Commits on Mar 12, 2024

1. Use before_all block to setup requests/cache_spec data …

   This spec is looping through various collections of endpoints and
   asserting different caching properties.
   
   Previously, the `before` block here was running before every example,
   and thus creating/dropping the full set of factories in every
   example-wrapping transaction.
   
   The change here is to use `before_all` which wraps the example group in
   a big transaction and creates the factories (which, in this entire spec,
   are very much like fixture data). The individual transactions around
   each example from rspec are still in place, so for the few examples
   which do modify any of the records, that will not pollute other specs.
   
   Some before/after summary to show factory creation reduction:
   
   BEFORE:
   ```
   Finished in 23.1 seconds (files took 2.78 seconds to load)
   253 examples, 0 failures
   
   [TEST PROF INFO] Factories usage
   
    Total: 1611
    Total top-level: 1611
    Total time: 00:12.677 (out of 00:23.892)
    Total uniq factories: 6
   
      total   top-level     total time      time per call      top-level time               name
   
        506         506        2.6162s            0.0052s             2.6162s             status
        264         264        3.4650s            0.0131s             3.4650s            account
        253         253        3.5710s            0.0141s             3.5710s             invite
        253         253        0.4301s            0.0017s             0.4301s               poll
        253         253        2.4347s            0.0096s             2.4347s               user
         82          82        0.1603s            0.0020s             0.1603s accessible_access_token
   ```
   
   AFTER:
   ```
   Finished in 9.54 seconds (files took 2.72 seconds to load)
   253 examples, 0 failures
   
   [TEST PROF INFO] Factories usage
   
    Total: 18
    Total top-level: 18
    Total time: 00:00.860 (out of 00:10.325)
    Total uniq factories: 6
   
      total   top-level     total time      time per call      top-level time               name
   
         12          12        0.6972s            0.0581s             0.6972s            account
          2           2        0.0997s            0.0499s             0.0997s             status
          1           1        0.0093s            0.0093s             0.0093s               user
          1           1        0.0232s            0.0232s             0.0232s             invite
          1           1        0.0134s            0.0134s             0.0134s               poll
          1           1        0.0175s            0.0175s             0.0175s accessible_access_token
   ```
   
   So, we are able to reduce:
   
   - Total factory creation from ~1600 DB records to just under 20
   - Percentage of time spent on factories from ~50% to under 10%
   - Overall spec file run time by more than 50%

   

   mjankowski committed Mar 12, 2024
   Configuration menu

   • View commit details
   • Copy full SHA for 4c90de6
   • Browse repository at this point

   Copy the full SHA

   4c90de6 View commit details

   Browse the repository at this point in the history