From 8a4186a5b5bffbfed254f5e7f59d3aeb14394452 Mon Sep 17 00:00:00 2001
From: Thibaut Lienart <tlienart@me.com>
Date: Wed, 29 May 2019 17:02:33 +1000
Subject: [PATCH] closes #130

---
 src/converter/lx.jl        |  2 +-
 src/converter/md_utils.jl  |  7 ++++---
 src/misc_utils.jl          | 12 ++++++++++--
 test/converter/markdown.jl | 15 +++++++++++++++
 test/misc.jl               | 11 +++++++++++
 test/runtests.jl           |  2 +-
 6 files changed, 42 insertions(+), 7 deletions(-)

diff --git a/src/converter/lx.jl b/src/converter/lx.jl
index 1ca9ac32e..5fb43a87a 100644
--- a/src/converter/lx.jl
+++ b/src/converter/lx.jl
@@ -57,7 +57,7 @@ JD_LOC_EQDICT_COUNTER
 Counter to keep track of equation numbers as they appear along the page, this helps with equation
 referencing.
 """
-const JD_LOC_EQDICT_COUNTER = "COUNTER_" * randstring(JD_LEN_RANDSTRING)
+const JD_LOC_EQDICT_COUNTER = "COUNTER_XC0q"
 
 """
 $(SIGNATURES)
diff --git a/src/converter/md_utils.jl b/src/converter/md_utils.jl
index 9004695ac..14a09ee32 100644
--- a/src/converter/md_utils.jl
+++ b/src/converter/md_utils.jl
@@ -10,7 +10,7 @@ function md2html(ss::AbstractString, stripp::Bool=false)::AbstractString
     isempty(ss) && return ss
 
     # Use the base Markdown -> Html converter and post process headers
-    partial = ss |> Markdown.parse |> Markdown.html |> make_headers_links
+    partial = ss |> Markdown.parse |> Markdown.html |> make_header_refs
 
     # In some cases, base converter adds <p>...</p>\n which we might not want
     stripp || return partial
@@ -56,14 +56,15 @@ By default the Base Markdown to HTML converter simply converts `## ...` into hea
 linkable ones; this is annoying for generation of table of contents etc (and references in
 general) so this function does just that.
 """
-function header_ref(h::String)::String
+function make_header_refs(h::String)::String
     io = IOBuffer()
     head = 1
     for m ∈ eachmatch(r"<h([1-6])>(.*?)</h[1-6]>", h)
         write(io, subs(h, head:m.offset-1))
         level = m.captures[1]
         name  = m.captures[2]
-        write(io, "<h$(level)><a id=\"$(refstring(name))\"></a>$(name)</h$(level)>")
+        ref   = refstring(name)
+        write(io, "<h$level><a id=\"$ref\" href=\"#$ref\">$name</a></h$level>")
         head = m.offset + lastindex(m.match)
     end
     write(io, subs(h, head:lastindex(h)))
diff --git a/src/misc_utils.jl b/src/misc_utils.jl
index c616a6828..74cb9c370 100644
--- a/src/misc_utils.jl
+++ b/src/misc_utils.jl
@@ -125,5 +125,13 @@ $(SIGNATURES)
 
 Takes a string `s` and replace spaces by underscores so that that we can use it
 for hyper-references. So for instance `"aa  bb"` will become `aa-bb`.
-"""
-refstring(s::AbstractString)::String = replace(lowercase(strip(s)), r"\s+" => "-")
+It also defensively removes any non-word character so for instance `"aa bb !"` will be `"aa-bb"`
+"""
+function refstring(s::AbstractString)::String
+    # remove non-word characters
+    st = replace(s, r"&#[0-9]+;" => "")
+    st = replace(st, r"[^a-zA-Z0-9_\-\s]" => "")
+    st = replace(lowercase(strip(st)), r"\s+" => "-")
+    isempty(st) && return string(hash(s))
+    return st
+end
diff --git a/test/converter/markdown.jl b/test/converter/markdown.jl
index 4d7ba7f34..752ff4064 100644
--- a/test/converter/markdown.jl
+++ b/test/converter/markdown.jl
@@ -111,3 +111,18 @@ end
     hstring = J.convert_inter_html(inter_html, blocks2insert, lxcontext)
     @test hstring == "<p>text A1 text A2 blah and \nescape B1\n  text C1 \\(\\mathrm{ b}\\) text C2  then part1: AA and part2: BB.</p>\n"
 end
+
+
+@testset "headers" begin
+    st = """
+        # Title
+        and then
+        ## Subtitle cool!
+        done
+        """
+    h = st |> Markdown.parse |> Markdown.html
+    r = J.make_header_refs(h)
+    @test occursin("<h1><a id=\"title\" href=\"#title\">Title</a></h1>", r)
+    @test occursin("<h2><a id=\"subtitle-cool\" href=\"#subtitle-cool\">Subtitle cool&#33;</a></h2>", r)
+    @test occursin("<p>done</p>", r)
+end
diff --git a/test/misc.jl b/test/misc.jl
index ef5856882..d1eb2a263 100644
--- a/test/misc.jl
+++ b/test/misc.jl
@@ -74,3 +74,14 @@ end
     m = match(r"\[done\s*(.*?)ms\]", r)
     @test parse(Float64, m.captures[1]) ≥ 500
 end
+
+
+@testset "refstring" begin
+    @test J.refstring("aa  bb") == "aa-bb"
+    @test J.refstring("aa  bb !") == "aa-bb"
+    @test J.refstring("aa-bb-!") == "aa-bb-"
+    @test J.refstring("aa 🔺 bb") == "aa-bb"
+    @test J.refstring("aaa 0 bb s:2  df") == "aaa-0-bb-s2-df"
+    @test J.refstring("🔺🔺") == string(hash("🔺🔺"))
+    @test J.refstring("blah&#33;") == "blah"
+end
diff --git a/test/runtests.jl b/test/runtests.jl
index 276a7477b..5eaab2c79 100644
--- a/test/runtests.jl
+++ b/test/runtests.jl
@@ -1,4 +1,4 @@
-using JuDoc, Test
+using JuDoc, Test, Markdown
 const J = JuDoc
 const D = joinpath(dirname(dirname(pathof(JuDoc))), "test", "_dummies")