Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2016-01-25

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

All times shown according to UTC.

Time Nick Message
00:05 smidge joined #darcs
00:28 pavonia Okay, that works. Thank you!
01:59 pavonia left #darcs
02:50 Riastradh joined #darcs
03:36 mizu_no_oto joined #darcs
03:42 Big_G joined #darcs
04:51 mizu_no_oto joined #darcs
09:24 gh_ joined #darcs
09:49 gh_ hi
10:28 gal_bolle joined #darcs
11:32 gal_bolle joined #darcs
11:41 ilbot3 joined #darcs
11:41 Topic for #darcs is now Sprint in Seville, Spain 15-17th January | http://darcs.net/ | logs: http://irclog.perlgeek.de/darcs/ | darcs 2.10.0 is out http://darcs.net/Releases/2.10
12:17 mizu_no_oto joined #darcs
13:39 mizu_no_oto joined #darcs
14:06 gh_ file move handling on git import is looking good so far
14:38 Big_G joined #darcs
14:42 gh_ the code: http://bugs.darcs.net/patch1439
15:08 gh_ so, the following patches are separating us from a 2.10.3 release: http://bugs.darcs.net/patch1440 , http://bugs.darcs.net/patch1438 and http://bugs.darcs.net/patch1436 (all related to git import)
15:10 gh_ and all ported from owen's darcs-bridge
15:12 gh_ anything more we want into 2.10.3 ?
15:35 sm hi gh_, it might be already done but the changelog should be named appropriately and mentioned in extra-source-files for hackage
15:36 gh_ sm, hi, it is currently named CHANGELOG and listed there
15:36 sm good
15:37 gh_ I wonder if we should do anything related to darcsden?
15:38 sm urgh
15:38 sm a good question
15:40 sm nothing comes to mind, except darcsden needs a release
15:42 sm though, more for tidiness I think than user needs
16:35 gh_ sm, was there a release that included dan's changes from last year?
16:38 gh_ hmm importing git's git repo with *all* branches takes a loot of time
16:38 gh_ and the "optimizing inventory" part at the end is quite cpu intensive
16:39 gh_ in the new (darcs) repo I had a growing tentative_hashed_inventory file of more than 20 MBytes rewritten again and again after each imported commit
16:40 gh_ this is an example of where "not always write the tentative state on disk" would work great
16:48 gh_ anyway that last step of Optimizing inventory is still going one and is going to eat all my memory, all of this without writing anything to the disk it seems
16:53 sm gh_: no, there hasn't been a release (just the deployment to darcs hub)
17:17 gh_ 'slightlyOptimizePatchset' on an RL of 47K patches is "slightly" slow
17:19 gh_ (and no tag)
17:28 gh_ ok, killed darcs, copied tentative_hashed_inventory to hashed_inventory, currently running darcs repair
17:28 gh_ I wonder if 'slightlyOptimizePatchset' can be optimized
17:52 Heffalump gh_ how does importing multiple branches work?
17:53 gh_ Heffalump, I have no idea. looks like the final order of patches is not very chronological
17:54 gh_ I'm sharing the resulting _darcs soon (when I get it compressed and uploaded)
17:54 Heffalump I suspect it doesn't work :-)
17:54 Heffalump oh, I guess the import could rewind to the right point. Maybe.
17:59 gh_ Heffalump, well the working copy is like a mix of the working copies of distinct branches so, yeah it does not work that simply
18:06 gh_ http://bugs.darcs.net/issue2491
18:50 Riastradh joined #darcs
19:32 smidge joined #darcs
20:18 vikraman joined #darcs
20:41 Heffalump gh_: do the git tests work for you? I'm seeing a couple of failures in screened.
22:56 gh_ joined #darcs
23:01 gh_ Heffalump, all tests pass for me. do you have git installed?
23:04 gh_ even if you don't the tests that depend on it should appear as "skipped"
23:08 Heffalump gh_: I'll double-check then, thanks
23:09 Heffalump looks like git is telling me to tell it who I am, not sure if that would mess things up
23:09 Heffalump | fatal: empty ident name (for <tester>) not allowed
23:16 gh_ aah didn't think of that. it should be configurable via an environment variable inside of the testsuite's lib I hope.
23:19 gh_ hmm looks like the <tester> you have does come from the harness
23:19 Heffalump if it's only happening on my machine I'll figure out how to fix it, easier than you trying to reproduce
23:21 gh_ I nuked my ~/.gitconfig, and I can see that it uses my linux user name as author name, and indeed <tester> as e-mail address (inside of the git repos created by the testsuiteĆ 
23:21 gh_ )
23:22 gh_ I have git version 2.7.0.rc3
23:22 gh_ https://git-scm.com/book/es/v2/Git-Internals-Environment-Variables
23:22 gh_ looks like we need to set GIT_AUTHOR_NAME to be sure
23:22 Heffalump 1.9.4 here
23:23 gh_ I'm not going to do it now but I can tomorrow
23:23 gh_ unless you do it by then. I have to go, thanks for the report
23:24 gh_ bye!
23:29 Heffalump setting GIT_AUTHOR_NAME and GIT_AUTHOR_EMAIL in the obvious place in test.hs doesn't help

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