Merged
Conversation
a65c75a to
253522c
Compare
253522c to
75179ea
Compare
This was referenced Jan 24, 2025
quachpas
added a commit
to typst-community/glossarium
that referenced
this pull request
Feb 8, 2025
Due to typst/typst#5746, captions are now blocks.
quachpas
added a commit
to typst-community/glossarium
that referenced
this pull request
Feb 8, 2025
Due to typst/typst#5746, captions are now blocks.
stelzo
pushed a commit
to stelzo/typst
that referenced
this pull request
Nov 21, 2025
git download method fixed warinings documentation cli updater adaptation and clippy fixes enhanced documentation add git downloader default impl migrating from git2 to gitoxide crate for git downloads Add support for `c2sc` OpenType feature in `smallcaps` (typst#5655) Just add MathText SyntaxKind Basic SymbolElem addition Use SymbolElem in more places and add `char` cast for content Add SymbolElem to realization Update math TextElem layout to separate out SymbolElem Handle boxes and blocks a bit better in HTML export (typst#5744) Co-authored-by: Martin Haug <3874949+reknih@users.noreply.github.com> Tweak HTML pretty printing (typst#5745) Semantic paragraphs (typst#5746) Fix space collapsing for explicit paragraphs (typst#5749) Support first-line-indent for every paragraph (typst#5768) Fixed typo in the new outline docs (typst#5772) Resolve bound name of bare import statically (typst#5773) Fix typo in scripting.md (typst#5783) Modular, multi-threaded, transitioning plugins (typst#5779) Include images from raw pixmaps and more (typst#5632) Co-authored-by: PgBiel <9021226+PgBiel@users.noreply.github.com> Co-authored-by: Laurenz <laurmaedje@gmail.com> Change type repr to short name (typst#5788) Disable cjk_latin_spacing in raw by default (typst#5753) Change the default math class of U+22A5 ⊥ UP TACK to Normal (typst#5714) Revert adding `flatten-text` to `image` (typst#5789) Refactor `Scope` (typst#5797) Enable HTML feature in docs generator (typst#5800) Scope deprecations (typst#5798) Fix typo in page documentation (typst#5804) Bump openssl from 0.10.66 to 0.10.70 (typst#5802) Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Bump codex to 0.1.0 (typst#5805) Bump dependencies (typst#5808) Fix small copy-paste oversight (typst#5811) Bump more dependencies (typst#5813) Export target docs (typst#5812) Co-authored-by: Martin Haug <3874949+reknih@users.noreply.github.com> 0.13 changelog (typst#5801) Release Candidate 1 Autocomplete content methods (typst#5822) Documentation fixes and improvements (typst#5816) Fix docs outline for nested definitions (typst#5823) Document removals in changelog (typst#5827) Fix unnecessary import rename warning (typst#5828) Don't crash on image with zero DPI (typst#5835) Add warning for `pdf.embed` elem used with HTML (typst#5829) Add smart quotes for Bulgarian (typst#5807) Respect `par` constructor arguments (typst#5842) Bump `typst-assets` Fix autocomplete and jumps in math (typst#5849) Update documentation for `float.{to-bits, from-bits}` (typst#5836) `Gradient::repeat`: Fix floating-point error in stop calculation (typst#5837) Lazy parsing of the package index (typst#5851) Remove Linux Libertine warning (typst#5876) Bring back type/str compatibility for 0.13, with warnings and hints (typst#5877) More robust SVG auto-detection (typst#5878) HTML export: Use `<code>` for inline `RawElem` (typst#5884) --make-deps fixes (typst#5873) Update changelog (typst#5894) Version bump Fix HTML export of table with gutter (typst#5920) Fix comparison of `Func` and `NativeFuncData` (typst#5943) HTML export: fix elem counting on classify_output (typst#5910) Co-authored-by: Laurenz <laurmaedje@gmail.com> Fix introspection of HTML root sibling metadata (typst#5953) Fix high CPU usage due to inotify watch triggering itself (typst#5905) Co-authored-by: Laurenz <laurmaedje@gmail.com> Fix false positive for type/str comparison warning (typst#5957) Fix paper name in page setup guide (typst#5956) Fix curve with multiple non-closed components. (typst#5963) Fix docs example with type/string comparison (typst#5987) Correct typo (typst#5971) Make `array.chunks` example more readable (typst#5975) Hotfix for labels on symbols (typst#6015) Replace `par` function call in tutorial (typst#6023) Mention that `sym.ohm` was removed in the 0.13.0 changelog (typst#6017) Co-authored-by: Laurenz <laurmaedje@gmail.com> Mark breaking symbol changes as breaking in 0.13.0 changelog (typst#6024) 0.13.1 changelog (typst#6025) Version bump dep min 1.81 use 1.81 in ci
