When an article is updated the associated collection is touched along with all its articles. This creates a deadlock where two transaction(nested) lock the article that was updated.
With this commit the collection wll be touched after the article updates transaction has been commited. Doing so we have two sequential transactions.
* wip
* Removed reference to cached_tag_list as the feed enpoint only returns tag_list now.
* Removed reference to cached_properties.
* Updated <CommentsCount /> and <TagList /> to support additinal className
* wip
* wip
* Undid init scrolling for Algolia articles
* Converted homePageFeed.jsx to an ERB template to access Sprockets.
* Now the homePageFeed.jsx files is an ERB so it can load Sprockets assets.
* wip
* Removed accidental commit of debugger statement in ERB.
* Now different feed time frames are loaded based on the homepage feed tab clicked.
* Prevent default Algolia paging when on the homepage.
* Added a tiny comment about feed time frame.
* Now feed loads properly for each time frame on home page.
* prop name refactor.
* Reneabled default Algolia scrolling.
* On the homepage, we no longer overwrite article HTML with Algolia data for initial articles.
* Assumption is the proper URL for main_image of featured article will be sent in the feed.
* Fixed Google Analytics for featured article.
* No longer need the new-feed ID for checking if we're on the homepage.
* Trying to make codeclimate happy.
* wip
* Removed reference to cached_tag_list as the feed enpoint only returns tag_list now.
* Removed reference to cached_properties.
* Updated <CommentsCount /> and <TagList /> to support additinal className
* wip
* wip
* Undid init scrolling for Algolia articles
* Converted homePageFeed.jsx to an ERB template to access Sprockets.
* Now the homePageFeed.jsx files is an ERB so it can load Sprockets assets.
* wip
* Removed accidental commit of debugger statement in ERB.
* Now different feed time frames are loaded based on the homepage feed tab clicked.
* Prevent default Algolia paging when on the homepage.
* Added a tiny comment about feed time frame.
* Now feed loads properly for each time frame on home page.
* prop name refactor.
* Reneabled default Algolia scrolling.
* On the homepage, we no longer overwrite article HTML with Algolia data for initial articles.
* Assumption is the proper URL for main_image of featured article will be sent in the feed.
* Fixed Google Analytics for featured article.
* No longer need the new-feed ID for checking if we're on the homepage.
* Added postcst episodes to <Feed /> state.
* Breaking up Article.jsx to make codeclimate happy.
* wip
* Created the <TodaysPodcastEpisodes /> and <PodcastEpisode /> components.
* Fixed issues with some components that were updated for the <FeaturedArticle />
* Added tests for <FeaturedArticle />
* Barreled up <FeaturedArticle />
* Added Storybook stories for <FeaturedArticle />
* Featured articles don't display org, so removed that test.
* Added proper today's podcasts components to feed.
* Converted some leftover ERB + removed pro checkmark as it doesn't exist yet.
* Now logic is same as master for feed timeframe.
* Updated <FeaturedArticle /> snapshots.
* On frontpage, bookmark button state is managed by Preact components now.
* bookmarkedFeedItems is now a prop on <Feed />.
* Fixed logic check for initializing reading list icons.
* Now <Article /> and <FeaturedArticle /> receive a bookmarkClick prop.
* Removed beginnings of scroll handler in <Feed /> as we're using what's there already for now.
* Removed logic to not show main image on other time frame feeds as it was not correct.
* Fixed issue with 0 rendering when no podcast episodes in feed.
* Converted <SaveButton /> to a stateful component.
* wip - <Feed /> render prop now passes bookmarked feed item ID set and a click event.
* Moved FEED_ICONS to an export so it's easier to use JSX/JS in <Feed />
* Moved <Feed /> into article components.
* Now reading list save buttons work properly in Preact components.
* Fixed issue with Save buttons in Algolia paging/<Feed />
* Updated snapshots.
* Took <Feed /> out of articles barrel file as jest was trying to import FEED_ICONs which is an ERB template.
* Now the Preact feed is dynamically imported.
* Now the Preact <Feed /> only loads for a logged on user.
* Fixed <FeaturedArticle /> so that the save button behaves properly now.
* Updated article stories for Storybook.
* Putting this back to the order it was because I've removed my code changes for this block.
* Fixed when renderFeed is imported for a logged on user.
* Add podcasts to feed and render home from server
* Add padding to feature laoding div
* Update tests
* Add more loading articles
* Fix conflicts
* Add readinglist javascript to initscrolling
* Ensure first article is one with image
* Organize feed items outside of feed instead of in it
Co-authored-by: Ben Halpern <bendhalpern@gmail.com>
Co-authored-by: Josh Puetz <josh@grorichpuetz.com>
* Add Transport module to Search
* Test public methods only
* Add Search::Client with exception handling
* Remove Transport and request do
* Replace SearchClient with Search::Client
* Custom social preview template for tags
* Added index on tags social_preview_template + small refactoring
* Add checks if index exists to tags migration
* Re-subscribe user to newsletter by setting MC status to pending
* Add guard clause and new spec
* Make trait name more explicit
* Fix guard clause and spec
* Use active in place of sent on internal broadcast routes
* Add and use active scope on broadcasts
* Update seeds + specs to use active broadcasts
* Add note about active broadcasts to /internal/broadcasts page
It only runs when markdown has underscored usernames.
Skip escaping underscore when content is in code or codeblock. It works by
going through all lines of a markdown content. Find underscored username.
Escape if we are not in the codeblock.
The (?<!) is negative lookbehind which rules out the case
when a underscored username is present in code, e.g., `@_dev_` will not be escaped.
* Pluck over map
* Explain why map makes sense there
* Refactor reactions controller a bit and iterate only once
* Use group by instead of N counting queries
* More positive
* Simplify BufferedArticlesController
* Less queries for MailchimpBot
* Use Rails instead of SQL
* Build comment_ids only when needed
