release: 0.11.0 - #166
Open
WaylonWalker wants to merge 38 commits into
Open
Conversation
Deploying markata with |
| Latest commit: | 6b6a59a |
| Status: | ✅ Deploy successful! |
| Preview URL: | https://6111230d.markata.pages.dev |
| Branch Preview URL: | https://develop.markata.pages.dev |
- Add get_templates_mtime() to track all template file changes including includes/extends - Update feeds.py cache keys to include: - All post metadata via post.to_dict() instead of just content - Template modification times to detect template changes - Update post_template.py cache keys to include template mtime - Fixes issue where feeds don't update when: - Post metadata changes (title, date, slug, published, description) - Published status flips from scheduled to published - Template files are modified (main or included templates) - New posts are added to feeds This eliminates the need for manual cache busting when making common changes.
- redirects.py: Include template mtime in cache key to detect template changes - jinja_md.py: Include post metadata and markata version in cache key Previously: - Changing redirect template wouldn't regenerate redirect files - Changing post metadata in jinja templates wouldn't invalidate cache - Changes to markata context wouldn't trigger re-render Now cache properly invalidates when: - Redirect template file is modified - Post metadata changes that affects jinja rendering - Markata version changes (API changes)
Improve cache key consistency across plugins: - heading_link.py: Replace expensive Path(__file__).read_text() with __version__ - render_markdown.py: Add __version__, backend, and extensions to cache key - auto_description.py: Add __version__ to cache key - seo.py: Add __version__ to cache key Previously: - Changing markdown backend/extensions wouldn't invalidate cache - Version changes wouldn't bust cache for descriptions and SEO Now cache properly invalidates when: - Markata version changes (API/behavior changes) - Markdown backend or extensions change - Plugin code changes (via version bump) This eliminates unnecessary cache hits with stale data and improves performance by removing file I/O from cache key generation.
…aths' - Add get_template_paths() helper function to extract paths from Jinja2 Environment - Update get_templates_mtime() in post_template.py and feeds.py to use helper - Jinja2 Environment stores paths in loader.searchpath, not as direct attribute
- Move get_template_paths(), get_templates_mtime(), and get_template() to jinja_env.py - Remove duplicate implementations from post_template.py and feeds.py - Update plugins to use centralized functions from jinja_env - Rename local wrappers to _get_cached_template() to avoid conflicts - Simplifies template handling and reduces code duplication (net -99 lines) This makes it easier for any plugin to work with templates without reimplementing the same utilities.
- Add @lru_cache to get_template() in jinja_env.py - Remove _get_cached_template() wrappers from post_template.py and feeds.py - Plugins now call get_template(markata.jinja_env, template) directly - Eliminates code duplication: -30 lines Now plugins just need one simple call - no need to implement their own caching wrappers. The centralized function handles everything.
- Document cache invalidation improvements across all plugins - Highlight breaking changes for plugin authors using internal get_template() - Show migration path to centralized jinja_env utilities
* feat: atom support
…attributes
(e.g. {.class-name}), admonitions (!!!, !!!+, ???, ???+), and HTML
comments for cleaner descriptionspaths relative to `output_dir`, preventing files from being written to project root
- Cache expensive feed.map() calls during hash generation (~7.7s savings) - Batch directory creation in feeds plugin (~2s savings) - Only read XSL files when they exist and need comparison - Only write files when content changes in: - to_json (markata.json) - service_worker (service-worker.js) - redirects (redirect HTML files) - jinja_env (head.html template) - post_template (CSS/JS/XSL templates) - feeds (XSL files) Prevents unnecessary file system modifications and downstream syncing.
- Increase diskcache size_limit to 5GB (from default 1GB) - Reduce cull_limit to 10 (from default 100) for faster evictions - Use lightweight post identifiers (slug + content_hash) instead of expensive str(post.to_dict()) in feed hash generation The profile showed 22.5s in a single cache.set() call with: - 3.9s culling (evicting entries at size limit) - 14.65s transaction overhead - 6.4s serializing post.to_dict() for hashing These changes reduce cache contention and eliminate expensive Pydantic repr calls during hash generation.
- Feed cache now uses slug+date+title instead of non-existent content_hash This fixes feeds showing posts but posts not being written - output_html validator now checks if path already contains output_dir before prepending it, preventing markout/markout/ duplication The content_hash attribute doesn't exist on posts, causing all feed cache keys to be identical (all empty strings), which prevented proper cache invalidation.
- Add parse_set_options() for dot notation config parsing - Add _deep_merge() for nested config merging - Support -c/--config for alternate config files - Support -o/--output-dir for output override - Support -s/--set for generic config overrides - Support MARKATA_* environment variables - Enable runtime theme switching and config override - Backward compatible with existing builds
- Document -s/--set flag for runtime config overrides with dot notation - Document -c/--config and -o/--output-dir flags - Document MARKATA_* environment variable support - Note runtime theme switching capability
## Summary Add pagination support to feeds with three pagination types and comprehensive security hardening. ## Features - Add feed pagination configuration (`items_per_page`, `pagination_type`, `enabled`) - Implement three pagination types: `htmx`, `manual`, `js` - HTMX infinite scroll with partial loading - JavaScript-based infinite scroll using Intersection Observer - Manual page navigation with prev/next controls - Improve feed name sanitization with Python identifier conversion ## Security - SHA-256 integrity verification for HTMX downloads (25+ versions supported) - Path traversal protection for feed slugs - XSS prevention in templates using `|tojson` filter - No CDN fallback - fail securely if HTMX download fails - Comprehensive security test suite ## Other Changes - Add pagination static assets (CSS/JS) - Add pagination templates (`feed_partial.html`, `feed_items_partial.html`, `pagination_controls.html`) - Add pagination implementation guide documentation - Bump Python minimum version due to pydantic requirements - Add helper script for managing HTMX integrity hashes
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for freeto join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.