From f3661dce09bc715a46c01f7ea57694e06b587f29 Mon Sep 17 00:00:00 2001 From: Michael Howell Date: Mon, 10 Jun 2024 15:01:31 -0700 Subject: rustdoc: word wrap CamelCase in the item list table This is an alternative to ee6459d6521cf6a4c2e08b6e13ce3c6ce5d55ed0. That is, it fixes the issue that affects the very long type names in https://docs.rs/async-stripe/0.31.0/stripe/index.html#structs. This is, necessarily, a pile of nasty heuristics. We need to balance a few issues: - Sometimes, there's no real word break. For example, `BTreeMap` should be `BTreeMap`, not `BTreeMap`. - Sometimes, there's a legit word break, but the name is tiny and the HTML overhead isn't worth it. For example, if we're typesetting `TyCtx`, writing `TyCtx` would have an HTML overhead of 50%. Line breaking inside it makes no sense. --- src/librustdoc/html/escape.rs | 44 ++++++++++++++++++++++++ src/librustdoc/html/escape/tests.rs | 57 ++++++++++++++++++++++++++++++++ src/librustdoc/html/format.rs | 3 +- src/librustdoc/html/render/print_item.rs | 6 ++-- 4 files changed, 106 insertions(+), 4 deletions(-) create mode 100644 src/librustdoc/html/escape/tests.rs (limited to 'src/librustdoc/html') diff --git a/src/librustdoc/html/escape.rs b/src/librustdoc/html/escape.rs index ea4b573aeb9..94414913163 100644 --- a/src/librustdoc/html/escape.rs +++ b/src/librustdoc/html/escape.rs @@ -5,6 +5,8 @@ use std::fmt; +use unicode_segmentation::UnicodeSegmentation; + /// Wrapper struct which will emit the HTML-escaped version of the contained /// string when passed to a format string. pub(crate) struct Escape<'a>(pub &'a str); @@ -74,3 +76,45 @@ impl<'a> fmt::Display for EscapeBodyText<'a> { Ok(()) } } + +/// Wrapper struct which will emit the HTML-escaped version of the contained +/// string when passed to a format string. This function also word-breaks +/// CamelCase and snake_case word names. +/// +/// This is only safe to use for text nodes. If you need your output to be +/// safely contained in an attribute, use [`Escape`]. If you don't know the +/// difference, use [`Escape`]. +pub(crate) struct EscapeBodyTextWithWbr<'a>(pub &'a str); + +impl<'a> fmt::Display for EscapeBodyTextWithWbr<'a> { + fn fmt(&self, fmt: &mut fmt::Formatter<'_>) -> fmt::Result { + let EscapeBodyTextWithWbr(text) = *self; + if text.len() < 8 { + return EscapeBodyText(text).fmt(fmt); + } + let mut last = 0; + let mut it = text.grapheme_indices(true).peekable(); + let _ = it.next(); // don't insert wbr before first char + while let Some((i, s)) = it.next() { + let pk = it.peek(); + let is_uppercase = || s.chars().any(|c| c.is_uppercase()); + let next_is_uppercase = + || pk.map_or(true, |(_, t)| t.chars().any(|c| c.is_uppercase())); + let next_is_underscore = || pk.map_or(true, |(_, t)| t.contains('_')); + if (i - last > 3 && is_uppercase() && !next_is_uppercase()) + || (s.contains('_') && !next_is_underscore()) + { + EscapeBodyText(&text[last..i]).fmt(fmt)?; + fmt.write_str("")?; + last = i; + } + } + if last < text.len() { + EscapeBodyText(&text[last..]).fmt(fmt)?; + } + Ok(()) + } +} + +#[cfg(test)] +mod tests; diff --git a/src/librustdoc/html/escape/tests.rs b/src/librustdoc/html/escape/tests.rs new file mode 100644 index 00000000000..f99a2a693a6 --- /dev/null +++ b/src/librustdoc/html/escape/tests.rs @@ -0,0 +1,57 @@ +// basic examples +#[test] +fn escape_body_text_with_wbr() { + use super::EscapeBodyTextWithWbr as E; + // extreme corner cases + assert_eq!