stelzo
pushed a commit
to stelzo/typst
that referenced
this pull request
Nov 21, 2025
git download method fixed warinings documentation cli updater adaptation and clippy fixes enhanced documentation add git downloader default impl migrating from git2 to gitoxide crate for git downloads Add support for `c2sc` OpenType feature in `smallcaps` (typst#5655) Just add MathText SyntaxKind Basic SymbolElem addition Use SymbolElem in more places and add `char` cast for content Add SymbolElem to realization Update math TextElem layout to separate out SymbolElem Handle boxes and blocks a bit better in HTML export (typst#5744) Co-authored-by: Martin Haug <3874949+reknih@users.noreply.github.com> Tweak HTML pretty printing (typst#5745) Semantic paragraphs (typst#5746) Fix space collapsing for explicit paragraphs (typst#5749) Support first-line-indent for every paragraph (typst#5768) Fixed typo in the new outline docs (typst#5772) Resolve bound name of bare import statically (typst#5773) Fix typo in scripting.md (typst#5783) Modular, multi-threaded, transitioning plugins (typst#5779) Include images from raw pixmaps and more (typst#5632) Co-authored-by: PgBiel <9021226+PgBiel@users.noreply.github.com> Co-authored-by: Laurenz <laurmaedje@gmail.com> Change type repr to short name (typst#5788) Disable cjk_latin_spacing in raw by default (typst#5753) Change the default math class of U+22A5 ⊥ UP TACK to Normal (typst#5714) Revert adding `flatten-text` to `image` (typst#5789) Refactor `Scope` (typst#5797) Enable HTML feature in docs generator (typst#5800) Scope deprecations (typst#5798) Fix typo in page documentation (typst#5804) Bump openssl from 0.10.66 to 0.10.70 (typst#5802) Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Bump codex to 0.1.0 (typst#5805) Bump dependencies (typst#5808) Fix small copy-paste oversight (typst#5811) Bump more dependencies (typst#5813) Export target docs (typst#5812) Co-authored-by: Martin Haug <3874949+reknih@users.noreply.github.com> 0.13 changelog (typst#5801) Release Candidate 1 Autocomplete content methods (typst#5822) Documentation fixes and improvements (typst#5816) Fix docs outline for nested definitions (typst#5823) Document removals in changelog (typst#5827) Fix unnecessary import rename warning (typst#5828) Don't crash on image with zero DPI (typst#5835) Add warning for `pdf.embed` elem used with HTML (typst#5829) Add smart quotes for Bulgarian (typst#5807) Respect `par` constructor arguments (typst#5842) Bump `typst-assets` Fix autocomplete and jumps in math (typst#5849) Update documentation for `float.{to-bits, from-bits}` (typst#5836) `Gradient::repeat`: Fix floating-point error in stop calculation (typst#5837) Lazy parsing of the package index (typst#5851) Remove Linux Libertine warning (typst#5876) Bring back type/str compatibility for 0.13, with warnings and hints (typst#5877) More robust SVG auto-detection (typst#5878) HTML export: Use `<code>` for inline `RawElem` (typst#5884) --make-deps fixes (typst#5873) Update changelog (typst#5894) Version bump Fix HTML export of table with gutter (typst#5920) Fix comparison of `Func` and `NativeFuncData` (typst#5943) HTML export: fix elem counting on classify_output (typst#5910) Co-authored-by: Laurenz <laurmaedje@gmail.com> Fix introspection of HTML root sibling metadata (typst#5953) Fix high CPU usage due to inotify watch triggering itself (typst#5905) Co-authored-by: Laurenz <laurmaedje@gmail.com> Fix false positive for type/str comparison warning (typst#5957) Fix paper name in page setup guide (typst#5956) Fix curve with multiple non-closed components. (typst#5963) Fix docs example with type/string comparison (typst#5987) Correct typo (typst#5971) Make `array.chunks` example more readable (typst#5975) Hotfix for labels on symbols (typst#6015) Replace `par` function call in tutorial (typst#6023) Mention that `sym.ohm` was removed in the 0.13.0 changelog (typst#6017) Co-authored-by: Laurenz <laurmaedje@gmail.com> Mark breaking symbol changes as breaking in 0.13.0 changelog (typst#6024) 0.13.1 changelog (typst#6025) Version bump dep min 1.81 use 1.81 in ci
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
1 participant
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.
