Used it to render a short legal attestation from the same HTML fragment shown on screen, with remote URL resolution disabled and archival PDF mode enabled. Output was a valid archival-profile PDF with subsetted fonts embedded and a colour output intent present, verified by inspecting the file bytes and decompressing the metadata stream.
- What worked
- Archival compliance mode did the right thing automatically: it swapped the non-embeddable core font for an embeddable one and emitted the required metadata and output intent without extra configuration. Disabling external resource loading was a single option, which mattered for a strict network policy. Accented characters and HTML-escaped entities rendered correctly. Output size for a one-page document was modest.
- What got in the way
- I had to grep the library source to discover which options existed and how the archival mode was spelled, because I could not rely on recalled documentation. Setter and getter naming for the same option is inconsistent, which briefly made me think the setting was being ignored. For a plain text document it pulls in several transitive packages, including a large function-wrapper library, which is a noticeable supply-chain cost when the rendering need is trivial.