(&E("").to_string(), ""); + assert_eq!(&E("a").to_string(), "a"); + assert_eq!(&E("A").to_string(), "A"); + // real(istic) examples + assert_eq!(&E("FirstSecond").to_string(), "FirstSecond"); + assert_eq!(&E("First_Second").to_string(), "First_Second"); + assert_eq!(&E("First_Second").to_string(), "First<T>_Second"); + assert_eq!(&E("first_second").to_string(), "first_second"); + assert_eq!(&E("MY_CONSTANT").to_string(), "MY_CONSTANT"); + assert_eq!(&E("___________").to_string(), "___________"); + // a string won't get wrapped if it's less than 8 bytes + assert_eq!(&E("HashSet").to_string(), "HashSet"); + // an individual word won't get wrapped if it's less than 4 bytes + assert_eq!(&E("VecDequeue").to_string(), "VecDequeue"); + assert_eq!(&E("VecDequeueSet").to_string(), "VecDequeueSet"); + // how to handle acronyms + assert_eq!(&E("BTreeMap").to_string(), "BTreeMap"); + assert_eq!(&E("HTTPSProxy").to_string(), "HTTPSProxy"); + // more corners + assert_eq!(&E("ṼẽçÑñéå").to_string(), "ṼẽçÑñéå"); + assert_eq!(&E("V\u{0300}e\u{0300}c\u{0300}D\u{0300}e\u{0300}q\u{0300}u\u{0300}e\u{0300}u\u{0300}e\u{0300}").to_string(), "V\u{0300}e\u{0300}c\u{0300}D\u{0300}e\u{0300}q\u{0300}u\u{0300}e\u{0300}u\u{0300}e\u{0300}"); + assert_eq!(&E("LPFNACCESSIBLEOBJECTFROMWINDOW").to_string(), "LPFNACCESSIBLEOBJECTFROMWINDOW"); +} +// property test +#[test] +fn escape_body_text_with_wbr_makes_sense() { + use itertools::Itertools as _; + + use super::EscapeBodyTextWithWbr as E; + const C: [u8; 3] = [b'a', b'A', b'_']; + for chars in [ + C.into_iter(), + C.into_iter(), + C.into_iter(), + C.into_iter(), + C.into_iter(), + C.into_iter(), + C.into_iter(), + C.into_iter(), + ] + .into_iter() + .multi_cartesian_product() + { + let s = String::from_utf8(chars).unwrap(); + assert_eq!(s.len(), 8); + let esc = E(&s).to_string(); + assert!(!esc.contains("")); + assert!(!esc.ends_with("")); + assert!(!esc.starts_with("")); + assert_eq!(&esc.replace("", ""), &s); + } +} diff --git a/src/librustdoc/html/format.rs b/src/librustdoc/html/format.rs index d6aed75103d..bb5ac303ffd 100644 --- a/src/librustdoc/html/format.rs +++ b/src/librustdoc/html/format.rs @@ -32,7 +32,7 @@ use crate::clean::utils::find_nearest_parent_module; use crate::clean::{self, ExternalCrate, PrimitiveType}; use crate::formats::cache::Cache; use crate::formats::item_type::ItemType; -use crate::html::escape::Escape; +use crate::html::escape::{Escape, EscapeBodyText}; use crate::html::render::Context; use crate::passes::collect_intra_doc_links::UrlFragment; @@ -988,6 +988,7 @@ pub(crate) fn anchor<'a, 'cx: 'a>( f, r#"{text}"#, path = join_with_double_colon(&fqp), + text = EscapeBodyText(text.as_str()), ) } else { f.write_str(text.as_str()) diff --git a/src/librustdoc/html/render/print_item.rs b/src/librustdoc/html/render/print_item.rs index 24476e80778..3f01c082ba9 100644 --- a/src/librustdoc/html/render/print_item.rs +++ b/src/librustdoc/html/render/print_item.rs @@ -29,7 +29,7 @@ use crate::clean; use