Merge branch 'liw/html-correctness' into 'main'

Various HTML output fixes Closes #336 See merge request subplot/subplot!355
author: Daniel Silverstone <dsilvers+gitlab@digital-scurf.org> 2023-10-10 07:08:21 +0000
committer: Daniel Silverstone <dsilvers+gitlab@digital-scurf.org> 2023-10-10 07:08:21 +0000
commit: 58350fed3216b3ec2875503f6036e1a7f982088e (patch)
tree: 9fb29d0b107d5e1edb78954e68e35270b5fc391c
parent: 478147c59de2547cb1e3e0a0cd5309c25f4453dc (diff)
parent: 76838e218f8c0943de43befce434fa8845a68e86 (diff)
download: subplot-58350fed3216b3ec2875503f6036e1a7f982088e.tar.gz
3 files changed, 73 insertions, 18 deletions
diff --git a/check b/check
index 58407aa..d1edc07 100755
--- a/check
+++ b/check
@@ -320,7 +320,9 @@ def check_subplots(r):
         base = os.path.basename(subplot)
         base, _ = os.path.splitext(subplot)
         base = os.path.join(output, base)
-        r.docgen(subplot, doc_template, base + ".html", cwd=dirname)
+        html = base + ".html"
+        r.docgen(subplot, doc_template, html, cwd=dirname)
+        r.runcmd(["tidy", "-errors", html], cwd=dirname)
 
 
 def tail(filename, numlines=100):
@@ -344,6 +346,7 @@ def check_tooling(r):
         "plantuml",
         "rustc",
         "rustfmt",
+        "tidy",
     ]
     for command in commands:
         if not r.got_command(command):
diff --git a/src/html.rs b/src/html.rs
index f863727..1e0dafa 100644
--- a/src/html.rs
+++ b/src/html.rs
@@ -7,10 +7,13 @@ use line_col::LineColLookup;
 use log::{debug, trace};
 use pulldown_cmark::{CodeBlockKind, Event, HeadingLevel, Options, Parser, Tag};
 use serde::{Deserialize, Serialize};
+use std::collections::HashMap;
 use std::fmt::Write as _;
 use std::io::Write;
 use std::path::{Path, PathBuf};
 
+const DOCTYPE: &str = "<!DOCTYPE html>";
+
 /// A HTML page, consisting of a head and a body.
 #[derive(Debug)]
 pub struct HtmlPage {
@@ -47,8 +50,11 @@ impl HtmlPage {
     pub fn serialize(&self) -> Result<String, HtmlError> {
         let mut html = Element::new(ElementTag::Html);
         html.push_child(Content::Elt(self.head.clone()));
-        html.push_child(Content::Elt(self.body.clone()));
-        html.serialize()
+        let mut body = Element::new(ElementTag::Body);
+        body.push_child(Content::Elt(self.body.clone()));
+        html.push_child(Content::Elt(body));
+        let html = html.serialize()?;
+        Ok(format!("{}\n{}", DOCTYPE, html))
     }
 
     /// Try to write an HTML page as text into a file.
@@ -82,7 +88,7 @@ pub fn parse(filename: &Path, markdown: &str) -> Result<Element, HtmlError> {
     let p = Parser::new_ext(markdown, options).into_offset_iter();
     let linecol = LineColLookup::new(markdown);
     let mut stack = Stack::new();
-    stack.push(Element::new(ElementTag::Body));
+    stack.push(Element::new(ElementTag::Div));
     for (event, loc) in p {
         trace!("event {:?}", event);
         let (line, col) = linecol.get(loc.start);
@@ -150,6 +156,7 @@ pub fn parse(filename: &Path, markdown: &str) -> Result<Element, HtmlError> {
                 Tag::Image(_, url, title) => {
                     let mut e = Element::new(ElementTag::Img);
                     e.push_attribute(Attribute::new("src", url.as_ref()));
+                    e.push_attribute(Attribute::new("alt", title.as_ref()));
                     if !title.is_empty() {
                         e.push_attribute(Attribute::new("title", title.as_ref()));
                     }
@@ -275,6 +282,13 @@ impl Element {
         self.attrs.push(attr);
     }
 
+    /// Drop all attributes with a given name.
+    pub fn drop_attributes(&mut self, unwanted: &[&str]) {
+        for uw in unwanted {
+            self.attrs.retain(|a| a.name() != *uw);
+        }
+    }
+
     /// Append a new child to the element.
     pub fn push_child(&mut self, child: Content) {
         self.children.push(child);
@@ -321,9 +335,11 @@ impl Element {
 
     fn fix_up_img_alt(&mut self) {
         if self.tag == ElementTag::Img {
-            let alt = as_plain_text(self.children());
-            self.push_attribute(Attribute::new("alt", &alt));
-            self.children.clear();
+            if !self.attrs.iter().any(|a| a.name() == "alt") {
+                let alt = as_plain_text(self.children());
+                self.push_attribute(Attribute::new("alt", &alt));
+                self.children.clear();
+            }
         } else {
             for child in self.children.iter_mut() {
                 if let Content::Elt(kid) = child {
@@ -376,9 +392,14 @@ impl Element {
     }
 
     fn serialize_attrs_to_buf(&self, buf: &mut String) -> Result<(), std::fmt::Error> {
+        let mut attrs = Attributes::default();
         for attr in self.attrs.iter() {
-            write!(buf, " {}", attr.name())?;
-            if let Some(value) = attr.value() {
+            attrs.push(attr);
+        }
+
+        for (name, value) in attrs.iter() {
+            write!(buf, " {}", name)?;
+            if !value.is_empty() {
                 write!(buf, "=\"{}\"", encode_double_quoted_attribute(value))?;
             }
         }
@@ -487,6 +508,32 @@ impl ElementTag {
     }
 }
 
