merge upstream#1
Open
emhl wants to merge 240 commits into
Open
Conversation
…465) Co-authored-by: Alexandra Zaharia <alexandra.zaharia@platform.sh>
Solves #455 by statically configuring the page source to use the source and description of the first image in bigimg. When the description is empty, also pre-hides the description box. When JavaScript is enabled, this theme's main.js (https://github.com/halogenica/beautifulhugo/blob/master/static/js/main.js) takes over and will overwrite this style attribute, changing the image in use.
Co-authored-by: Henry Schreiner <HenrySchreinerIII@gmail.com>
Ox-hugo uses an array for authors by default which breaks this theme. Some themes already support multiple authors e.g. ananke
Adding Disclaimer Text to the footer. Signed-off-by: Robert Terakedis <rterakedis@vmware.com> Co-authored-by: Robert Terakedis <rterakedis@vmware.com>
Signed-off-by: Henry Schreiner <henryschreineriii@gmail.com>
Signed-off-by: Henry Schreiner <henryschreineriii@gmail.com>
Signed-off-by: Henry Schreiner <henryschreineriii@gmail.com>
Signed-off-by: Henry Schreiner <henryschreineriii@gmail.com>
This reverts commit 1d98d70.
This reverts commit 3c55b34.
Signed-off-by: Henry Schreiner <henryschreineriii@gmail.com>
Signed-off-by: Henry Schreiner <henryschreineriii@gmail.com>
* Convert theme to hugo module * ci: bump PR build too Signed-off-by: Henry Schreiner <henryschreineriii@gmail.com> * Bump hugo SSG to latest version * Update .github/workflows/ci.yml --------- Signed-off-by: Henry Schreiner <henryschreineriii@gmail.com> Co-authored-by: Henry Schreiner <henryschreineriii@gmail.com>
* seo/article: Fix dateModified and add timezone * Remove new line at the end of the file
* Fixes Upstream Issue 349 - tag link has double forward slash * Updated fix for Upstream Issue 349 - tag link has double forward slash
* Add cusdis support * update example website config.toml
added logo manager for asset, if is assets, optimize the logo in Hugo pipelines
* Site.Author is deprecated, use Params.author instead * Update footer.html * Update footer.html * Update footer.html * Update footer.html * Update footer.html * Update footer.html * Update footer.html * Update ci.yml * Update ci.yml * Update footer.html * Update ci.yml * Update ci.yml --------- Co-authored-by: Henry Schreiner <HenrySchreinerIII@gmail.com>
* Fixed deprecated site.isServer param to hugo.isServer param * Also fixed another deprecated function for disqus * Updated README.md In the face of that we will update to 0.120 it would also be nice to update the README to reference the hugo.* file as the config instead of the config.* file. I also changed the description of the disqus feature to match the updated version of the code * Added checks for hugoVersion and disqus feature * Fixed accidental delete * Removed new line
* fix: correct paths for subdirectory deployments - Remove absolute '/index.json' fallback in search-fuse.js that bypassed relURL for subdirectory deployments - Guard footer commit link with .Site.Params.commit check so 'commit = false' doesn't render 'false' + hash in the URL - Clean up redundant 'and' in show-source.html GitInfo conditional Assisted-by: opencode:glm-5 * fix: default search provider to fuse instead of none When [Params.search] is absent from config, the provider defaulted to "none" instead of "fuse", disabling search unexpectedly. The search-provider partial now defaults to "fuse" so search works out of the box. Assisted-by: opencode:glm-5 * Apply suggestion from @henryiii
relURL with a leading slash (e.g. "/index.json") produces a server-root path, ignoring the baseURL subpath. Remove the leading slash so relURL correctly prepends the path component (e.g. "/beautifulhugo/index.json" instead of "/index.json"). Also add comment= to example site config. Assisted-by: OpenCode:glm-5
Add callout, include-code, and no-ai-summary shortcodes. Add recipe content type, TOC panel, robot/AI meta tags, view source button, custom HTML hooks, and MathJax support. Replace outdated 'Home page callout' with slide-out TOC panel. Assisted-by: OpenCode:glm-5
Same relURL leading-slash issue as the search index URL: "/css/search.css" and "/js/search.js" with relURL produce server-root paths that ignore the baseURL subpath. Remove leading slashes so they resolve correctly under subpath deployments like /beautifulhugo/. Assisted-by: OpenCode:glm-5
Signed-off-by: Henry Schreiner <henryfs@princeton.edu>
…#744) The search page uses type:search, so the header partial generates a search-heading class, but CSS only had rules for page-heading, tags-heading, and categories-heading. This caused the search page title to be left-aligned with inconsistent font sizing compared to other pages. Assisted-by: OpenCode:glm-5
The vendored Bootstrap 5.3.5 CSS does not include [data-bs-theme="dark"]
alert overrides. Since the Beautiful Hugo theme uses a custom
data-theme="dark" attribute, the six main Bootstrap alert types
(primary, secondary, info, success, warning, danger) retained their
light-mode pastel backgrounds and dark text in dark mode, making them
nearly unreadable.
