Use rb_wait_for_single_fd instead of deprecated rb_thread_select #206
Conversation
|
It looks like we're failing on the travis-ci tests for 2.1. Do you have any idea what's going on there? |
|
That's a weird one - it didn't look to me as if the extension was even compiled on travis. Will investigate further |
|
Couldn't reproduce on my home machine but could reproduce it in a linux vm with that particular test seed. it seems like assert_abort is somehow failing to catch the call to abort in test_metrics. Weirdly it doesn't happen if you run those tests on their own. |
…ctory in invalid state
|
That was an order dependant spec failure - tests are now running in a random order. There may well be more of these to flash out |
|
I don't write C, but I tested this out on OSX (ruby |
|
wow I just found this bugs happens and hope you test and merge soon 👍 |
|
Has there been any consideration given to this patch? 2.2.0 is now considered stable and this popular gem is completely dead in the water without it. Thoughts? Concerns? |
|
This looks good to me, but I'm not sure who is in-charge of god releases these days. @eric did you do the last gem release? |
|
I believe I did. I can release another one if it needs to be. |
|
I guess we should add ruby 2.2 to the travis-ci tests. |
Use rb_wait_for_single_fd instead of deprecated rb_thread_select
|
Okay. It's been released: |
|
You guys rock. Seriously. Leaving a comment, going grocery shopping for the week, and coming home to a fix and a push. Huge thanks all around. |
rb_thread_select has been removed in ruby 2.2 (see #194 for example). The old code will still be used if
rb_wait_for_single_fdis not available (which I believe is since 1.9.3)I've done a quick check that everything still works as intended (linux & os x) when using events, just using the
simple.rbexample in the god documentation.There doesn't seem to be any test coverage for this though, so extra pairs of eyes etc. appreciated