+#[derive(Debug, Default, Clone)]
+struct Attributes {
+    attrs: HashMap<String, String>,
+}
+
+impl Attributes {
+    fn push(&mut self, attr: &Attribute) {
+        if let Some(new_value) = attr.value() {
+            if let Some(old_value) = self.attrs.get_mut(attr.name()) {
+                assert!(!old_value.is_empty());
+                old_value.push(' ');
+                old_value.push_str(new_value);
+            } else {
+                self.attrs.insert(attr.name().into(), new_value.into());
+            }
+        } else {
+            assert!(!self.attrs.contains_key(attr.name()));
+            self.attrs.insert(attr.name().into(), "".into());
+        }
+    }
+
+    fn iter(&self) -> impl Iterator<Item = (&String, &String)> {
+        self.attrs.iter()
+    }
+}
+
 /// An attribute of an HTML element.
 #[derive(Clone, Debug)]
 pub struct Attribute {
diff --git a/src/md.rs b/src/md.rs
index 554cc00..8573167 100644
--- a/src/md.rs
+++ b/src/md.rs
@@ -324,6 +324,7 @@ fn extract_scenario(e: &[StructureElement]) -> Result<(Option<Scenario>, usize),
 }
 
 mod typeset {
+    const UNWANTED_ATTRS: &[&str] = &["add-newline"];
 
     use crate::html::{Attribute, Content, Element, ElementTag};
     // use crate::parser::parse_scenario_snippet;
@@ -338,7 +339,7 @@ mod typeset {
     use base64::prelude::{Engine as _, BASE64_STANDARD};
 
     pub(crate) fn typeset_element(e: &Element) -> Result<Element, SubplotError> {
-        match e.tag() {
+        let new = match e.tag() {
             ElementTag::Pre if e.has_attr("class", "scenario") => typeset_scenario(e),
             ElementTag::Pre if e.has_attr("class", "file") => typeset_file(e),
             ElementTag::Pre if e.has_attr("class", "example") => typeset_example(e),
@@ -360,7 +361,10 @@ mod typeset {
                 }
                 Ok(new)
             }
-        }
+        };
+        let mut new = new?;
+        new.drop_attributes(UNWANTED_ATTRS);
+        Ok(new)
     }
 
     fn typeset_scenario(e: &Element) -> Result<Element, SubplotError> {
@@ -378,19 +382,19 @@ mod typeset {
     fn typeset_dot(e: &Element) -> Result<Element, SubplotError> {
         let dot = e.content();
         let svg = DotMarkup::new(&dot).as_svg()?;
-        Ok(svg_to_element(svg))
+        Ok(svg_to_element(svg, "Dot diagram"))
     }
 
     fn typeset_plantuml(e: &Element) -> Result<Element, SubplotError> {
         let markup = e.content();
         let svg = PlantumlMarkup::new(&markup).as_svg()?;
-        Ok(svg_to_element(svg))
+        Ok(svg_to_element(svg, "UML diagram"))
     }
 
     fn typeset_pikchr(e: &Element) -> Result<Element, SubplotError> {
         let markup = e.content();
         let svg = PikchrMarkup::new(&markup, None).as_svg()?;
-        Ok(svg_to_element(svg))
+        Ok(svg_to_element(svg, "Pikchr diagram"))
     }
 
     fn typeset_roadmap(e: &Element) -> Result<Element, SubplotError> {
@@ -400,12 +404,12 @@ mod typeset {
         let roadmap = roadmap::from_yaml(&yaml)?;
         let dot = roadmap.format_as_dot(WIDTH)?;
         let svg = DotMarkup::new(&dot).as_svg()?;
-        Ok(svg_to_element(svg))
+        Ok(svg_to_element(svg, "Road map"))
     }
 
-    fn svg_to_element(svg: Svg) -> Element {
+    fn svg_to_element(svg: Svg, alt: &str) -> Element {
         let url = svg_as_data_url(svg);
-        let img = html_img(&url);
+        let img = html_img(&url, alt);
         html_p(vec![Content::Elt(img)])
     }
 
@@ -422,9 +426,10 @@ mod typeset {
         new
     }
 
-    fn html_img(src: &str) -> Element {
+    fn html_img(src: &str, alt: &str) -> Element {
         let mut new = Element::new(ElementTag::Img);
         new.push_attribute(Attribute::new("src", src));
+        new.push_attribute(Attribute::new("alt", alt));
         new
     }
 }
author	Daniel Silverstone <dsilvers+gitlab@digital-scurf.org>	2023-10-10 07:08:21 +0000
committer	Daniel Silverstone <dsilvers+gitlab@digital-scurf.org>	2023-10-10 07:08:21 +0000
commit	58350fed3216b3ec2875503f6036e1a7f982088e (patch)
tree	9fb29d0b107d5e1edb78954e68e35270b5fc391c
parent	478147c59de2547cb1e3e0a0cd5309c25f4453dc (diff)
parent	76838e218f8c0943de43befce434fa8845a68e86 (diff)
download	subplot-58350fed3216b3ec2875503f6036e1a7f982088e.tar.gz