Podlite 2.0 released

Podlite 2.0 is tagged. Podlite is a block-based markup language built around typed blocks and explicit boundaries — the same document is meant to read cleanly whether a person or a tool parses it. This release adds eight blocks and attributes and changes two parsing rules.

The specification is at podlite.org/specification; the full changelog sits inside the spec under =head2 v2.0.

The Coming in Podlite 2.0 article from the review window covered what is new in depth. This post focuses on what to do now: how to migrate existing documents and where to find the rest. For most documents the answer is short — a well-formed v1.0 document renders unchanged under v2.0.

Breaking changes

Two changes parse differently than before. Neither touches a well-formed document — if anything needs updating, it is a parser, not your text.

Legacy attribute syntax removed

A few outdated string attribute formats are gone. The bracket form (:key<value>) and the parenthesized form (:key('value')) remain.

If a document uses the current syntax, nothing changes.

=include is now a directive

=include always behaved like a directive, but the spec previously listed it under block types. Tokenization rules differ between directives and blocks. Parsers built against the v1.0 spec must move =include into the directive dispatch path alongside =config and =alias.

For document authors: no change. =include still takes the same syntax and produces the same output. The reclassification matters only for tools that build ASTs.

New features at a glance

Eight additions ship in v2.0. Existing documents render unchanged.

=boundary: a typed section divider. Renders as a horizontal rule, exposes structure to tools.
=set: pre-configure attributes for the next block. Multiline values, inline markup, lexical scope.
G<> + :masked: content masking. Inline mark or whole-block attribute; hidden by default, revealed by render condition.
=data-table block: renders CSV or TSV as a table. Three source forms (inline body, external :src<file:X>, =data reference). :columns selects projection, :rename{} overrides display names.

  =begin data-table :caption('Quarterly sales') :mime-type<'text/csv; header=present'>
  quarter,product,revenue
  Q1,Widget,12000
  Q1,Gadget,18000
  Q2,Widget,15400
  Q2,Gadget,21200
  =end data-table

Extended :mime-type syntax: accepts MIME parameters per RFC 6838, with header=present|absent for CSV/TSV header indication per RFC 4180 §3.
Markdown fenced code block attributes: the string after the language tag is now parsed as Podlite attributes.
Per-link configuration attributes for L<>: settings stay with the link rather than scattering through prose.
=table error recovery: explicit warnings for malformed rows (cell count mismatch, mixed separators), graceful fallback instead of silent corruption.

For full examples and rationale, the Coming in Podlite 2.0 article is still accurate.