summaryrefslogtreecommitdiff
path: root/config
AgeCommit message (Collapse)Author
6 hoursAdd revert!/discard and rebuild nodes#edit around the new hierarchyerdgeist
revert! discards exactly the topmost non-empty layer -- autosave if present, else draft -- and reveals whatever's beneath it, releasing the lock only once nothing is left to protect. Guards against destroying a brand-new, never-published node's only draft, which would violate the (head | draft) invariant every other method here assumes holds; the view's Destroy/Discard button is gated the same way. nodes#edit now calls lock_for_editing! instead of find_or_create_draft, and always displays autosave || draft || head, resurrecting an abandoned session's unsaved content by default with an explicit flash explaining what's shown and how to get back to the last saved version. The view drops content_for :subnavigation entirely: Show becomes "Unlock + Back", Preview stays a plain link, metadata's own <details> already replaced the old toggle, and Publish moves off this page for good, per the earlier decision to manage the publish lifecycle entirely from nodes#show. Save Draft and Save + Unlock + Exit appear both above and below the form, given the body field alone runs 600px on desktop.
7 hoursRoute Save and autosave through the new Node methods (addendum)erdgeist
3 daysAdd public preview links for draftserdgeist
Lets editors share a draft with people outside the admin system, via a random per-page token rather than an enumerable id - /preview/<token>. shared_previews#show is intentionally unauthenticated. Redirects to the real public URL once the page is published. Surfaced on nodes#show (Admin Preview + Public Preview, next to Public Link) as generate/revoke buttons.
3 daysDisable verify_foreign_keys_for_fixtures for test enverdgeist
Adding the occurrences.event_id FK constraint broke fixture loading entirely (every test failing at setup) with a misleading "Foreign key violations found in your fixture data" error. The real cause: this Rails feature does its own proactive integrity check during fixture insertion, independent of the constraint's own deferrability, by directly flipping pg_constraint.convalidated and re-validating - which needs catalog-level privilege the app's DB role correctly doesn't have. No actual fixture data was ever invalid; the check itself couldn't run. Scoped to test.rb only - this never affected real application behavior, and Postgres still fully enforces the actual constraint everywhere, in every environment, regardless of this setting.
3 daysSurface node-kind conventions in the admin UIerdgeist
nodes#new: each kind's radio button now shows a label plus a smaller, muted .field_hint explaining what happens automatically (parent, tags, template) - previously a single undifferentiated string mixing a short label with a long parenthetical, cluttering the smaller kinds (erfa/chaostreff) worst. Title field gets the same hint treatment, telling editors up front that the slug is auto-generated and where to adjust it later - the previous silence here meant editors only discovered slug generation after already committing to a title, with no visible way to fix it. New "Resulting path" row live-previews the full URL (origin + kind's path prefix + a debounced call to the new parameterize_preview route, reusing the exact same slug_for helper "create" now calls, so the preview can never diverge from what actually gets created) as the editor types, plus a copy-to-clipboard button. For "generic", the prefix comes from whichever parent gets chosen via the search widget, not the registry - parent_search.link_closure now stashes the chosen node's unique_name as a data attribute for exactly this. New read-only admin/conventions view dumps the NODE_KINDS registry as a reference table for whoever's confused later about how a given kind behaves - deliberately just renders the same label/hint text used on nodes#new rather than re-describing parent/tags/template separately, since decomposing the parent Proc would mean either invoking it (Update's has a real side effect - creates the year folder) or re-stating its logic a second time in a different shape, either of which risks drifting from the actual behavior. nodes#edit's Template field gets a one-line hint noting the value may already be pre-filled based on how the node was created - the one place the previous commit's new inheritance mechanism actually surfaces on this view.
