Avoid duplicate anchors (#159)

Fixes #158 Signed-off-by: Dave Thaler <[email protected]>
dthaler · Feb 29, 2024 · 9ced488 · 9ced488
1 parent e8f151d
commit 9ced488
Show file tree

Hide file tree

Showing 3 changed files with 197 additions and 28 deletions.
diff --git a/lib/rst2rfcxml.cpp b/lib/rst2rfcxml.cpp
@@ -303,6 +303,33 @@ _replace_all(string line, string from, string to)
     return line;
 }
 
+string
+rst2rfcxml::define_anchor(string value)
+{
+    string anchor;
+    if (_anchors.find(value) == _anchors.end()) {
+        // Create a new anchor.
+        anchor = _anchor(value);
+    } else {
+        // This is a duplicate anchor definition so create a new one and
+        // map all future lookups to this latest one.
+        anchor = _anchors[value] + "-";
+    }
+    _anchors[value] = anchor;
+    return anchor;
+ }
+
+string
+rst2rfcxml::lookup_anchor(string value)
+{
+    if (_anchors.find(value) != _anchors.end()) {
+        return _anchors[value];
+    }
+
+    // Undefined anchor.
+    return _anchor(value);
+}
+
 string
 rst2rfcxml::replace_term_links(string line)
 {
@@ -325,7 +352,7 @@ rst2rfcxml::replace_term_links(string line)
             term = middle.substr(label_end + 4, term_end - label_end - 4);
         }
 
-        line = fmt::format("{}<xref target=\"term-{}\">{}</xref>{}", before, _anchor(term), label, after);
+        line = fmt::format("{}<xref target=\"{}\">{}</xref>{}", before, lookup_anchor("term-" + term), label, after);
     }
     return line;
 }
@@ -425,7 +452,7 @@ rst2rfcxml::replace_reference_links(string line)
             }
         }
 
-        line = fmt::format("{}<xref target=\"{}\">{}</xref>{}", before, _anchor(middle), middle, after);
+        line = fmt::format("{}<xref target=\"{}\">{}</xref>{}", before, lookup_anchor(middle), middle, after);
     }
     return line;
 }
@@ -802,11 +829,14 @@ rst2rfcxml::handle_section_title(int level, string marker, string current, strin
             push_context(output_stream, xml_context::MIDDLE);
         }
         string title = handle_escapes_and_links(current);
-        push_context(
-            output_stream,
-            xml_context::SECTION,
-            current_indentation,
-            fmt::format("anchor=\"{}\" title=\"{}\"", _anchor(title), title));
+        string anchor = define_anchor(title);
+        string attributes;
+        if (anchor.empty()) {
+            attributes = fmt::format("title=\"{}\"", title);
+        } else {
+            attributes = fmt::format("anchor=\"{}\" title=\"{}\"", anchor, title);
+        }
+        push_context(output_stream, xml_context::SECTION, current_indentation, attributes);
         return true;
     }
     if (current.starts_with(marker) && current.find_first_not_of(marker, 0) == string::npos &&
@@ -972,8 +1002,13 @@ rst2rfcxml::process_line(string current, string next, ostream& output_stream)
         if (!in_context(xml_context::DEFINITION_LIST)) {
             push_context(output_stream, xml_context::DEFINITION_LIST, current_indentation);
         }
-        string attributes = fmt::format("anchor=\"term-{}\"", _anchor(_trim(current)));
-        push_context(output_stream, xml_context::DEFINITION_TERM, current_indentation, attributes);
+        string anchor = define_anchor("term-" + _trim(current));
+        if (anchor.empty()) {
+            push_context(output_stream, xml_context::DEFINITION_TERM, current_indentation);
+        } else {
+            string attributes = fmt::format("anchor=\"{}\"", anchor);
+            push_context(output_stream, xml_context::DEFINITION_TERM, current_indentation, attributes);
+        }
     }
 
     // Handle artwork.

diff --git a/lib/rst2rfcxml.h b/lib/rst2rfcxml.h
@@ -131,6 +131,10 @@ class rst2rfcxml
     std::string
     replace_term_links(std::string line);
     std::string
+    define_anchor(std::string term);
+    std::string
+    lookup_anchor(std::string term);
+    std::string
     handle_escapes_and_links(std::string line);
     void
     output_table_row(std::ostream& output_stream);
@@ -140,6 +144,7 @@ class rst2rfcxml
     std::string _ipr;
     std::string _category;
     std::vector<size_t> _column_indices;
+    std::map<std::string, std::string> _anchors;
     std::map<std::string, author> _authors;
     std::string _submission_type;
     std::string _abbreviated_title;

diff --git a/test/basic_tests.cpp b/test/basic_tests.cpp
@@ -9,6 +9,25 @@
 
 using namespace std;
 