Add explicit [data-theme="dark"] overrides for .alert-{type}.callout
border, background, and text colors (and their .alert-heading colors)
to static/css/dark.css, matching Bootstrap 5's dark emphasis palette.
Assisted-by: OpenCode:Kimi-K2.6
Override Bootstrap tooltip background color from black (#000) to dark grey (#343a40) for better visual appearance. Assisted-by: OpenCode:NRP/kimi
* feat(search): decode HTML entities in search results Add a `decodeHtml` helper function to ensure that HTML-encoded characters in page titles and excerpts are properly decoded before being escaped and rendered in the search results dropdown. This prevents double-encoding or raw entity display in the UI. * refactor: replace DOM element creation with DOMParser for HTML decoding
Signed-off-by: Henry Schreiner <henryfs@princeton.edu>
…#753) * fix(seo): self-canonicalize paginated pages without breaking listings The canonical link added in #731 used `.Permalink`, which makes every paginated page (`/page/2/`, `/post/page/2/`, …) canonicalize to page 1. Google now recommends paginated pages self-canonicalize instead. Switching to `.Paginator.URL` (as proposed in #750) fixes the URLs but introduces a silent bug: the SEO partial runs in <head>, before the body, so calling `.Paginator` there builds the default, *unfiltered* paginator first. Hugo caches one paginator per page and hands that cached set to the body's filtered `.Paginate`, ignoring the filter — `hidden` posts then leak onto the home page. Fix by centralizing the list collection so the <head> canonical and the <body> listing paginate the exact same set: - Add `partials/func/list-pages.html`: single source of truth for a list node's collection (home → mainSections regular pages; section/term → `.Pages`), with `hidden` pages filtered out before pagination so they neither show nor leave gaps in the page numbering. - Add `partials/seo/canonical.html`: overridable canonical tag. Resolution order is front-matter `canonicalURL` → paginator URL (home/section/term) → permalink. Honors `disableCanonical` (site-wide or per-page). The taxonomy index keeps using `.Permalink`, so no phantom paginated pages. - Route `seo/main.html`, `index.html`, and `list.html` through the shared collection partial. - Document canonical behavior in the example site (seo-and-i18n.md, configuration.md). Verified against the example site: the 6 paginated pages now self-canonicalize, all other canonicals and every list page's contents are unchanged, and a `hidden` post no longer leaks into any listing. Closes #751. Supersedes #750. Assisted-by: ClaudeCode:claude-opus-4.8 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(seo): use absLangURL for canonical links on multilingual sites Per PR review: relative `canonicalURL` overrides ran through `absURL`, which resolves against the global baseURL and drops the language prefix on multilingual sites. Switch to `absLangURL` so relative overrides stay language-correct. Verified on a two-language test site that this is safe in every case: - paginator URLs already carry the language prefix, and absLangURL does not double it (…/fr/post/page/2/ stays correct); - relative overrides become language-aware (/fr/post/original/); - absolute override URLs (https://other.com/…) are emitted unchanged. Single-language sites are unaffected. Doc wording updated to match. Assisted-by: ClaudeCode:claude-opus-4.8 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* chore: Update partial to return a dict of schema data Add @id to be reference by WebPage Add publisher which references #organization Remove @context as this will be contained within the @graph Chnage partial to return a dict which will be used when building the @graph schema. This will merge all ld+json into a single context. * chore: Update partial to return a dict and add missing data Remove @context as this will be part of a @graph context Add @id #webpage Add isPartOf which links the WebPage to the #website * chore: Update partial to return a dict This makes it ready to be used as part of a new @graph schema context * chore: Update partial to return a dict This just makes it ready to use with a new @graph context, unifying the schema. * typo: Wording fix * fix: Provide correct support for pagination and breadcrumb creation The previous implemenation had several issues. - the url for home "/" was not unique across homepage pagination. This meant the @id was not truely unique for #breadcrumb. - The absolute "/" should not contain a #breadcrumb. Its a bit of circular reference, as breadcrumbs are intended to show how to get to the current page from the sites root page. However, breadcrumbs should exist for the homepage where pagination has ocurred. Such as /page/2 /page/3 and so on This new implemenation provides these elements. - "/" does not contain #breadcrumb - Paginated pages such as /page/2 now have unique #breadcrumb @ids in the form of /page/2/#breadcrumb and so on. * feat: Add CollectionPage for .IsNode content Add a #collectionpage @id - A CollectionPage supports #breadcrumb. This