What is this about?
When you add inline-level content to your document, Typst will automatically wrap it in paragraphs. However, a typical document also contains some text that shouldn't semantically be part of a paragraph, for example in a heading or caption. To express this difference, this pull request introduces a few rules that let Typst distinguish between "just text" and a semantic paragraph.
The new rules for when Typst wraps inline-level content in a paragraph are as follows:
All text at the root of a document is wrapped in paragraphs.
Text in a container (like a
block) is only wrapped in a paragraph if the container holds any block-level content. If all of the contents are inline-level, no paragraph is created.In the laid-out document, it's not immediately visible whether text became part of a paragraph. However, it is still important for various reasons:
Certain paragraph styling like
first-line-indentwill only apply to proper paragraphs, not any text. Similarly,parshow rules of course only trigger on paragraphs.A proper distinction between paragraphs and other text helps people who rely on assistive technologies (such as screen readers) navigate and understand the document properly. Currently, this only applies to HTML export since Typst does not yet output accessible PDFs, but support for this is planned for the near future.
HTML export will generate a
<p>tag only for paragraphs.Going forward, when creating custom reusable components, users can and should take charge over whether Typst creates paragraphs. By wrapping text in a [
block] instead of just adding paragraph breaks around it, one can force the absence of a paragraph. Conversely, by adding a [parbreak] after some content in a container, one can force it to become a paragraph even if it's just one word. This is, for example, what non-tightlists do to force their items to become paragraphs.Changes
Concretely, this PR makes the following changes:
The
ParElembecomes a properly constructible element. Previously, calling#par[..]was sort of fake and internally just added parbreaks and applied the relevant styles. In particular, it could result in multiple paragraphs. With this PR,#par[..]creates a real paragraph. If its body happens to contain block-level content, Typst will ignore it and emit a warning (as is the usual way for handling elements that cannot be processed, same as withhtml.elemin layout mode).Paragraphs resulting from realization also become
ParElems. To do that, the already realized[(&Content, StyleChain)]pairs are repacked into sequences and styled elements, yieldingbody: Contentfor the paragraph.When the paragraph is laid out during flow layout, its body is realized again with a new paragraph realization mode. If the body is repacked (previously realized) content (the common case), this mostly just restores the flat list (we pay a small performance price for this duplicated work). However, even in this case, the realization can have an effects due to
parshow rules that might have been applied in the meantime.Fragment (i.e. non top-level) layout and HTML fragment generation change as follows: During fragment realization, we observe whether any block-level content occured. If not, we don't terminate a potentially opened paragraph grouping, instead directly emitting the inline content. We also let the caller know whether there was any block-level content. If there was, layout proceeds as usual. If not, flow layout is put into inline mode where it expects inline elements. These go through normal inline layout, but with knowledge that they do not form a semantic paragraph. In a follow-up PR, I plan to support a new first-line-indent mode that activates even on non-consecutive paragraphs.
Non-paragraph inline layout now ignores hanging-indent and first-line-indent. However, not all paragraph properties are ignored. We respect
leading(after all, we need some value if it breaks over multiple lines). We also respectjustify,linebreaks, andpar.line. I'm not 100% sure how to deal with these, but if I disable them now, there'd be no way at all to get them, so they are untouched for now. And I'm not certain they truly should be tied to semantic paragraphs. For the indents, the situation is more clear cut. We'll need to see.The default show rules of some built-in elements like lists, quotes, etc. are adjusted to ensure they produce/don't produce paragraphs as appropriate. (This was my motivation for the outline work -- outline entries needed to become blocks. Side track much.)
The paragraph documentation was expanded to explain the new distinction.
Notes
ParElemlocatable/queryable now. However, I'm holding off on this for now because it will put a lot of strain on the introspector and we should first evaluate the performance impact more thoroughly. Down the road, I plan to optimize the introspection data structures to make it possible to make many more things (everything?) locatable. Note that, for now, it's still possible to count semantic paragraphs with aparshow rule that dispatches a counter update.