3 daysAdd erfa/chaostreff node-kind creation conventionserdgeist
lib/ccc_conventions.rb: NODE_KINDS registry replaces the kind-specific branches previously scattered across nodes_controller#create (tags), unique_path-position check). Each kind declares its parent lookup (a Proc - Update's "update"/"press_release" continue to genuinely find-or-create their year-folder via Update.find_or_create_parent; erfa/chaostreff use a plain find_by_unique_name!, since their parent nodes are fixed and a missing one should fail loudly, not be silently created), its tags, and (new) its default template. Node gains a default_template_name column (migration, with backfill for existing update-tree nodes via node.update? - reusing that method rather than re-deriving its unique_path check in raw SQL). Page#set_template now inherits from node.default_template_name, falling back to the old update?-based check only when the column is blank, and only fills in template_name when nothing's already been explicitly chosen - a deliberate change from the previous behavior, which unconditionally overwrote template_name on every save regardless of manual selection. Node#update? itself is unchanged and still used as-is by admin_controller's sitemap filtering - a genuinely different, still valid use of that check. "generic" stays special-cased in the controller, parametrized by params[:parent_id] at request time - doesn't fit "kind implies a fixed lookup" and isn't in the registry. nodes#new's four hardcoded radio buttons and nodes_controller's kind-specific case/when branches are both replaced by iterating/looking up this one registry - adding a new kind now means one new hash entry, not four scattered edits.
6 daysChange open-today widget labelerdgeist
6 daysAdd 'open today' sidebar widget and open days in chapter listserdgeist
- open_erfas_today helper: samples 3 random open-day-tagged occurrences of the current day; widget hidden entirely on days without open chapters; heading shows current weekday - content/_open_erfas_today partial rendered in right column above tags and featured articles - _chapter partial: open-day schedules (event_schedule_text) shown inline in aggregated chapter lists with 'Offene Tage' label - ccc.css: open_erfas_today added to existing right-column widget selector groups (headings, lists, links)
7 daysAdd tagging support to Event modelerdgeist
- Event: acts_as_taggable_on :tags — uses existing polymorphic taggings table, no migration needed - events/edit and events/new: tag_list field added, tag_list permitted in events_controller strong params - locales: add open_days (Offene Tage / Open days), upcoming_events (Weitere Termine / Upcoming events), event_schedule_time, event_schedule_unrecognized, event_schedule_none keys
8 daysAdd RRULE humanizer and wire events into nodes#showerdgeist
- app/models/concerns/rrule_humanizer.rb: new concern included into Event, renders recurring schedule as natural-language German or English from RRULE string; handles WEEKLY/MONTHLY, biweekly (INTERVAL=2), ordinal weekday positions (1TU, -1TH, -2WE), BYMONTH single-month exclusions (December pause convention); gracefully returns nil for COUNT/UNTIL/unrecognized shapes - test/models/concerns/rrule_humanizer_test.rb: 15 tests covering all distinct RRULE shapes found in production data - app/helpers/nodes_helper.rb: add event_schedule_text helper combining humanize_rrule with start_time formatting - app/views/nodes/show.html.erb: add events row, conditionally rendered when node has associated events - config/locales/de.yml, en.yml: add event_schedule_time, event_schedule_unrecognized, event_schedule_none keys
8 daysRemove recent_changes RSS feed. It leaks unpublished drafts and the basic auth was only symbolicerdgeist
10 daysRestore and improve admin node searcherdgeist
- Admin search broke when routing-filter was removed: hardcoded /admin/search and /admin/menu_search URLs in admin_search.js no longer matched the locale-scoped routes. Fix by emitting locale-aware URLs from the layout as JS variables. - Also fixes form submission (POST -> GET, missing = on form_tag), jQuery .attr("value") -> .val() for typeahead input reading, and template name for Rails 8 compatibility. - Adds a visible "search" link to the admin menu so editors can discover the feature without knowing the Alt+F shortcut. Search results now show node path alongside title and link directly to the edit view. Named route admin_menu_search added to routes.rb.