means that WebPage is not needed in the schema for these pages. - Correctly avoid adding #breadcrumb for "/" but adds it for "/page/2" and so on. - Pagination is limited to pages that are of "post". This might need to be revisist? Or it may impact how others use the theme if they are heavily custimised. This is just speculation on my part. * fix: Update partial to return a dict and improve data Remove @context remove "isPartOf" and replace with "mainEntityOfPage" to main link to #webpage * feat: Switch to a single @graph schema Rework schema to generate a single inified @graph schema. This makes use of the partials that have been updated to return dictionaries of their related data. *May want to consider remove the "recipe" from the theme layout and instead creating a custom schema.html for the example site within its own layout strucuture* * fix: tag for multiple author support If there are multiple author the "#author" alone does not provide uniqueness. - Create unique id using #athor-name name constructed using the urlize function - Each author has their own "Person" object created with their unique @id. - The article then gets an "author" attribute that is an array of "@id"s to link to the "Person" objects. This ensure uniquness. #Note: This still supports the case where there is only a single author. The array structure does not create any issues. * delete: Remove "publisher" within the article schema Remove the "publisher" organization object. This is no unique as it uses the "#organization" tag. Multiple Organization blocks can be used but each must be unique. The Site publisher gets a "#organization" tag. You might then create another Organization denoted where the author works if they are not work for the same organization. This new organization would have a differe @id such as #author-org. This is merely a tag and could be anything as long as it is unique. The author object would then need to contain a "worksFor" attribute that would have an @id referecnes #author-org. * fix: Add the publisher attribute using @id * fix: Update recipe.html to be similar to article.html * fix: keywords list correctly as done in article.html * fix(issue 2): This addresses the issue 2 of the copilot review * fix(issue 3): This addresses issue 3 reported by copilot * fix: Switch to using func/list-pages.html * refactor(seo): clean up structured-data partials - Remove dead old-breadcrumb.html backup partial - Drop unused $currentPageUrl variable in schema.html - Reuse the .Store paginator in collection-page.html instead of re-paginating with a hardcoded 'post' section, keeping the collection consistent with the canonical link and body listing Assisted-by: ClaudeCode:claude-opus-4.8 --------- Co-authored-by: Adam.S <Bas-Man@users.noreply.github.com> Co-authored-by: Henry Schreiner <henryfs@princeton.edu>
Update the CI test matrix to replace 0.161.1 with 0.163.0 as the latest Hugo release. Also fix AGENTS.md docs to reflect the current matrix: 0.146.2, 0.155.2, 0.163.0. Assisted-by: OpenCode:glm-5
* Added uptional dates * Update to docs * Updates * Tweak * Done
Signed-off-by: Henry Schreiner <henryfs@princeton.edu>
Inside the comment range loop the dot is the comment data map, so .Site.Params.dateformat silently resolved to nil and comments always used the shortdateFormat fallback. Use $.Site to reach the page context. Assisted-by: ClaudeCode:claude-sonnet-4-6
The dateformat validator only recognized the numeric (01) and full (January) month forms, so valid Go layouts like "Jan 2, 2006" - the canonical Go reference format - failed the build. Accept the abbreviated (Jan) form as well, in both post_meta and page_meta. Assisted-by: ClaudeCode:claude-sonnet-4-6
seo/schema.html built the paginator from the unfiltered .Pages for non-home nodes. Since it runs in <head>, it won the one-paginator-per- page race and the filtered .Paginate in list.html was silently ignored, so pages with hidden: true appeared in section/term listings and their JSON-LD. Route all node kinds through func/list-pages, which already handles the filtering. Assisted-by: ClaudeCode:claude-sonnet-4-6
* feat: add mathjax switch Assisted-by: OpenCode:Kimi-K2.6 Signed-off-by: Henry Schreiner <henryfs@princeton.edu> * chore: minor touchup Signed-off-by: Henry Schreiner <henryfs@princeton.edu> * feat: per-page settings Assisted-by: OpenCode:Kimi-K2.6 Signed-off-by: Henry Schreiner <henryfs@princeton.edu> * fix: load MathJax even when selfHosted is enabled head.html emitted the window.MathJax config for mathEngine=mathjax regardless of selfHosted, but scripts.html only loaded the MathJax library in the CDN branch, so selfHosted sites got a config and no library - math silently broke. MathJax is not vendored, so load it from the CDN in both branches and document the behavior. Assisted-by: ClaudeCode:claude-sonnet-4-6 --------- Signed-off-by: Henry Schreiner <henryfs@princeton.edu>
The inline script for delayed Disqus loading never closed its
$(function(){ wrapper, so the whole script failed to parse and the
"Show comments" button did nothing.
