Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2017-12-16

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

All times shown according to UTC.

Time Nick Message
00:57 idiosyncrat Demat!
00:59 idiosyncrat amon has suggested adding Travis Cl config files to the Marpa::R3 repo.  What do others think of that?
01:00 idiosyncrat amon: Am I correct in understanding that adding Travis Cl compatibility is just a matter of having the appropriate config file, and beyond that has no impact?
01:07 koom joined #marpa
02:58 ilbot3 joined #marpa
02:58 Topic for #marpa is now Start here: http://savage.net.au/Marpa.html - Code paste/run: https://f.perlbot.pl/#marpa - Jeffrey's Marpa site: http://jeffreykegler.github.io/Marpa-web-site/ - IRC log: http://irclog.perlgeek.de/marpa/today - Youtube channel: https://www.youtube.com/channel/UCYKVfGBtfTqbs1JdYq-dc5g
03:25 ernimril joined #marpa
10:00 amon_ joined #marpa
10:05 amon_ Re Travis: This involves the .travis.yml file to describe how the project should be built, and signing up on the travis-ci.org site. After connecting Travis with your Github account, you can activate automated builds.
10:06 amon_ Then, whenever a commit is pushed to Github, a build will be launched. This has the following benefits:
10:07 amon_ 1. the build is performed in a clean environment, which can discover missing dependencies or problems in the build process
10:07 amon_ 2. a number of configurations (perl versions, compilers) can be tested. This is significantly faster but also far less comprehensive that CPAN Testers
10:09 amon_ (Travis only supports Linux containers on x86_64 and macOS, nothing exotic like Windows or Sparc)
10:10 amon_ After a build, the committer is sent an email with the results. This can be annoying, but I find the confidence useful.
10:12 amon_ Also, Travis integrates at some points with the Github UI
10:12 amon_ This is most useful on larger projects where pull requests will be automatically built by Travis. If the build fails, that should be fixed before merging.
10:14 amon_ Marpa::R3 is probably not yet at a stage in its lifecycle where that would be necessary, but it might still be useful.
10:19 amon_ The most involved Travis build I've done was for my Async::Trampoline module. This was complicated by using Dist::Zilla, XS, and C++11. So I wanted to test different compilers and Perl variants – more than I want to install locally.
10:20 amon_ Here's the (unusually complicated) .travis.yml: https://github.com/latk/p5-Async-Trampoline/blob/master/.travis.yml
10:21 amon_ And here are the build results: https://travis-ci.org/latk/p5-Async-Trampoline (click on one job to see the build log)
16:40 koom joined #marpa
16:49 idiosyncrat joined #marpa
16:49 idiosyncrat amon_: re https://irclog.perlgeek.de/marpa/2017-12-16#i_15587056 -- Sounds nice.
16:51 idiosyncrat What if there are commits for which I do not want a Travis build?  That is, I sometimes commit versions that are interim versions, and which I know are broken.  I'd just as soon *not* see the build results for them.
16:56 amon_ Travis builds are triggered per push, not per commit. So as long as the HEAD commit is good, everything is fine. You can also suppress the trigger by including "[skip ci]" in the commit message.
17:01 idiosyncrat "[skip ci]" is what I was looking for.  Great!
17:02 idiosyncrat If you're willing to volunteer to do a travis.yaml file, I'm happy to try it out.
17:02 idiosyncrat I actually develop in continuous intergration style already.
17:04 idiosyncrat The "t++" and "t-3" notations at the end of my commit messages record the latest test results.  "t++" means all tests OK.  "t-3" means all but 3 tests fail.
17:05 idiosyncrat amon_: In the process, you may discover some bugs&misfeatures in my build system.
17:07 amon_ ok, I'll activate Travis on my fork and send a PR when it works
18:04 amon_ Done :)
18:04 amon_ I considered using the "make cpan_dir_full_test" target for Travis, but that depends on a "realclean" target which doesn't exist.
18:59 koom joined #marpa
19:02 Cheery joined #marpa
19:09 amon_ In other news, I wrote a small "imarpa" script as a simple command line driver for Marpa R2/R3: https://github.com/latk/p5-MarpaX-App-imarpa
19:09 amon_ I'd be interested in any feedback before I upload to CPAN.
19:09 amon_ Specifically, can anyone suggest a better name?
19:09 amon_ I chose "imarpa" in analogy to irb or ghci, by this isn't really interactive at all. I don't want to hog the "marpa" name.
21:13 koom joined #marpa
21:36 amon_ left #marpa
22:01 ronsavage joined #marpa
23:11 idiosyncrat joined #marpa
23:12 idiosyncrat amon_: Thanks for your consideration when it comes to namespace.
23:12 idiosyncrat I hereby delegate the imarpa script name to your project. :-)
23:13 idiosyncrat I haven't looked at it carefully, but I have confidence you're making good use of it based on your past work.
23:13 idiosyncrat If I do have suggests, I hope they'll receive due consideration. :-)
23:16 idiosyncrat _amon: A question re "[ci skip]" --
23:16 idiosyncrat Builds you said are per "git push" ...
23:19 idiosyncrat but the '[ci skip]' tag is in a commit message.  Github says that such a commit is "ignored by Travis CI".  Is that forever in the sense that Travis Cl will cherry-pick commits and leave the ones tagged "[ci skip]" out?
23:20 idiosyncrat Or is the only relevant tag the one on the most recent commit?
23:49 idiosyncrat I am currently trying to set up for Travis, and have a "wait for a few seconds" screen that's been up for several minutes now.
23:52 idiosyncrat The message is "Just a few more seconds as we talk to GitHub to find out which repositories belong to you." and it got a GIF(?) with a rotating face of (I assume) Travis himself.

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