Fix for #326, avoid double escaping when pre-rendering a html code bl…

…ock (#328)

* closes #326

Project.toml

@@ -1,7 +1,7 @@
 name = "JuDoc"
 uuid = "4ca9428c-4c75-11e9-2efb-bf5cb6c1e8f8"
 authors = ["Thibaut Lienart <[email protected]>"]
-version = "0.4"
+version = "0.4.1"
 
 [deps]
 Dates = "ade2ca70-3891-5945-98fb-dc099432e06a"

src/converter/js_prerender.jl

@@ -63,11 +63,16 @@ function js_prerender_highlight(hs::String)::String
         # core code
         cs = subs(hs, nextind(hs, matchrange(co).stop), prevind(hs, matchrange(cc).start))
         cs = escape_string(cs)
-
+        # lang
         lang = co.captures[2]
         if isnothing(lang)
             write(jsbuffer, """console.log("<pre><code class=hljs>$cs</code></pre>");\n""")
         else
+            if lang == "html" && is_html_escaped(cs)
+                # corner case, see https://github.com/tlienart/JuDoc.jl/issues/326
+                # highlight will re-escape the string further.
+                cs = html_unescape(cs) |> escape_string
+            end
             # add to content of jsbuffer
             write(jsbuffer, """console.log("<pre><code class=\\"$lang hljs\\">" + hljs.highlight("$lang", "$cs").value + "</code></pre>");""")
         end
@@ -113,3 +118,27 @@ function js2html(hs::String, jsbuffer::IOBuffer, matches::Vector{RegexMatch},
 
     return String(take!(htmls))
 end
+
+
+# Copied from https://github.com/JuliaLang/julia/blob/acb7bd93fb2d5adbbabeaed9f39ab3c85495b02f/stdlib/Markdown/src/render/html.jl#L25-L31
+const _htmlescape_chars = LittleDict(
+            '<' => "&lt;",
+            '>' => "&gt;",
+            '"' => "&quot;",
+            '&' => "&amp;"
+            )
+for ch in "'`!\$%()=+{}[]"
+    _htmlescape_chars[ch] = "&#$(Int(ch));"
+end
+
+const _htmlesc_to = values(_htmlescape_chars) |> collect
+
+is_html_escaped(cs::String) = !isnothing(findfirst(ss -> occursin(ss, cs), _htmlesc_to))
+
+function html_unescape(cs::String)
+    # this is a bit inefficient but whatever, `cs` shouldn't  be very long.
+    for (ssfrom, ssto) in _htmlescape_chars
+        cs = replace(cs, ssto => ssfrom)
+    end
+    return cs
+end

test/global/html_esc.jl

@@ -0,0 +1,8 @@
+# See https://github.com/tlienart/JuDoc.jl/issues/326
+
+@testset "Issue 326" begin
+    h1 = "<div class=\"hello\">Blah</div>"
+    h1e = Markdown.htmlesc(h1)
+    @test J.is_html_escaped(h1e)
+    @test J.html_unescape(h1e) == h1
+end

test/runtests.jl

@@ -49,6 +49,7 @@ println("GLOBAL")
 include("global/cases1.jl")
 include("global/cases2.jl")
 include("global/ordering.jl")
+include("global/html_esc.jl")
 
 begin
     # create temp dir to do complete integration testing (has to be here in order