crate::config::ModuleSorting; use crate::formats::item_type::ItemType; use crate::formats::Impl; -use crate::html::escape::Escape; +use crate::html::escape::{Escape, EscapeBodyTextWithWbr}; use crate::html::format::{ display_fn, join_with_double_colon, print_abi_with_space, print_constness_with_space, print_where_clause, visibility_print_with_space, Buffer, Ending, PrintWithSpace, @@ -423,7 +423,7 @@ fn item_module(w: &mut Buffer, cx: &mut Context<'_>, item: &clean::Item, items: "
{}extern crate {} as {};", visibility_print_with_space(myitem, cx), anchor(myitem.item_id.expect_def_id(), src, cx), - myitem.name.unwrap(), + EscapeBodyTextWithWbr(myitem.name.unwrap().as_str()), ), None => write!( w, @@ -520,7 +520,7 @@ fn item_module(w: &mut Buffer, cx: &mut Context<'_>, item: &clean::Item, items: {stab_tags}\
\ {docs_before}{docs}{docs_after}", - name = myitem.name.unwrap(), + name = EscapeBodyTextWithWbr(myitem.name.unwrap().as_str()), visibility_and_hidden = visibility_and_hidden, stab_tags = extra_info_tags(myitem, item, tcx), class = myitem.type_(), -- cgit 1.4.1-3-g733a5 From 9186001f3491c0eb996de6f61a457cecfb089333 Mon Sep 17 00:00:00 2001 From: Michael Howell Date: Sat, 6 Jul 2024 19:05:28 -0700 Subject: rustdoc: avoid redundant HTML when there's already line breaks --- src/librustdoc/html/escape.rs | 6 ++++++ src/librustdoc/html/escape/tests.rs | 4 ++++ 2 files changed, 10 insertions(+) (limited to 'src/librustdoc/html') diff --git a/src/librustdoc/html/escape.rs b/src/librustdoc/html/escape.rs index 94414913163..3e20c5b322b 100644 --- a/src/librustdoc/html/escape.rs +++ b/src/librustdoc/html/escape.rs @@ -97,6 +97,12 @@ impl<'a> fmt::Display for EscapeBodyTextWithWbr<'a> { let _ = it.next(); // don't insert wbr before first char while let Some((i, s)) = it.next() { let pk = it.peek(); + if s.chars().all(|c| c.is_whitespace()) { + // don't need "First Second"; the space is enough + EscapeBodyText(&text[last..i]).fmt(fmt)?; + last = i; + continue; + } let is_uppercase = || s.chars().any(|c| c.is_uppercase()); let next_is_uppercase = || pk.map_or(true, |(_, t)| t.chars().any(|c| c.is_uppercase())); diff --git a/src/librustdoc/html/escape/tests.rs b/src/librustdoc/html/escape/tests.rs index f99a2a693a6..7933f23eb74 100644 --- a/src/librustdoc/html/escape/tests.rs +++ b/src/librustdoc/html/escape/tests.rs @@ -9,6 +9,10 @@ fn escape_body_text_with_wbr() { // real(istic) examples assert_eq!(&E("FirstSecond").to_string(), "FirstSecond"); assert_eq!(&E("First_Second").to_string(), "First_Second"); + assert_eq!(&E("First Second").to_string(), "First Second"); + assert_eq!(&E("First HSecond").to_string(), "First HSecond"); + assert_eq!(&E("First HTTPSecond").to_string(), "First HTTPSecond"); + assert_eq!(&E("First SecondThird").to_string(), "First SecondThird"); assert_eq!(&E("First_Second").to_string(), "First<T>_Second"); assert_eq!(&E("first_second").to_string(), "first_second"); assert_eq!