11 daysrss: add per-tag Atom feed at /rss/tags/:tag/updates.xmlerdgeist
- rss#tag_updates action: filters Page.heads by tag name, default locale, 20 items, same caching as updates feed - tag_updates.xml.builder: Atom feed with CGI.escapeHTML on title and summary, consistent with updates.xml.builder - tags/show.html.erb: add subscription link above article list - routes: two routes per existing pattern (format-less + .:format constrained to /xml/)
11 daysStage 7 click-testing fixeserdgeist
- aggregate: switch to shortcode syntax [aggregate ...]; fix paragraph wrapping by excising the shortcode and its surrounding <p> before sanitize, concatenating collection output outside sanitized content - page.rb: remove aggregate XML unwrapping from rewrite_links_in_body (no longer needed with shortcode approach) - rss builders: explicit CGI.escapeHTML on title/abstract; Builder 3.3.0 does not escape when target buffer is html_safe (ActionView default) - tinymce: disable menubar and promotion nag; add code plugin, remove paste plugin (built into TinyMCE 8 core); configure via admin_interface.js directly (config/tinymce.yml affects tinymce() helper only, not tinymce.init)
11 daysFix routes after removing the routing-filtererdgeist
11 daysStage 7: Rails 7.2 → 8.1 on Ruby 3.2.11erdgeist
- Bump Rails to 8.1.3 (Ruby unchanged at 3.2.11, new gemset rails8-upgrade) - config.load_defaults 8.1; merge app:update diffs for all environment files - Remove routing-filter 0.7.0; replace with native scope '(:locale)' in routes.rb and default_url_options in ApplicationController - Delete config/initializers/routing_filter_rails71_patch.rb - Replace vendored TinyMCE 3.x (~200 files) with tinymce-rails ~> 8.3; migrate admin_interface.js from jQuery .tinymce()/advanced theme to tinymce.init(); add config/tinymce.yml; note: TinyMCE 7+ is GPL - rails-i18n ~> 8.0 added explicitly (previously indirect dependency) - awesome_nested_set, acts-as-taggable-on pinned to git main/master (gemspec activerecord < 8.1 ceiling; no functional incompatibility; repin to version once upstream releases updated gemspecs) - globalize ~> 7.0, libxml-ruby ~> 5.0, nokogiri ~> 1.18, pg ~> 1.5 - sass-rails, coffee-rails, uglifier moved from :assets group to main (Sprockets 4 convention; :assets group no longer meaningful) - Node: head, draft, lock_owner marked belongs_to optional: true - Page: node, user, editor marked belongs_to optional: true - Static assets in public/images/ and public/javascripts/ referenced via plain HTML tags; Rails 8 load_defaults raises on pipeline helpers for undeclared assets - sessions_controller_test.rb: remove stale require and dead rescue_action - users_controller_test.rb: assert button[type=submit] not input[type=submit] (Rails 8 button_to renders <button> not <input>) - test_helper.rb: node.reload after children.create! (awesome_nested_set 3.9.0 does not refresh parent in memory after callback) - 129 runs, 339 assertions, 3 failures, 0 errors — identical baseline to 7.2
12 daysAdd minimal storage.yml to satisfy Active Storage engine in productionerdgeist
12 daysFix RSS feeds, xmlparser initializererdgeist
12 daysStage 5 click-testing fixeserdgeist
- Fix link_to :method → button_to for all PUT/DELETE actions - Add button_to CSS reset to admin.css for visual consistency - Fix admin layout: replace broken jquery/jquery_ujs pipeline refs with admin_bundle via sprockets; add sprockets-rails, jquery-ui-rails gems - Add app/assets/javascripts/admin_bundle.js pipeline manifest - Fix event_information helper: use safe_join to avoid double-escaping - Fix nodes_helper: to_s(:db) → to_fs(:db) for event times - Fix revisions view: eliminate nested forms; diff button uses vanilla JS to collect radio button values before POST - Fix config/environments/development.rb and test.rb: cache_classes → enable_reloading - Add routing_filter_rails71_patch.rb version guard - Move LockedByAnotherUser to own file for Zeitwerk autoloading - Fix Globalize fallbacks via config.i18n.fallbacks in application.rb
