actually his design
00:41 haarg not sure there's much point rescuing it
01:30 FROGGS_ joined #metacpan
01:35 klapperl_ joined #metacpan
03:00 oiami joined #metacpan
03:10 vanstyn_ joined #metacpan
03:22 ether rescue to put out of its misery?
03:27 haarg i don't see much benefit to breaking stable (well, stagnant) code
03:28 haarg best that could be done with the module itself is adding a deprecation warning
03:33 mst bigass deprecation warning in the *POD* would be nice
03:33 mst last project I was aware of using it was SQL::Translator which ilmari ported to Moo
03:40 haarg joined #metacpan
04:40 vanstyn joined #metacpan
05:11 oiami1 joined #metacpan
05:12 oiami joined #metacpan
06:59 neilb joined #metacpan
08:03 neilb joined #metacpan
08:40 ribasushi re the rebase wars - my workflow after the *always massive* amount of rebases is, right before things hit travis to do: ~/devel/dbic$ while true; do (perl Makefile.PL && make realclean && perl Makefile.PL)&>/dev/null && prove -lrswTj10 && git rebase --continue || break ; done
08:40 ribasushi and in the meantime I go do something else
08:44 ribasushi trs: the dbic repository btw is a good example of before and after - that is the svn/svk history was meticulously preserved (haarg++)
08:45 ribasushi trs: so look at the repo history now and around 2009 and tell me what you rather work with ;)
09:01 haarg that conversion was an attempt to maintain all relevant information while allowing people to ignore that it was ever using svn/svk
09:01 haarg hopefully it has been successful in that.  but the svk history gets rather messy.
09:04 mst I miss svk, somehwat. it was insane, but it did everything except rebase that git does for me now.
09:06 haarg i tried to use svk once.  then i moved my checkout directory.
09:06 haarg gave up at that point
09:11 mst I ran about four years of DBIx::Class merge master work using svk
09:11 mst if you knew how to treat it right, it ... worked
09:12 ribasushi the only problem with rebasing is that it hides all the sorrow and anguish of the actual work, which was always there to see with svn
09:13 ribasushi which is a net benefit - because I get to do a *lot* more of crazy shit if I know nobody will ever see it
09:13 ribasushi also I have global gc.auto=0 receive.autogc=0
09:13 ribasushi so that reflogs are eternally there to remind me of the pain ;)
09:13 haarg i got a lot of mileage out of git-svn, and i gather svk had similar capabilities in may ways.
09:15 haarg i'm just glad i never had to deal with svk directly.  seemed like it was trying to live in two different worlds.
09:15 mst haarg: yeah, git-svn didn't exist when I did that
09:15 mst svk made perfect sense to me, but I had a lot of trouble teaching it
09:15 haarg and yeah, at the time there weren't many alternatives to svk
09:15 mst I think somehow it fit my brain but not other people's, hence why I didn't tend to recommend it, even though I relied on it absolutely
09:17 haarg i bought in to svn fully, but at the time i didn't have a lot of experience.
09:17 haarg it was usable enough, but i ended up using it as a file backup mechanism
09:17 haarg where git is a useful development tool
09:18 ribasushi svn is still the only answer for distributed dev with varying levels of trust
09:18 haarg ish
09:18 ribasushi git's "move snapshots of everything around" just doesn't fit in a corp
09:18 ribasushi by distributed I didn't mean distributed scm, I meant distributed teams
09:19 haarg the git idea is different areas of concern should be different repos
09:19 haarg and it doesn't always work out that way
09:19 haarg and it also depends on if you are taking about read vs write
09:20 haarg because write permissions can be enforced as fine grained as you want with git
09:21 haarg kind of got off on my git ranting mode to people who already understand, but welp
09:25 ribasushi but write permissions to what ;)
09:25 mport joined #metacpan
09:26 haarg to specific files?  to specific regions of files?  however you want.
09:34 neilb joined #metacpan
09:52 rashi joined #metacpan
10:12 neilb joined #metacpan
10:19 neilb joined #metacpan
10:37 rashi joined #metacpan
10:48 alnewkirk joined #metacpan
11:02 oiami left #metacpan
11:38 neilb joined #metacpan
11:56 DerAlex joined #metacpan
12:18 neilb joined #metacpan
12:39 DerAlex joined #metacpan
12:49 neilb joined #metacpan
13:01 DerAlex joined #metacpan
14:14 mport joined #metacpan
14:38 oiami joined #metacpan
16:32 ether ribasushi: regarding your while command and rebasing -- why would you be getting stopped at rebases if there's no action to be taken?  a rebase generally only stops at a commit if there's a conflict, so git rebase --continue won't do anything on its own
16:33 ether ribasushi++ # "I get to do a *lot* more of crazy shit if I know nobody will ever see it"
16:33 ether it really does give a lot of freedom, especially for those with the personality type or experience level to be really really paranoid about being seen by others to make mistakes
16:33 ether for that reason I try to emphasize the advantages of rebasing to newbies
16:35 ether varying degrees of trust can be handled with subrepos, if the codebase really does need to be shared
16:35 ether so you have a subrepo with exclusive r/w perms and everyone else gets a r/o copy of that stuck in their repo
16:35 ether but yeah.. split up the code into discrete libraries with clean APIs
18:40 FROGGS joined #metacpan
19:41 rashi joined #metacpan
20:33 haarg you can do file level permission checks with hooks
23:10 jwang joined #metacpan

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