Ordered lists that carry an explicit style (lower-roman or lower-alpha) render with the correct marker in body text, but when the same list is nested inside a prf:* directive (theorem, lemma, proposition, corollary, definition, proof, exercise) the style is dropped and the browser falls back to decimal 1, 2, 3. This is the renderer-side counterpart to the body-list marker support added in #50.
Why it matters (book-dp2). DP2 sets \setlist[enumerate,1]{label=(\roman*)}, so list parts are (i), (ii), (iii) and the prose refers to them as "part (viii)", "part (i)", etc. In body text this now renders correctly. Inside theorem/lemma/proof boxes the markers show 1, 2, 3, which directly contradicts the surrounding prose. Reported by a reader in QuantEcon/book-dp-public#29 (Lemma 2.1.1, Theorem 5.2.18, and several appendix examples).
The markup is correct — only the rendering drops the style. In the built AST, the list node inside the proof/theorem admonition already carries the style:
{ "type": "list", "ordered": true, "start": 1, "style": "lower-roman" }but the emitted HTML is a bare decimal list (no type / list-style-type):
<ol start="1"><li>…</li></ol>
Outside a prf:* directive, an identical style:"lower-roman" node renders with roman markers as expected.
Expected. A style:"lower-roman" / "lower-alpha" list renders with roman / alpha markers regardless of whether it is nested inside a prf:* directive.
Actual. Inside prf:* the list renders decimal.
Scope. Affects both lower-roman and lower-alpha; affects all prf:* admonition types; body lists are unaffected. Fixing this single rendering path clears three separate reader-reported items at once: enumerated parts in theorem/lemma/proposition/exercise boxes, the appendix "index labels", and the Theorem 5.2.18 parts.
Renderer in use: QuantEcon fork qe-v8.
Ordered lists that carry an explicit
style(lower-romanorlower-alpha) render with the correct marker in body text, but when the same list is nested inside aprf:*directive (theorem, lemma, proposition, corollary, definition, proof, exercise) the style is dropped and the browser falls back to decimal1, 2, 3. This is the renderer-side counterpart to the body-list marker support added in #50.Why it matters (book-dp2). DP2 sets
\setlist[enumerate,1]{label=(\roman*)}, so list parts are (i), (ii), (iii) and the prose refers to them as "part (viii)", "part (i)", etc. In body text this now renders correctly. Inside theorem/lemma/proof boxes the markers show1, 2, 3, which directly contradicts the surrounding prose. Reported by a reader in QuantEcon/book-dp-public#29 (Lemma 2.1.1, Theorem 5.2.18, and several appendix examples).The markup is correct — only the rendering drops the style. In the built AST, the list node inside the proof/theorem admonition already carries the style:
{ "type": "list", "ordered": true, "start": 1, "style": "lower-roman" }but the emitted HTML is a bare decimal list (no
type/list-style-type):Outside a
prf:*directive, an identicalstyle:"lower-roman"node renders with roman markers as expected.Expected. A
style:"lower-roman"/"lower-alpha"list renders with roman / alpha markers regardless of whether it is nested inside aprf:*directive.Actual. Inside
prf:*the list renders decimal.Scope. Affects both
lower-romanandlower-alpha; affects allprf:*admonition types; body lists are unaffected. Fixing this single rendering path clears three separate reader-reported items at once: enumerated parts in theorem/lemma/proposition/exercise boxes, the appendix "index labels", and the Theorem 5.2.18 parts.Renderer in use: QuantEcon fork
qe-v8.