13 daysStage 5 fixes: RouteWithParams removal, Globalize fallbacks, search stub, to_s(:db) → to_fs(:db), LockedByAnotherUser autoload, test environment configerdgeist
- Remove safe_path helper and content_path shim from link_helper.rb - Update all safe_path call sites in views to use named route helpers directly - Fix Globalize fallbacks via config.i18n.fallbacks in application.rb, remove i18n initializer - Stub Node.search returning none (search disabled pending PostgreSQL upgrade) - Replace to_s(:db) with to_fs(:db) in node.rb, nodes_helper.rb, link_helper.rb, admin view - Move LockedByAnotherUser to app/models/locked_by_another_user.rb for Zeitwerk autoloading - Fix config/environments/test.rb: config.assets removed, cache_classes → enable_reloading, test_order removed, minitest pinned to ~> 5.25 - Fix config/environments/development.rb: cache_classes → enable_reloading - Park search vector migration in doc/ pending PostgreSQL and plpgsql availability
13 daysStage 5: Rails 6.1 → 7.2 on Ruby 3.2.11erdgeist
- Bump Rails to 7.2.3, Ruby to 3.2.11 (new gemset rails7-upgrade) - pg pinned to 1.4.6 (bridge: Ruby 3.2 compatible, PostgreSQL 9.6 tolerant) - acts-as-taggable-on → 12.x (Rails 7.2 + Ruby 3.2 support) - awesome_nested_set → 3.7.0 (Rails 7.2 support, avoids 3.9.0 lft/rgt bug) - globalize → 7.0 (Rails 7.x required) - libxml-ruby → 5.x (Ruby 3.2 required) - unicorn → 6.x; fix unicorn.rb: RAILS_ROOT→Rails.root, RAILS_ENV→ENV, File.exists?→File.exist? - Add puma for development server - Remove dead initializers: assets.rb, ruby2.rb, backtrace_silencers.rb, new_rails_defaults.rb - Fix File.exists? → File.exist? in page.rb and authors_importer.rb - Add routing_filter_rails71_patch.rb (restores params[:locale] on Rails 7.1+)
13 daysStage 4: Rails 5.2 -> 6.1 on Ruby 2.7.2erdgeist
- routing-filter 0.6.3 -> 0.7.0 (Rails 6.1 compatibility) - RSS named routes rss_xml/rss_rdf added - RouteWithParams workarounds: will_paginate_patch, content_path shim, safe_path helper - Paperclip removed, replaced with FileAttachment concern (preserves URL scheme) - Assets resource moved to /admin/assets (Sprockets middleware conflict) - ApplicationRecord base class added, all models migrated - Strong parameters added to Assets, Occurrences, Events, MenuItems controllers - update_attributes -> update throughout - render :nothing -> head :ok/:not_found throughout - language_selector rewritten (removes :overwrite_params) - Environment files updated for Rails 6.1 (eager_load, public_file_server, ActionMailer) - Arel::Visitors::DepthFirst and Integer/Float duration patches removed from test_helper - AssetsController tests added (10 tests covering upload, variants, destroy, auth) - ImageMagick geometry: 460x250! for headline crop (not # which is invalid in IM6) 129 runs, 311 assertions, 5 failures (all pre-existing), 0 errors
13 daysUpgrade to Rails 5.2.8.1 on Ruby 2.5.8erdgeist
- Bump rails 4.2.11.3 → 5.2.8.1, ruby 2.4.10 → 2.5.8 - Upgrade acts-as-taggable-on ~> 3.5 → ~> 6.0 - Upgrade exception_notification ~> 4.4 → ~> 4.5 - Upgrade globalize ~> 5.0 → ~> 5.2.0 - Upgrade pg ~> 0.17 → ~> 1.0 - Upgrade sass-rails ~> 4.0 → ~> 5.0 - Upgrade libxml-ruby to ~> 3.2 (5.x requires Ruby 3.2+) - Pin awesome_nested_set ~> 3.4.0 (3.9 has lft/rgt update bug) - Add rails-controller-testing gem - Add assets initializer for jquery precompile - Add acts_as_taggable missing indexes migration - Set eager_load, serve_static_files, active_record settings
14 daysUpgrade to Rails 4.2.11.3erdgeist
- Bump rails 3.2.22.5 → 4.2.11.3 - Replace globalize3 with globalize ~> 5.0 (gem renamed at 5.0) - Upgrade routing-filter ~> 0.3 → ~> 0.6 - Upgrade sass-rails, coffee-rails to 4.x - Upgrade awesome_nested_set 2.x → 3.x (Rails 4 required) - Add jquery-rails for UJS support - Pin nokogiri ~> 1.10.10, loofah ~> 2.20.0, rails-html-sanitizer ~> 1.4.4 - Add config/secrets.yml (gitignored), eager_load, serve_static_files - Fix routes: add via: to all match calls, remove legacy catch-all routes - Add admin named route, fix rvm dotfiles - Fix ActiveRecord::FixtureSet rename in test_helper - Set active_support.test_order and active_record.raise_in_transactional_callbacks
