Skip to content
Snippets Groups Projects

Repository graph

You can move around the graph by using the arrow keys.
Select Git revision
  • 127-eliminate-more-references-to-hive_posts_view
  • 221-vacuum-hivemind-tables
  • 267-rebased-onto-develop
  • 267-update-notification-cache
  • 676-as-tiny-assets
  • 72-postgres-warning-about-wrong-collation-in-recursive_deps-2
  • abw_ecv_op_experiment
  • abw_max_retries
  • abw_post_delete_fix
  • abw_rshares_experiment
  • add-git-to-requirements
  • add-vote-info
  • arc-get-content-deleted
  • as-tmp-to-remove
  • asuch-limit-follows
  • asuch-postgrest-fixing-problems
  • asuch-replace-python-with-sql-get-follow-count
  • asuch-tmp-wip-condenser-get-blog
  • autoexplain-python
  • bridge_api.get_ranked_posts_fixes
  • 1.27.11rc2 protected
  • 1.27.11rc1 protected
  • 1.27.10 protected
  • 1.27.9 protected
  • 1.27.8 protected
  • 1.27.7rc16 protected
  • 1.27.7rc15 protected
  • 1.27.7rc14 protected
  • 1.27.7rc13 protected
  • 1.27.7rc12 protected
  • 1.27.7rc11 protected
  • 1.27.7rc10 protected
  • 1.27.6rc9 protected
  • 1.27.6rc8 protected
  • 1.27.6rc7 protected
  • 1.27.6rc6 protected
  • 1.27.6rc5 protected
  • 1.27.6rc3 protected
  • 1.27.5 protected
  • 1.27.5rc9 protected
40 results
Created with Raphaël 2.2.05Jan432129Dec282725232221181716151411109875432130Nov2827262524232221201918171614131213121110987654Hardcode path to test id - user has no need to set it manuallyTavern test dir as parameter, renamed original python benchmarks script.[DLN] validate author name using find_account_id, then use id instead of author name in query[DLN] not sure it makes sense to rank promoted comments as well as posts, but that's old behavior and probably no one cares anyways[DLN] fix typo[DLN] added new live_post views, hopefully fixing problem where counter_deleted wasn't specified in some some of the new CTEsResolved further conflicts accidently missed previously.branch 'release_candidate_v1_24' rebased onto masterRebase to current developSet python3 as default interpreter for tox env. Fixes issue when there are more than one interpreter installed.HIVEMIND_BENCHMARKS_IDS_FILE added to tox.iniSet tests ids file from env variable[DLN] eliminate duplicate language specification[DLN] slightly older version of bridge_get_discussion is working better, don't feel like anlayzing this late, so just going to revert it[DLN] minor code cleanup, should be no change[DLN] fix typo and minor code cleanup[DLN] fix invalid rename of source to something more logical that breaks the interface + more code cleanup[DLN] fix another typo[DLN] fix compile error in bridge_get_ranked_post_by_payout_comments[DLN] created some more CTEs assoicated with get_post_view_by_id and various minor code cleanups[DLN] add a limit to how many replies to fetch, and more importantly, don't return replies that are deleted.[DLN] fix another typo[DLN] fix another query error from code refactor[DLN] fix another typo[DLN] fix typo[DLN] fix misc compile errors[DLN] some more cleanup, still probably won't compile[DLN] Replaced more usages of hive_posts_view and added more CTEs around subqueries that join to get_hive_posts_by_id[DLN] Several of these queries could potentially have the same problem as bridge_get_ranked_post_by_created_for_tag, so decided to encapsulate their post id subqueries in CTEs as their no real chance for it to be disadvantageous to do so.[DLN] Query planner got unexpectedly stupid after replacing hive_posts_view with get_post_view_by_id (against all apparent reason, since it's job should be easier): it stopped using a bitmap index for the tags. Putting the created table into a CTE fixed the issue.[DLN] so just found out db_upgrade.sh doesn't run by default in CI, which makes sense, so add ref to new sql file in schema.py[DLN] yet another typo[DLN] fix a few more typos[DLN] fix typo (get_hive_post_by_id instead of get_post_view_by_id)[DLN] upgrades for index changes, still needs change to upgrade_runtime_migration.sql[DLN] Add get_post_view_by_id to replace hive_post_view which has too long of a planning time. All the joins in hive_post_view are essentially one-to-one (no hive_posts get filtered out by the joins) so order was never super critical within the view itself. But the filtering of hive_posts based on id is critical to performance and needs to be done before any of the joins. I tried to do this without using a function, but wasn't able to figure out how, so ultimately I went created the get_post_view_by_id function and set the join_collapse_limit locally within the function to 1 (locallly settable value was another benefit of using a function). Since functions can't take a table as input, I just pass in the id of a single post. It's possible performance could be increased by passing in an array of ids when there's a lot of ids, but I didn't bother to experiment with this since generally we never process more than about 2000 hive_post ids anyways. I also haven't experimented much with reordering of the manually set joins within get_post_view_by_id, I just chose an order than seemed logical and performance appears good enough so far. Other sql code that joins to hive_post_view should probably also be later be changed to use this function instead and probably remove hive_post_view in favor of a composite type definition, but for now I just changed the usages that were trivally easy to replace. The trivial edits haven't been compiled/tested, so likely to fail in CI.[DLN] Add get_post_view_by_id to replace hive_post_view which has too long of a planning time. All the joins in hive_post_view are essentially one-to-one (no hive_posts get filtered out by the joins) so order was never super critical within the view itself. But the filtering of hive_posts based on id is critical to performance and needs to be done before any of the joins. I tried to do this without using a function, but wasn't able to figure out how, so ultimately I went created the get_post_view_by_id function and set the join_collapse_limit locally within the function to 1 (locallly settable value was another benefit of using a function). Since functions can't take a table as input, I just pass in the id of a single post. It's possible performance could be increased by passing in an array of ids when there's a lot of ids, but I didn't bother to experiment with this since generally we never process more than about 2000 hive_post ids anyways. I also haven't experimented much with reordering of the manually set joins within get_post_view_by_id, I just chose an order than seemed logical and performance appears good enough so far. Other sql code that joins to hive_post_view should probably also be later be changed to use this function instead and probably remove hive_post_view in favor of a composite type definition, but for now I just changed the usages that were trivally easy to replace. The trivial edits haven't been compiled/tested, so likely to fail in CI.[DLN] fix compile error[DLN] we don't care about deleted replies, so just exclude counter_deleted != 0 from index to make it smaller (2 fields instead of 3)Merge branch 'dn_fix_invalid_query' into 'develop'