[libcxx-commits] [libcxx] [libc++][TZDB] Fixes parsing interleaved rules. (PR #84808)

[Mark de Wever via libcxx-commits]

https://github.com/mordante updated https://github.com/llvm/llvm-project/pull/84808

>From 557788da3adf697a0162c1ae2e7cc186a3d6eb23 Mon Sep 17 00:00:00 2001
From: Mark de Wever <koraq at xs4all.nl>
Date: Mon, 11 Mar 2024 19:15:30 +0100
Subject: [PATCH 1/2] [libc++][TZDB] Fixes parsing interleaved rules.

Typically the rules in the database are contiguous, but that is not a
requirement. This fixes the case when they are not.
---
 libcxx/src/tzdb.cpp                           | 27 ++++++++++++++++---
 .../time.zone/time.zone.db/rules.pass.cpp     | 24 +++++++++++++++++
 2 files changed, 47 insertions(+), 4 deletions(-)

diff --git a/libcxx/src/tzdb.cpp b/libcxx/src/tzdb.cpp
index 2bb801e48694ee..49985b870d8525 100644
--- a/libcxx/src/tzdb.cpp
+++ b/libcxx/src/tzdb.cpp
@@ -511,14 +511,33 @@ static string __parse_version(istream& __input) {
   return chrono::__parse_string(__input);
 }
 
+[[nodiscard]]
+static __tz::__rule& __create_entry(__tz::__rules_storage_type& __rules, const string& __name) {
+  auto __result = [&]() -> __tz::__rule& {
+    auto& __rule = __rules.emplace_back(__name, vector<__tz::__rule>{});
+    return __rule.second.emplace_back();
+  };
+
+  if (__rules.empty())
+    return __result();
+
+  // Typically rules are in contiguous order in the database.
+  // But there are exceptions. For example, NZ and Chatham are interleaved.
+  if (__rules.back().first == __name)
+    return __rules.back().second.emplace_back();
+
+  if (auto __it = ranges::find(__rules, __name, [](const auto& __r) { return __r.first; });
+      __it != ranges::end(__rules))
+    return __it->second.emplace_back();
+
+  return __result();
+}
+
 static void __parse_rule(tzdb& __tzdb, __tz::__rules_storage_type& __rules, istream& __input) {
   chrono::__skip_mandatory_whitespace(__input);
   string __name = chrono::__parse_string(__input);
 
-  if (__rules.empty() || __rules.back().first != __name)
-    __rules.emplace_back(__name, vector<__tz::__rule>{});
-
-  __tz::__rule& __rule = __rules.back().second.emplace_back();
+  __tz::__rule& __rule = __create_entry(__rules, __name);
 
   chrono::__skip_mandatory_whitespace(__input);
   __rule.__from = chrono::__parse_year(__input);
diff --git a/libcxx/test/libcxx/time/time.zone/time.zone.db/rules.pass.cpp b/libcxx/test/libcxx/time/time.zone/time.zone.db/rules.pass.cpp
index 5ae2ed1e91ebe5..4814f4aad87fbc 100644
--- a/libcxx/test/libcxx/time/time.zone/time.zone.db/rules.pass.cpp
+++ b/libcxx/test/libcxx/time/time.zone/time.zone.db/rules.pass.cpp
@@ -550,6 +550,29 @@ R a 0 1 - Ja Su>=31 1w 2s abc
   assert(result.rules[0].second[2].__letters == "abc");
 }
 
+static void test_mixed_order() {
+  // This is a part of the real database. The interesting part is that the
+  // rules NZ and Chatham are interleaved. Make sure the parse algorithm
+  // handles this correctly.
+  parse_result result{
+      R"(
+# Since 1957 Chatham has been 45 minutes ahead of NZ, but until 2018a
+# there was no documented single notation for the date and time of this
+# transition.  Duplicate the Rule lines for now, to give the 2018a change
+# time to percolate out.
+Rule NZ  1974    only    -   Nov Sun>=1  2:00s   1:00    D
+Rule Chatham 1974    only    -   Nov Sun>=1  2:45s   1:00    -
+Rule NZ  1975    only    -   Feb lastSun 2:00s   0   S
+Rule Chatham 1975    only    -   Feb lastSun 2:45s   0   -
+Rule NZ  1975    1988    -   Oct lastSun 2:00s   1:00    D
+Rule Chatham 1975    1988    -   Oct lastSun 2:45s   1:00    -
+)"};
+
+  assert(result.rules.size() == 2);
+  assert(result.rules[0].second.size() == 3);
+  assert(result.rules[1].second.size() == 3);
+}
+
 int main(int, const char**) {
   test_invalid();
   test_name();
@@ -560,6 +583,7 @@ int main(int, const char**) {
   test_at();
   test_save();
   test_letter();
+  test_mixed_order();
 
   return 0;
 }

>From eac58f6cacea5639b8a938b9d603203a275449ff Mon Sep 17 00:00:00 2001
From: Mark de Wever <zar-rpg at xs4all.nl>
Date: Tue, 12 Mar 2024 17:26:25 +0100
Subject: [PATCH 2/2] Update libcxx/src/tzdb.cpp

Co-authored-by: Louis Dionne <ldionne.2 at gmail.com>
---
 libcxx/src/tzdb.cpp | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/libcxx/src/tzdb.cpp b/libcxx/src/tzdb.cpp
index 49985b870d8525..0307f754caabf3 100644
--- a/libcxx/src/tzdb.cpp
+++ b/libcxx/src/tzdb.cpp
@@ -522,7 +522,7 @@ static __tz::__rule& __create_entry(__tz::__rules_storage_type& __rules, const s
     return __result();
 
   // Typically rules are in contiguous order in the database.
-  // But there are exceptions. For example, NZ and Chatham are interleaved.
+  // But there are exceptions, some rules are interleaved.
   if (__rules.back().first == __name)
     return __rules.back().second.emplace_back();