+constexpr const char* BASIC_PREAMBLE = R"(<?xml version="1.0" encoding="UTF-8"?>
+  <?xml-stylesheet type="text/xsl" href="rfc2629.xslt"?>
+  <!-- generated by https://github.com/dthaler/rst2rfcxml version 0.1 -->
+
+<!DOCTYPE rfc [
+]>
+
+<?rfc rfcedstyle="yes"?>
+<?rfc toc="yes"?>
+<?rfc tocindent="yes"?>
+<?rfc sortrefs="yes"?>
+<?rfc symrefs="yes"?>
+<?rfc strict="yes"?>
+<?rfc comments="yes"?>
+<?rfc inline="yes"?>
+<?rfc text-list-symbols="-o*+"?>
+<?rfc docmapping="yes"?>
+)";
+
 void
 test_rst2rfcxml(const char* input, const char* expected_output)
 {
@@ -99,6 +118,40 @@ TEST_CASE("titles", "[basic]")
     test_rst2rfcxml("Foo\n~~~\n", "<section anchor=\"foo\" title=\"Foo\">\n</section>\n");
 }
 
+TEST_CASE("duplicate titles", "[basic]")
+{
+    test_rst2rfcxml(
+        R"(.. header::
+
+foo
+===
+
+bar
+
+foo
+===
+
+baz
+)", (std::string(BASIC_PREAMBLE) + R"(
+<rfc ipr="" docName="" category="" submissionType="">
+ <front>
+ </front>
+ <middle>
+  <section anchor="foo" title="foo">
+   <t>
+    bar
+   </t>
+  </section>
+  <section anchor="foo-" title="foo">
+   <t>
+    baz
+   </t>
+  </section>
+ </middle>
+</rfc>
+)").c_str());
+}
+
 TEST_CASE("line block", "[basic]")
 {
     test_rst2rfcxml(
@@ -315,6 +368,101 @@ baz :term:`foo-ish<foo>` baz
 )");
 }
 
+TEST_CASE("definition list duplication", "[basic]")
+{
+    test_rst2rfcxml(
+        R"(
+foo
+  description
+
+bar
+  description
+
+baz :term:`foo-ish<foo>` baz
+
+foo
+  description 2
+
+bar
+  description 2
+
+baz :term:`foo-ish<foo>` baz
+
+foo
+  description 3
+
+bar
+  description 3
+
+baz :term:`foo-ish<foo>` baz
+
+)",
+        R"(<dl>
+ <dt anchor="term-foo">
+  foo
+ </dt>
+ <dd>
+  <t>
+   description
+  </t>
+ </dd>
+ <dt anchor="term-bar">
+  bar
+ </dt>
+ <dd>
+  <t>
+   description
+  </t>
+ </dd>
+</dl>
+<t>
+ baz <xref target="term-foo">foo-ish</xref> baz
+</t>
+<dl>
+ <dt anchor="term-foo-">
+  foo
+ </dt>
+ <dd>
+  <t>
+   description 2
+  </t>
+ </dd>
+ <dt anchor="term-bar-">
+  bar
+ </dt>
+ <dd>
+  <t>
+   description 2
+  </t>
+ </dd>
+</dl>
+<t>
+ baz <xref target="term-foo-">foo-ish</xref> baz
+</t>
+<dl>
+ <dt anchor="term-foo--">
+  foo
+ </dt>
+ <dd>
+  <t>
+   description 3
+  </t>
+ </dd>
+ <dt anchor="term-bar--">
+  bar
+ </dt>
+ <dd>
+  <t>
+   description 3
+  </t>
+ </dd>
+</dl>
+<t>
+ baz <xref target="term-foo--">foo-ish</xref> baz
+</t>
+)");
+}
+
 TEST_CASE("definition list with glossary label", "[basic]")
 {
     test_rst2rfcxml(
@@ -775,25 +923,6 @@ Paragraph two.
 )");
 }
 
-constexpr const char* BASIC_PREAMBLE = R"(<?xml version="1.0" encoding="UTF-8"?>
-  <?xml-stylesheet type="text/xsl" href="rfc2629.xslt"?>
-  <!-- generated by https://github.com/dthaler/rst2rfcxml version 0.1 -->
-
-<!DOCTYPE rfc [
-]>
-
-<?rfc rfcedstyle="yes"?>
-<?rfc toc="yes"?>
-<?rfc tocindent="yes"?>
-<?rfc sortrefs="yes"?>
-<?rfc symrefs="yes"?>
-<?rfc strict="yes"?>
-<?rfc comments="yes"?>
-<?rfc inline="yes"?>
-<?rfc text-list-symbols="-o*+"?>
-<?rfc docmapping="yes"?>
-)";
-
 TEST_CASE("empty header", "[basic]")
 {
     string expected_output = BASIC_PREAMBLE;