(&E("MY_CONSTANT").to_string(), "MY_CONSTANT"); -- cgit 1.4.1-3-g733a5 From 1d339b07ca84743710dc87dc0bc4c0597006ed59 Mon Sep 17 00:00:00 2001 From: Michael Howell Date: Sat, 6 Jul 2024 19:07:22 -0700 Subject: rustdoc: use `` in sidebar headers This also improves sidebar layout, so instead of BTreeM ap you get this BTree Map --- src/librustdoc/html/layout.rs | 2 ++ src/librustdoc/html/render/mod.rs | 2 +- src/librustdoc/html/render/sidebar.rs | 16 ++++++++++++++++ src/librustdoc/html/templates/page.html | 2 +- src/librustdoc/html/templates/sidebar.html | 6 ++++-- 5 files changed, 24 insertions(+), 4 deletions(-) (limited to 'src/librustdoc/html') diff --git a/src/librustdoc/html/layout.rs b/src/librustdoc/html/layout.rs index 7dfcc88398f..780cda9b1cd 100644 --- a/src/librustdoc/html/layout.rs +++ b/src/librustdoc/html/layout.rs @@ -69,6 +69,8 @@ struct PageLayout<'a> { display_krate_version_extra: &'a str, } +pub(crate) use crate::html::render::sidebar::filters; + pub(crate) fn render( layout: &Layout, page: &Page<'_>, diff --git a/src/librustdoc/html/render/mod.rs b/src/librustdoc/html/render/mod.rs index aaac8678264..b5cc495ce41 100644 --- a/src/librustdoc/html/render/mod.rs +++ b/src/librustdoc/html/render/mod.rs @@ -30,7 +30,7 @@ mod tests; mod context; mod print_item; -mod sidebar; +pub(crate) mod sidebar; mod span_map; mod type_layout; mod write_shared; diff --git a/src/librustdoc/html/render/sidebar.rs b/src/librustdoc/html/render/sidebar.rs index 6e826446c0e..101cc839f09 100644 --- a/src/librustdoc/html/render/sidebar.rs +++ b/src/librustdoc/html/render/sidebar.rs @@ -77,6 +77,22 @@ impl<'a> Link<'a> { } } +pub(crate) mod filters { + use std::fmt::Display; + + use rinja::filters::Safe; + + use crate::html::escape::EscapeBodyTextWithWbr; + use crate::html::render::display_fn; + pub(crate) fn wrapped(v: T) -> rinja::Result> + where + T: Display, + { + let string = v.to_string(); + Ok(Safe(display_fn(move |f| EscapeBodyTextWithWbr(&string).fmt(f)))) + } +} + pub(super) fn print_sidebar(cx: &Context<'_>, it: &clean::Item, buffer: &mut Buffer) { let blocks: Vec> = match *it.kind { clean::StructItem(ref s) => sidebar_struct(cx, it, s), diff --git a/src/librustdoc/html/templates/page.html b/src/librustdoc/html/templates/page.html index cdf01fa7a97..65c4304e202 100644 --- a/src/librustdoc/html/templates/page.html +++ b/src/librustdoc/html/templates/page.html @@ -98,7 +98,7 @@ {# #} {% endif %}

{# #} - {{display_krate}} {# #} + {{display_krate|wrapped|safe}} {# #} {% if !display_krate_version_number.is_empty() %} {{+ display_krate_version_number}} {% endif %} diff --git a/src/librustdoc/html/templates/sidebar.html b/src/librustdoc/html/templates/sidebar.html index 3251b4c14c9..025220ab415 100644 --- a/src/librustdoc/html/templates/sidebar.html +++ b/src/librustdoc/html/templates/sidebar.html @@ -1,6 +1,6 @@ {% if !title.is_empty() %}

{# #} - {{title_prefix}}{{title}} {# #} + {{title_prefix}}{{title|wrapped|safe}} {# #}

{% endif %} diff --git a/tests/rustdoc-gui/label-next-to-symbol.goml b/tests/rustdoc-gui/label-next-to-symbol.goml index 1fa0a120ada..a8363f29dd5 100644 --- a/tests/rustdoc-gui/label-next-to-symbol.goml +++ b/tests/rustdoc-gui/label-next-to-symbol.goml @@ -27,7 +27,8 @@ compare-elements-position-near: ( ".item-name .stab.deprecated", {"y": 2}, ) -compare-elements-position: ( +// "Unix" part is on second line +compare-elements-position-false: ( ".item-name .stab.deprecated", ".item-name .stab.portability", ["y"], -- cgit 