[Tile-serving] [openstreetmap/osm2pgsql] Avoid parallel execution for pending items when there are only a few of them (PR #1827)

[Sarah Hoffmann]

Worker handling for parallel processing of pending items has quite a large overhead of around 1s. This is not much of a problem when handling larger update files. It is noticable in the tests where we process update batches with one or two items only.

This avoids parallel processing entirely when only a few items are involved. I've set an completely arbitrary limit of 100 items. 
You can view, comment on, or merge this pull request online at:

  https://github.com/openstreetmap/osm2pgsql/pull/1827

-- Commit Summary --

  * do not use workers on small lists of pending objects
  * improve log output when not using threads on pending processing
  * add a comment why the sequential processing

-- File Changes --

    M src/osmdata.cpp (62)

-- Patch Links --

https://github.com/openstreetmap/osm2pgsql/pull/1827.patch
https://github.com/openstreetmap/osm2pgsql/pull/1827.diff

-- 
Reply to this email directly or view it on GitHub:
https://github.com/openstreetmap/osm2pgsql/pull/1827
You are receiving this because you are subscribed to this thread.

Message ID: <openstreetmap/osm2pgsql/pull/1827 at github.com>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tile-serving/attachments/20221118/23a46679/attachment.htm>