package river
All checked builds pass sectionYPositions = computeSectionYPositions($el), 10)"
x-init="setTimeout(() => sectionYPositions = computeSectionYPositions($el), 10)"
>
RSS2 and Atom feed aggregator for OCaml
Install
dune-project
Dependency
Authors
Maintainers
Sources
0.6.1.tar.gz
md5=455810b0b99d3b2944380cf73f6065b9
sha512=a5827aa9557fdd93577ba57f2c8ff0cabc2f43c54c5e98032c07f2efc737c293334b7791eac32043045082229fe5fc852d77eb2ab130c12023e97dfbaaf38189
doc/CHANGES.html
0.6.1 - 2026-11-08
- Raise the OCaml lower bound to 4.11
Fix post descriptions (#22):
summaryno longer runs words together across block elements (e.g. list items) and collapses whitespace, including non-breaking spaces; scripts and styles are dropped, and Atomtype="xhtml"summaries keep their block boundaries.meta_descriptionnow finds the standard<meta name="description">tag (case-insensitively), keepingproperty="description"as a fallback, and skips empty tags.- Behaviour change: an RSS2
<description>is only treated as a summary when the item also has a non-blank<content:encoded>. Otherwise it is the post content (often the whole article),summaryisNone, andmeta_descriptionfetches the post's page to read its meta tags.
0.6 - 2026-09-02
- Add an opt-in
?repairflag toRiver.fetch(and a newRiver.of_string) that recovers feeds which are valid Atom/RSS apart from unclosed HTML void elements (e.g.<img>) intype="xhtml"content, by self-closing them before parsing. Off by default, so well-formed feeds are never modified. Also exposesRiver.sanitize_void_elements. - Extend
?repairto also recover well-formed Atom feeds that Syndic rejects because an<entry>is missing the mandatory<updated>element: it is defaulted to the entry's<published>, else the feed-level<updated>(#20). Also exposesRiver.repair_missing_updated. - Respect the feed's own entry identity (Atom
<id>, RSS2 guid) when generating Atom entries, resolve relative entry links to absolute URLs, and deduplicate merged posts (#13) - Use the feed's own
<summary>(Atom) /<description>(RSS2) when available: add asummaryaccessor and makemeta_descriptionprefer it over scraping the origin HTML page (#12)
0.5 - 2026-08-25
- Expose timeout and user-agent in the public API (#16)
- Remove unused ocamlnet dependency
0.4 - 2024-11-08
- Replace ocamlnet HTML parser with Lambda Soup (#15, @aantron)
0.3 - 2023-11-21
- Fall back to entry id if entry links doesn't exist (#11, @sabine)
0.2 - 2022-04-14
- Build with dune.
- Make the types abstract and add accessor functions.
- Support fetching meta description and SEO image from the posts links.
0.1.3 - 2015-07-28
- Make river compatible with the latest syndic API
0.1.2 - 2015-03-24
- Refactoring modules.
0.1.1 - 2015-03-19
- Upgrading version number.
0.1 - 2015-03-15
- Initial release
sectionYPositions = computeSectionYPositions($el), 10)"
x-init="setTimeout(() => sectionYPositions = computeSectionYPositions($el), 10)"
>