Fixes #13095: Correctly generate nav in the reference doc

src/reference/Makefile

@@ -17,7 +17,7 @@ content: $(GRAPHVIZ_IMAGES) $(ADOC_DYN_FILES) modules/ROOT/nav.adoc
 	cp dependencies/$@ modules/ROOT/pages/_partials/dyn/$@
 
 modules/ROOT/nav.adoc: $(ADOC_SRC_FILES)
-	./tools/generate-nav.sh
+	./tools/generate-nav.py > modules/ROOT/nav.adoc
 
 modules/ROOT/assets/images/graphviz/%.png: modules/ROOT/assets/graphviz/%.dot
 	mkdir -p modules/ROOT/assets/images/graphviz

src/reference/tools/generate-nav.py

@@ -0,0 +1,56 @@
+#!/usr/bin/env python
+
+import os
+import re
+
+PAGESDIR = "modules/ROOT/pages/"
+TITLE = re.compile("^(=+) (.+)$")
+ID = re.compile(r"^[[(.+),?.*]]$")
+
+def remove_duplicate_underscore(string):
+    prev = ' '
+    res = []
+    for char in string:
+        if prev == '_' and char == '_':
+            pass
+        else:
+            res.append(char)
+        prev = char
+    return ''.join(res)
+
+# reproduce asciidoc's behavior
+def slugify(s):
+    s = "_" + s
+    s = s.lower()
+    s = s.strip()
+    s = re.sub('\W', '_', s)
+    s = remove_duplicate_underscore(s)
+    return s
+
+os.chdir(PAGESDIR)
+
+# Get all standalone .adoc pages, sorted alphanumerically
+# We exclude files in root (the index of the doc), _partials which are not actual pages
+files = sorted([root.split('/', 1)[-1]+"/"+file for root, dirs, files in os.walk('.') for file in files if file.endswith(".adoc") and not "_partials" in root and not root == "."])
+
+result = ["// Automatically generated list of content - do not edit"]
+
+for file in files:
+    with open(file) as f:
+        content = f.read().splitlines()
+
+    prev = ""
+    for line in content:
+        search_title = TITLE.search(line)
+        if search_title:
+            level = search_title.group(1).count("=")
+            title = search_title.group(2)
+            search_id = ID.search(prev)
+            if search_id:
+                page_id = search_id.group(1)
+            else:
+                page_id = slugify(title)
+            result.append("*" * level + " xref:" + file + "#" + page_id + "[" + title + "]")
+        prev = line
+
+print("\n".join(result))