Assisted-by: ClaudeCode:claude-sonnet-4-6
The noscript tracking pixel hardcoded http://, which is blocked as mixed content on https sites. Assisted-by: ClaudeCode:claude-sonnet-4-6
The fuse search index URL used relURL, which always resolves to the default-language /index.json. Multilingual sites emit one index per language, so non-default languages searched the wrong index. Use relLangURL, matching the adjacent searchPageURL. Assisted-by: ClaudeCode:claude-sonnet-4-6
Params.comment was a typo'd leftover of Params.commit (set to false above it) and is read by nothing. The commented AISearchSummary block referenced a config scheme that was never implemented; the real feature is the [Params.seo] ai-summary-limit map. Assisted-by: ClaudeCode:claude-sonnet-4-6
…768) StumbleUpon shut down in 2018. The email icon used fab fa-at, but fa-at is a solid icon so nothing rendered. The Twitter share URL emitted a dangling via= when author.twitter was unset. Assisted-by: ClaudeCode:claude-sonnet-4-6
printf "%d" emitted max-snippet:%!d(string=50) when the front matter value was a quoted number. Cast to int before formatting. Assisted-by: ClaudeCode:claude-sonnet-4-6
relURL/absURL treat a leading-slash input as host-root relative, so user-config paths like logo = "/img/avatar.png" rendered without the baseURL subpath and 404'd on sites deployed under a path prefix. Strip the leading slash before relURL/absURL for user-supplied image params (favicon, logo, bigimg, post previews, figure shortcodes, gallery, twitter/JSON-LD images), leaving page-relative figure paths untouched since RelPermalink already carries the prefix. Assisted-by: ClaudeCode:claude-sonnet-4-6
- seo/opengraph.html had no callers - layouts/index.html duplicated _default/list.html minus fullWidth; the home page now falls back to list.html and gains fullWidth - layouts/index.json inlined the search_index partial verbatim; home JSON falls back to _default/index.json - _default/list.json made every section emit an identical copy of the site-wide search index that nothing fetches - terms.html carried an unreachable range-else fallback - header.html had a fileExists conditional with identical branches Assisted-by: ClaudeCode:claude-sonnet-4-6
- func/date-format.html: dateformat resolution + validation shared by post_meta and page_meta - seo/structured/author-schema.html: author Person/@id JSON-LD builder shared by article and recipe schemas - func/menu-name.html: menu i18n lookup shared by top-level and dropdown entries in the navbar - func/toc-state.html: TOC/posts-panel availability shared by the navbar toggle and the panel itself Built example site is byte-identical before and after. Assisted-by: ClaudeCode:claude-opus-4-8
- footer: compute social href once instead of duplicating the <a> tag - github-buttons: drop identity and()/redundant gating - share-links: trust the caller's socialShare gate (single.html already uses .Param, which has identical precedence) - search index: .Plain is already plain text; plainify was re-parsing it as HTML and stripping code tokens from the index - meta-robots: append each tag once instead of accumulating duplicates for uniq to clean up - article/recipe schema: replace assignment-inside-return with a plain pipeline - post_meta: emit separators based on whether anything rendered before, not on whether the immediately preceding feature is on Assisted-by: ClaudeCode:claude-opus-4-8
- paginator prev/next now use .Paginator.URL, which respects a custom pagination path and links page 1 at the canonical section URL instead of the /page/1/ alias redirect - archive selects mainSections pages by Type, matching every other listing in the theme - in-site page links use .RelPermalink / relLangURL per the theme's relative-URL convention; JSON-LD, share, comment, and translation links intentionally stay absolute Assisted-by: ClaudeCode:claude-opus-4-8
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 free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
20 participants
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.