1.4.1-3-g733a5 From ac303df4e21eabfbf692f2d1959f7403f4887a31 Mon Sep 17 00:00:00 2001 From: Michael Howell Date: Mon, 22 Jul 2024 09:49:49 -0700 Subject: rustdoc: move the wbr after the underscore, instead of before --- src/librustdoc/html/escape.rs | 8 ++++---- src/librustdoc/html/escape/tests.rs | 8 ++++---- tests/rustdoc/item-desc-list-at-start.item-table.html | 2 +- 3 files changed, 9 insertions(+), 9 deletions(-) (limited to 'src/librustdoc/html') diff --git a/src/librustdoc/html/escape.rs b/src/librustdoc/html/escape.rs index f8199cd6e0e..691f86847b5 100644 --- a/src/librustdoc/html/escape.rs +++ b/src/librustdoc/html/escape.rs @@ -108,13 +108,13 @@ impl<'a> fmt::Display for EscapeBodyTextWithWbr<'a> { || pk.map_or(true, |(_, t)| t.chars().any(|c| c.is_uppercase())); let next_is_underscore = || pk.map_or(true, |(_, t)| t.contains('_')); let next_is_colon = || pk.map_or(true, |(_, t)| t.contains(':')); - if (i - last > 3 && is_uppercase() && !next_is_uppercase()) - || (s.contains('_') && !next_is_underscore()) - { + if i - last > 3 && is_uppercase() && !next_is_uppercase() { EscapeBodyText(&text[last..i]).fmt(fmt)?; fmt.write_str("")?; last = i; - } else if s.contains(':') && !next_is_colon() { + } else if (s.contains(':') && !next_is_colon()) + || (s.contains('_') && !next_is_underscore()) + { EscapeBodyText(&text[last..i + 1]).fmt(fmt)?; fmt.write_str("")?; last = i + 1; diff --git a/src/librustdoc/html/escape/tests.rs b/src/librustdoc/html/escape/tests.rs index e2d81cf5c27..a09649e9e18 100644 --- a/src/librustdoc/html/escape/tests.rs +++ b/src/librustdoc/html/escape/tests.rs @@ -14,16 +14,16 @@ fn escape_body_text_with_wbr() { assert_eq!(&E(" ").to_string(), " "); // real(istic) examples assert_eq!(&E("FirstSecond").to_string(), "FirstSecond"); - assert_eq!(&E("First_Second").to_string(), "First_Second"); + assert_eq!(&E("First_Second").to_string(), "First_Second"); assert_eq!(&E("First Second").to_string(), "First Second"); assert_eq!(&E("First HSecond").to_string(), "First HSecond"); assert_eq!(&E("First HTTPSecond").to_string(), "First HTTPSecond"); assert_eq!(&E("First SecondThird").to_string(), "First SecondThird"); - assert_eq!(&E("First_Second").to_string(), "First<T>_Second"); - assert_eq!(&E("first_second").to_string(), "first_second"); + assert_eq!(&E("First_Second").to_string(), "First<T>_Second"); + assert_eq!(&E("first_second").to_string(), "first_second"); assert_eq!(&E("first:second").to_string(), "first:second"); assert_eq!(&E("first::second").to_string(), "first::second"); - assert_eq!(&E("MY_CONSTANT").to_string(), "MY_CONSTANT"); + assert_eq!(&E("MY_CONSTANT").to_string(), "MY_CONSTANT"); // a string won't get wrapped if it's less than 8 bytes assert_eq!(&E("HashSet").to_string(), "HashSet"); // an individual word won't get wrapped if it's less than 4 bytes diff --git a/tests/rustdoc/item-desc-list-at-start.item-table.html b/tests/rustdoc/item-desc-list-at-start.item-table.html index ab8b1508b55..cff4f816529 100644 --- a/tests/rustdoc/item-desc-list-at-start.item-table.html +++ b/tests/rustdoc/item-desc-list-at-start.item-table.html @@ -1 +1 @@ - \ No newline at end of file + \ No newline at end of file -- cgit 1.4.1-3-g733a5