datetime: primary and aliases zones in Olson

* localtime.c refactored to make tzalloc/tzfree public;
* Modified Perl script to generate primary Olson timezones and aliases.
  Get rid of NYI;

NO_DOC=refactoring
NO_CHANGELOG=refactoring
NO_TEST=refactoring

src/lib/tzcode/CMakeLists.txt

@@ -4,8 +4,10 @@ target_link_libraries(tzcode)
 
 set(TZCODE_PROJECT tzcode-iana)
 set(TZCODE_BUILD_DIR ${CMAKE_CURRENT_BINARY_DIR})
-set(TZCODE_PREFIX ${TZCODE_BUILD_DIR}/tz)
+set(TZCODE_PREFIX_DIR tz)
+set(TZCODE_PREFIX ${TZCODE_BUILD_DIR}/${TZCODE_PREFIX_DIR})
 set(TZDATA_TARGET main.zi)
+set(TZCODE_SRC ${PROJECT_SOURCE_DIR}/src/lib/tzcode)
 
 set(TZDATA_FULL_TARGET
     ${TZCODE_PREFIX}/src/${TZCODE_PROJECT}/${TZDATA_TARGET})
@@ -15,7 +17,9 @@ ExternalProject_Add(${TZCODE_PROJECT}
     GIT_TAG 2022a
     GIT_SHALLOW YES
 
+    EXCLUDE_FROM_ALL 1
     INSTALL_DIR ${TZCODE_BUILD_DIR}/install
+    PREFIX ${TZCODE_PREFIX_DIR}
 
     UPDATE_COMMAND ""
     CONFIGURE_COMMAND ""
@@ -30,3 +34,21 @@ ExternalProject_Add(${TZCODE_PROJECT}
     INSTALL_COMMAND make DESTDIR=<INSTALL_DIR> install
     BUILD_BYPRODUCTS ${TZDATA_FULL_TARGET}
 )
+
+find_program(PERL perl)
+
+add_custom_target(gen-timezones DEPENDS ${TZCODE_PROJECT})
+add_custom_command(TARGET gen-timezones
+    COMMAND
+    ${PERL} ${TZCODE_SRC}/gen-zone-abbrevs.pl ${TZCODE_SRC}/zone-abbrevs.txt
+    ${TZDATA_FULL_TARGET} > ${TZCODE_SRC}/timezones.h
+    COMMENT "Generating timezones.h"
+)
+
+# We excluded build of gen-timezones and tzcode-iana (main.zi) targets
+# from ALL dependencies (see EXCLUDE_FROM_ALL=1) if you, by either reason
+# need to regenerate timezones.h then invoke target manually:
+# ```sh
+#   make gen-timezones VERBOSE=1
+# ```
+# it will regenerate `timezones.h` in-place, in their repository location.

src/lib/tzcode/gen-zone-abbrevs.pl

@@ -129,6 +129,7 @@ my $EntryRx = do {
 };
 
 my %ZoneAbbrevs;
+my %ZoneAbbrevNames;
 
 # read zone-abbrevs.txt with definition of all currently known
 # timezone abbreviations.
@@ -176,36 +177,43 @@ sub read_abbrevs_file($) {
         }
 
         $ZoneAbbrevs{$encoded} = [ $flags, $offset, $name ];
+        $ZoneAbbrevNames{$name}++;
     }
 }
 
 my %ZoneNames;
+my %ZoneIds;
+my %Aliases;
 
 sub read_zi_file($) {
     my $filename = shift;
     open( my $fh, '<:raw', $filename )
       or die qq<Could not open '$filename' for reading: '$!'>;
 
-    my $zone;
-    my $format;
+
     while (<$fh>) {
         next if /^#/;
         if (/^Zone\s+([A-Za-z\-\_\/]+)\s+/) {
-            $zone   = $1;
-            $ZoneNames{$zone}++ if $zone =~ q|/|;
-            next;
+            my $zone = $1;
+            next unless $zone =~ q|/|;
+            $ZoneNames{$zone}++;
+        } elsif (/^[#]?Link\s+(\S+)\s+(\S+)/) {
+            my $zone = $1;
+            my $link = $2;
+            # do not create alias if it's known abbreviation
+            next if defined $ZoneAbbrevNames{$link};
+            $Aliases{$link} = $zone;
         }
     }
 
