recapi 0.1.9 → 0.1.10

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 28f52d6426fd83d0b14d106535a06fca9e6d29ac56b4a3b366ae31f8ca7e22df
-  data.tar.gz: 93361e6a7d6dcbbc86be19b92ccc7fb4bc3b3695f99fa701ad08cfea17ccb57f
+  metadata.gz: e9d8c7e28e6929bfef64625ac5ba48ba5bec683d27b3d4c91646b67f189cadeb
+  data.tar.gz: b17b8ad7945fac43147ce43aca7cd27a1b62105a155fdc5ce15d8eccc7e6fca9
 SHA512:
-  metadata.gz: 54a7bf72a2b956ed847fe57af69c3241af18c638a199a86aa3220b2f0726d495402a3faa4698e011f815f57881e5969090bf9047b28b60bc0b6d11767f88eacf
-  data.tar.gz: a7825355e8be66451b687bc342a8c9961f8b7f8bc4c04fb7c57a38442ab59ed12daf67be7b04a39cba4b7a76843c5faec071e35dbcc15a5b3a1aba6f5154ae4b
+  metadata.gz: 01a65170536dc4f5f33596f509fc7102670ba72a566954b1443d7d4dc29f01137206c4a3b3c49822c4bf85bc6b3200751655a49480a2afa46fe0e10ba3fa45e5
+  data.tar.gz: d6d63d879a0eed194f410983abef9770b3296606184cf822a4824c56c2a0dd5a6350922595af45257c9ddf3bb3905be35de694ba9f230a9cc2d5d7a7b6676e7e

data/README.md

@@ -30,7 +30,7 @@ line with your address): save it as `lib/recapi.rb` and
 `require Rails.root.join("lib/recapi").to_s` instead.
 
 ```sh
-curl -fLo lib/recapi.rb https://api.<your-domain>/sdk/ruby/recapi-0.1.9.rb
+curl -fLo lib/recapi.rb https://api.<your-domain>/sdk/ruby/recapi-0.1.10.rb
 ```
 
 ## Quick start
@@ -74,7 +74,7 @@ and a log line.
 |---|---|
 | `track(item_id, level, user_id:, anon_id:, recommendation_id:)` | Records an interaction: `level` -10 to 10 or `:view` (3), `:cart` (5), `:purchase` (10), `:reject` (-10: never recommend the item to this user again until a positive event on it). The history changes at once; a background thread sends the event within 2 s (at once when a batch of 1,000 is full). An `item_id` the API would refuse (1-256 bytes of UTF-8, no control character; also `nil`, an Array or a Hash) drops the event with a log line: nothing raises, and the other events are sent as usual. A bad `level`, `user_id`, `anon_id` or `recommendation_id` still raises `ArgumentError`. |
 | `flush` | Sends the queued events now, with retries (end of a job). A network error, 429 or 5xx is retried 3 times (0.5 s, 1 s, 2 s with jitter, or the `Retry-After`, at most 8 s), then the events stay queued and the sender backs off up to a minute; any other error drops them (logged). |
-| `close` | Stops the sender and sends what is queued, one try per batch. It runs by itself at exit. |
+| `close` | Stops the sender and sends what is queued, one try per batch. It runs by itself at exit, also for a classic Sinatra app (`ruby app.rb`), whose server runs from its own exit hook. Don't `track` while such an app loads, before its server starts: the client then closes before the first request, and every later `track` is sent on the page's thread. |
 | `recommend(user_id:, anon_id:, count:, placement:, tags_any:, tags_all:, exclude:, context_item_id:, discovery:)` | A `Recapi::Recommendations` (`recommendation_id`, `fallback?`, `items`). At most 1.5 s; on a network error, timeout, 5xx or 429 it returns popular items (`fallback?`, no id), or no item when those fail too: hide the slot; on a 401 or 403 no item at once. A `context_item_id` the API would refuse (also an Array or a Hash, as `params[:id]` gives for `?id[]=a`) isn't sent: you get the same popular list and a log line, with no recommendation request. Any other failure (a bad request, a reply that isn't the API's) raises. |
 | `identify(anon_id, user_id)` | At login: the visitor's history and shown items become the user's. The user's first restore from recapi then leaves that visitor's browsing out, so it doesn't count twice. |
 | `popular(count:, tags_any:, tags_all:)` | Popular items, for pages without a user. On a network error, timeout, 5xx, 429, 401 or 403: an empty list and a log line. Other failures as in `recommend`. |
@@ -122,6 +122,15 @@ and logs one line at start:
 so a visitor who never returns stays on disk (about 8 KB each): run
 `Rails.cache.cleanup` on a schedule, or leave `store:` out.
 
+## Changes
+
+- 2026-11-05: version 0.1.10. `track` stays queued and in the background in a
+  classic Sinatra app
+  (`ruby app.rb`). The exit hook used to close the client before the server
+  started, so every event was sent on the page's thread, and a hung API
+  stalled each page for 5 s. The hook now starts with the sender, at the first
+  `track`.
+
 ## Development
 
 `../../scripts/ruby.sh -w test/run.rb [filter]` runs the tests (`make

data/lib/recapi.rb

@@ -25,7 +25,7 @@ rescue LoadError
 end
 
 module Recapi
-  VERSION = "0.1.9"
+  VERSION = "0.1.10"
 
   # Decayed user history, exactly as docs/history-spec.md (times in unix ms).
   # A port of sdk/python's history.py, snapshot merging included.
@@ -433,8 +433,8 @@ module Recapi
     TIMEOUT_ERRORS = [Net::OpenTimeout, Net::ReadTimeout, Net::WriteTimeout].freeze
     # A connection the server closed while idle fails this way, before any reply.
     STALE_ERRORS = [EOFError, Errno::ECONNRESET, Errno::EPIPE, Errno::ECONNABORTED].freeze
-    # Made here, not in initialize, whose block would hold the client through
-    # self: the weak reference lets a dropped client be collected.
+    # Made here, not in enqueue, whose block would hold the client through self:
+    # the weak reference lets a dropped client be collected.
     CLOSE_AT_EXIT = lambda do |ref|
       at_exit do
         ref.__getobj__.close
@@ -480,7 +480,6 @@ module Recapi
       @sleeper = method(:wait)
       @random = -> { rand }
       init_threading
-      CLOSE_AT_EXIT.call(WeakRef.new(self))
     end
 
     # Records an interaction: the user's history changes at once, the event
@@ -817,6 +816,12 @@ module Recapi
           @thread = Thread.new { run }
           @thread.name = "recapi-sender"
           @thread.report_on_exception = false
+          # Here, not in initialize: Sinatra's classic mode serves from its own
+          # at_exit, and exit hooks run last registered first, so a hook from
+          # the app's load would close the client before the first request.
+          # One registered while that server runs goes after the server ends.
+          # A forked worker starts its own sender, so it gets its own hook too.
+          CLOSE_AT_EXIT.call(WeakRef.new(self))
         end
         @queue << event
         drop_overflow if @queue.size > MAX_QUEUE

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: recapi
 version: !ruby/object:Gem::Version
-  version: 0.1.9
+  version: 0.1.10
 platform: ruby
 authors:
 - recapi