This commit extracts the heading ID generation logic into a shared utility
and updates both the route generation script and the markdown pipeline to use it.
This ensures consistency between the generated routes and the rendered
documentation, and fixes an issue where custom heading IDs (`{#id}`)
were ignored during route generation.
Fixes#67200
(cherry picked from commit bedfcb5644)
The tokenizer regex pattern `[^<]*` was consuming all non-`<` content
before custom HTML tags, causing lost content.
Changed to `\s*` which only allows leading whitespace, letting marked
properly tokenize preceding content.
(cherry picked from commit bc473cf60f)
Previously our system only found the first matching symboling to link to within a code block on each line, now we set up a link for all of the discovered symbols on each line
Fixes#65403
The list of valid links is generated from navigation data configuration in the ADEV app.
Redirections are knowingly exclude so we stop referencing them.
Adds the ability to hide the copy button on code snippets.
Updates documentation to use the new `hideCopy` option with the new
markdown code fence syntax.
Previously we extracted the regions and only made the visible region visible on the page, but since there is not a need for the
non-visible code to be presented, we don't need to include it in the generated code snippets.
Move the header id tracking into the renderer context rather than a global state. Since the RendererContext is
initialized for each execution of the marked parser, we can store the context of which header ids have been seen within
a single page there which allows us to run these parse interactions in parallel.
Rather than manually calling the transformer we use for headings from within a render function, we now properly call into the Renderer to do them
via the tokens they are require instead. This is being done to allow us to have a per Renderer instance of state instead of a global state, which will
allow us to run marked async and hopefully speed up the pipeline
Improves the symbol linking logic to handle Angular component selectors (e.g., ngCombobox). It attempts to convert Angular selector patterns to their corresponding class names, improving navigation to Angular API documentation.
Migrate docs-alert to be inline instead of block, this ensures that the content is placed within a <p> tag as expected. This is important
for ensuring that the iconography is placed correctly within the wrapper div generated for the alert. Additionally, we refactor the matcher
code to be more efficient, running with fewer loops.
Code blocks with `bash` language identifier were not rendering the `$`
prefix, while `shell` blocks did.
This ensures consistent command-line prompt rendering across both
`bash` and `shell` code blocks in the documentation.
Adds a new landing page for developer events, starting with the v21 page. It also includes an ICS file link for calendar invites. Also updated the tests for the docs-pill to support two new attributes so developers can add the event to their calendars.
The ExampleViewer component was extracting only innerHTML from code
blocks in multifile examples, which lost the shell class applied by
the formatCode function. This caused the $ prompt to not appear for
shell commands in multifile blocks even when language="shell" was set.
Modified the Snippet interface to track shell language state and
updated getCodeSnippetsFromMultifileWrapper and getStandaloneCodeSnippet
methods to preserve the shell class. Updated example-viewer template
to conditionally apply the shell class to the code wrapper element.
Fixes inconsistency between standalone and multifile shell code blocks.