 }
 
-sub gen_header() {
-    printf "/* Automatically generated by gen-zone-abbrevs.pl */\n";
-
-    # first we generate abbreviations list
+# first we generate abbreviations list
+sub gen_abbrevs() {
     foreach my $encoded ( sort { $a <=> $b } keys %ZoneAbbrevs ) {
         my ( $flags, $offset, $name ) = @{ $ZoneAbbrevs{$encoded} };
         my $nextid = next_id($name);
+        $ZoneIds{$nextid} = $name;
 
         if ( $flags != 0 ) {
             my @names;
@@ -221,14 +229,34 @@ sub gen_header() {
         printf "ZONE_ABBREV(%4d, %4d, \"%s\", %s)\n",
           $nextid, $offset, $name, $flags;
     }
+}
 
-    # second we enumerate all known from main.zi primary zone names
+# second we enumerate all known from main.zi primary zone names
+sub gen_primary_zones() {
     foreach my $zonename (sort keys %ZoneNames) {
         my $nextid = next_id($zonename);
+        $ZoneIds{$zonename} = $nextid;
         printf "ZONE_UNIQUE(%4d, \"%s\")\n", $nextid, $zonename;
     }
 }
 
+# third we enumerate all aliases for primary zones
+sub gen_aliases() {
+    foreach my $alias ( sort keys %Aliases ) {
+        my $zonename = $Aliases{$alias};
+        printf "ZONE_ALIAS(%4d, \"%s\", \"%s\")\n", $ZoneIds{$zonename},
+            $alias, $zonename;
+    }
+}
+
+sub gen_header() {
+    printf "/* Automatically generated by gen-zone-abbrevs.pl */\n";
+
+    gen_abbrevs();
+    gen_primary_zones();
+    gen_aliases();
+}
+
 read_abbrevs_file($zone_abbrevs_file);
 read_zi_file($tzdata_zi_file);
 gen_header();

src/lib/tzcode/localtime.c

@@ -15,8 +15,9 @@
 #include <unistd.h>
 #define LOCALTIME_IMPLEMENTATION
 #include "private.h"
-
 #include "tzfile.h"
+#include "tzcode.h"
+
 #include <fcntl.h>
 
 #ifndef TZ_ABBR_MAX_LEN
@@ -66,10 +67,10 @@
 #define WILDABBR "   "
 #endif /* !defined WILDABBR */
 
-static const char wildabbr[] = WILDABBR;
-
 static const char gmt[] = "GMT";
 
+static const char wildabbr[] = WILDABBR;
+
 /*
 ** The DST rules to use if TZ has no rules and we can't load TZDEFRULES.
 ** Default to US rules as of 2017-05-07.

src/lib/tzcode/timezone.c

@@ -31,7 +31,9 @@ timezone_flags(const struct date_time_zone *zone)
 static const char * zone_names[] = {
 #define ZONE_ABBREV(id, offset, name, flags) [id] = name,
 #define ZONE_UNIQUE(id, name) [id] = name,
+#define ZONE_ALIAS(id, alias, name)
 #include "timezones.h"
+#undef ZONE_ALIAS
 #undef ZONE_UNIQUE
 #undef ZONE_ABBREV
 };
@@ -45,9 +47,10 @@ timezone_name(int64_t index)
 
 static struct date_time_zone zone_abbrevs[] = {
 #define ZONE_ABBREV(id, offset, name, flags) { name, id, flags, offset },
-#define ZONE_UNIQUE(id, name) { name, id, TZ_NYI, 0 },
-
+#define ZONE_UNIQUE(id, name) { name, id, TZ_OLSON, 0 },
+#define ZONE_ALIAS(id, alias, name) { alias, id, TZ_OLSON|TZ_ALIAS, 0 },
 #include "timezones.h"
+#undef ZONE_ALIAS
 #undef ZONE_UNIQUE
 #undef ZONE_ABBREV
 };

src/lib/tzcode/timezone.h

@@ -17,6 +17,8 @@ enum {
 	TZ_MILITARY = 0x04,
 	TZ_AMBIGUOUS = 0x08,
 	TZ_NYI = 0x10,
+	TZ_OLSON = 0x20,
+	TZ_ALIAS = 0x40,
 };
 
 /**

src/lib/tzcode/tzcode.h

@@ -66,6 +66,13 @@ char *
 tnt_strptime(const char *__restrict buf, const char *__restrict fmt,
 	     struct tnt_tm *__restrict tm);
 
+typedef struct state *timezone_t;
+
+/** Find installed timezone definition and load it to memory */
+timezone_t tzalloc(const char *);
+/** Free loaded timezone definition */
+void tzfree(timezone_t);
+
 #if defined(__cplusplus)
 } /* extern "C" */
 #endif /* defined(__cplusplus) */