1. 17 Dec, 2019 1 commit
    • Jason Madden's avatar
      Get the spawn benchmark working again [skip ci] · 3c453c72
      Jason Madden authored
      Current results:
      
      | Benchmark             | gevent-libev-cext | gevent-libev-cffi             | gevent-libuv-cffi               |
      |-----------------------|-------------------|-------------------------------|---------------------------------|
      | gevent spawn          | 12.5 us           | 13.7 us: 1.10x slower (+10%)  | 14.2 us: 1.14x slower (+14%)    |
      | gevent sleep          | 2.09 us           | 3.16 us: 1.51x slower (+51%)  | 74.4 us: 35.64x slower (+3464%) |
      | geventpool sleep      | 3.46 us           | 6.20 us: 1.79x slower (+79%)  | 151 us: 43.58x slower (+4258%)  |
      | geventraw spawn       | 5.34 us           | 6.54 us: 1.22x slower (+22%)  | 6.38 us: 1.19x slower (+19%)    |
      | geventraw sleep       | 949 ns            | 1.66 us: 1.75x slower (+75%)  | 43.9 us: 46.25x slower (+4525%) |
      | geventpool join       | 1.57 us           | 3.43 us: 2.18x slower (+118%) | 88.2 us: 56.07x slower (+5507%) |
      | gevent spawn kwarg    | 14.2 us           | 12.4 us: 1.15x faster (-13%)  | 14.8 us: 1.04x slower (+4%)     |
      | geventraw spawn kwarg | 7.68 us           | 8.38 us: 1.09x slower (+9%)   | 8.47 us: 1.10x slower (+10%)    |
      | none spawn kwarg      | 771 ns            | 738 ns: 1.04x faster (-4%)    | 734 ns: 1.05x faster (-5%)      |
      
      Not significant (3): geventpool spawn; none spawn; geventpool spawn kwarg
      
      Refs #1493
      3c453c72
  2. 16 Dec, 2019 2 commits
  3. 12 Dec, 2019 1 commit
  4. 11 Dec, 2019 2 commits
  5. 08 Dec, 2019 4 commits
  6. 06 Dec, 2019 5 commits
  7. 05 Dec, 2019 2 commits
  8. 04 Dec, 2019 1 commit
  9. 03 Dec, 2019 2 commits
  10. 02 Dec, 2019 3 commits
  11. 18 Nov, 2019 8 commits
  12. 02 Nov, 2019 2 commits
  13. 22 Oct, 2019 5 commits
  14. 11 Sep, 2019 1 commit
    • Jason Madden's avatar
      Support TLS1.3, OpenSSL 1.1.1 (and Python 3.7.4) (#1460) · 9d27d269
      Jason Madden authored
      * Fix tests with TLS1.3
      
      There were hidden assumptions about the order in which greenlets would run.
      
      TLS1.3 changed the handshake in a way that broke those assumptions.
      
      * Attempting to fix the appveyor backdoor test timeouts.
      
      Suspect a buffering issue. Trying to fix by disabling Nagle's algorithm.
      
      Also a cleaner separation of concerns in the backdoor server.
      
      * Brute-force debugging.
      
      * Tweak the backlog and simultaneous accept settings, and drop the extra event listeners we don't need.
      
      * max_accept=1 and dropping the events fixed test__backdoor on windows.
      
      probably it was just the events. test that.
      
      * Brute force debugging for the SSL issues.
      
      * Some of the windows ssl issues may be due to fileno reuse and hence incorrectly multiplexed watchers?
      
      * Even more care with socket lifetimes.
      
      On Python 3.7, use the newer wrapping form that the stdlib uses to create the _sslobj.
      
      * More debugging.
      
      * The preferred method of unwrapping differs on 3.6 and 3.7+
      
      * Checking ordering.
      
      * Partial success, hopefully.
      
      * Lots of progress. Try again.
      
      * More Windows work.
      
      * D'oh. Test cases from 3.7.4 handle some of these issues more gracefully so we will just use them.
      
      * Tweaks.
      
      * Having a hard time getting 3.7 updated on Travis. Try to force the issue.
      
      * Yet another weird error from appveyor.
      9d27d269
  15. 08 Sep, 2019 1 commit