2026-06-24Upgrade acts-as-taggable-on to 3.5, add Rails 3.2 config fixeserdgeist
- Bump acts-as-taggable-on from 2.4.1 to 3.5.0 - Fake engine migrations for pre-existing schema - Set serve_static_assets in development - Fix request.fullpath in authenticated_system (request_uri removed in Rails 3.x) - Fix Paperclip path format to match existing file layout
2026-06-24Stage 1 complete: Rails 2.3.5 to Rails 3.2.22.5 upgradeerdgeist
- Converted plugins to gems (Gemfile) - Updated config structure (application.rb, boot.rb, environment.rb) - Converted routes to Rails 3 DSL - Converted named_scope to scope throughout models - Converted find(:all, :conditions) to where() chains - Fixed has_many :order to use ordering scope - Updated session store and secret token configuration - Fixed exception_notification middleware configuration - Patched Ruby 2.4 / Rails 3.2 incompatibilities: - Integer/Float duration arithmetic (ActiveSupport) - Arel visit_Integer for PostgreSQL adapter - create_database String/Integer coercion - ActionController consider_all_requests_local - Migrated taggings schema for acts-as-taggable-on - Replaced dynamic_form gem with custom form_error_messages helper - Fixed Rails 3 block helper syntax (form_for, form_tag, fields_for) - Fixed admin layout yield - Updated test suite for Rails 3 APIs
2026-06-23Update deployed state to what's currently runningerdgeist
2025-01-28Big overhaul patch and style changesCharlie Root
2012-02-18small style improvements. more button addedhukl
2012-02-18show number of gallery images if more than onehukl
2011-02-10removed thinking_sphinx plugin and replaced it with gem.hukl
also tuned dependencies
2010-12-29use sanitize instead of only html_escape!User
2010-07-20adding sponsors footerhukl
2010-05-17added i18n initializer to re-enable fallbacks - jesusUser
2010-01-14updated to latest rails stable versionhukl
2009-11-18added template for old RDF feed - needs to be validated thoughhukl
2009-11-16added action mailer config for production environmenthukl
2009-11-16changed email address for exception notifierhukl
2009-11-14nodes :show now also includes a direct link to to the live version of the node. Additionally, I changed the default route to a named route so it is much ↵hukl
easier to generate proper public paths and urls.
2009-11-14added headline to Tag :show view to suggest that it is a list rather than a singe page / article when only one page is tagged with a certain tag.hukl
I still want to avoid thos "more / mehr" links
2009-10-31beta badge and link to old ccc.de pagehukl
2009-10-31added exception notifier plugin to catch all exceptionshukl
2009-10-17added sanitize statements to hopefully all critical templates to protect against cross site scripting.hukl
added section to environment.rb listing the whitelisted tags and attributes and examples on how to extend the list
2009-10-07refactored revsions controller to act as a nested resource of nodes. boy that cleaned up some stuff quite a bit. also having tests for that is just feeling ↵hukl
great
2009-09-12removed sqlite3 sample config as its no longer supported by mehukl
2009-09-12added another gem requirement and updated the chaos_importer task to work with the current node validationshukl
2009-09-10added public rss controller plus template to render the latest 20 updates into a neat little atom feedhukl
2009-09-10going back to raisl 2.3.3 as 2.3.4 has weird issues when translationg active record validation errorshukl
also changing the german error messages to english as the interface is english only so far
2009-09-07fixing searchd configurationhukl
2009-09-07added searchd production settingshukl