ioquatix (Samuel Williams)
- Login: ioquatix
- Email: email@example.com
- Registered on: 07/10/2011
- Last connection: 01/18/2022
- 01:39 AM Ruby master Bug #16908: Strange behaviour of Hash#shift when used with `default_proc`.
- Thank you @jeremyevans0
- 05:52 AM Ruby master Bug #18485 (Closed): Even though init a blocking Fiber with Fiber.new(blocking: true) but scheduler is invoked
- This is working as intended.
f = Fiber.new(blocking: true) do
- 05:49 AM Ruby master Bug #18484: Fiber should return it self when blocked by IO
irb(main):001:1* f = Fiber.new do
irb(main):003:1* Fiber.yield 123
- 05:48 AM Ruby master Bug #18465: Make `IO#write` atomic.
- I would personally like to simplify `IO` implementation but I'm not sure if major refactor is acceptable especially g...
- 08:56 PM Ruby master Bug #18482 (Rejected): Fiber can not disable scheduler
- Did you read the documentation?
- 07:54 PM Ruby master Bug #18465: Make `IO#write` atomic.
- By the way, even calling `write` directly is no guarantee of synchronous output between threads/processes - on Linux ...
- 07:53 PM Ruby master Bug #18465: Make `IO#write` atomic.
- Thanks for all that information.
This is a bug fix, but what you are proposing sounds like a feature request.
- 08:19 PM Ruby master Bug #18465: Make `IO#write` atomic.
- > I think it'd better to guarantee atomicity for puts and write, even if the same fd is used by multiple IO instances...
- 08:01 AM Ruby master Bug #18465: Make `IO#write` atomic.
- Even though `stderr` should not be buffered, I feel like it would be advantageous to use a write lock too, to avoid i...
- 06:46 AM Ruby master Bug #18465: Make `IO#write` atomic.
- I tried to make a micro-benchmark measuring this.
> make benchmark ITEM=io_write
Also available in: Atom