From ed0fd9937fa5f77494c728827226a94d4295f670 Mon Sep 17 00:00:00 2001
From: "Steinar H. Gunderson" <sgunderson@bigfoot.com>
Date: Mon, 6 Nov 2023 23:32:39 +0100
Subject: [PATCH] Use diacritic-ignoring, proper UCA comparison.

---
 bin/sync.pl | 27 +++++++++++++++++----------
 1 file changed, 17 insertions(+), 10 deletions(-)

diff --git a/bin/sync.pl b/bin/sync.pl
index ef07ee7..4351817 100644
--- a/bin/sync.pl
+++ b/bin/sync.pl
@@ -9,6 +9,7 @@ use DBI;
 use POSIX;
 use Time::HiRes;
 use IO::Select;
+use Unicode::Collate;
 binmode STDOUT, ':utf8';
 binmode STDERR, ':utf8';
 use utf8;
@@ -16,6 +17,7 @@ use utf8;
 require '../include/config.pm';
 
 my @log = ();
+my $uca = Unicode::Collate->new(level => 1);
 
 my %rgb = (
 	yellow => {
@@ -44,6 +46,11 @@ sub log_timing {
 	printf "%s: %.0f ms.\n", $msg, 1e3 * $elapsed;
 }
 
+sub sort_key {
+	my $m = shift;
+	return $uca->getSortKey($m);
+}
+
 sub get_oauth_bearer_token {
 	my ($dbh, $ua) = @_;
 	my $now = time();
@@ -107,18 +114,18 @@ sub get_spreadsheet_name {
 
 sub matches_name {
 	my ($slack_name, $spreadsheet_name) = @_;
-	if (lc($slack_name) eq lc($spreadsheet_name)) {
+	if (sort_key($slack_name) eq sort_key($spreadsheet_name)) {
 		return 1;
 	}
 
 	my @ap = split /\s+/, $slack_name;
 	my @bp = split /\s+/, $spreadsheet_name;
-	if (scalar @ap >= 2 && scalar @bp >= 2 && lc($ap[0]) eq lc($bp[0])) {
+	if (scalar @ap >= 2 && scalar @bp >= 2 && sort_key($ap[0]) eq sort_key($bp[0])) {
 		# First name matches, try to match some surname
 		my $found = 0;
 		for my $ai (1..$#ap) {
 			for my $bi (1..$#bp) {
-				$found = 1 if (lc($ap[$ai]) eq lc($bp[$bi]));
+				$found = 1 if (sort_key($ap[$ai]) eq sort_key($bp[$bi]));
 			}
 		}
 		if ($found) {
@@ -278,7 +285,7 @@ sub find_where_each_name_is {
 		for my $val (@{$row->{'values'}}) {
 			my $name = get_spreadsheet_name($val);
 			if (defined($name)) {
-				push @{$seen_names{lc $name}}, [$name, $rowno, $colno];
+				push @{$seen_names{sort_key($name)}}, [$name, $rowno, $colno];
 			}
 			++$colno;
 		}
@@ -397,10 +404,10 @@ sub find_diff {
 	}
 	for my $real_name (keys %$have_colors) {
 		next if (exists($want_colors->{$real_name}));
-		if (!exists($seen_names->{lc $real_name})) {
+		if (!exists($seen_names->{sort_key($real_name)})) {
 			# TODO: This can somehow come if we try to add someone who's not in the sheet, too?
 			skv_log("Ønsket å fjerne at $real_name skulle på trening, men de var ikke i regnearket lenger.");
-		} elsif (scalar @{$seen_names->{lc $real_name}} > 1) {
+		} elsif (scalar @{$seen_names->{sort_key($real_name)}} > 1) {
 			# Don't touch them.
 		} else {
 			skv_log("Fjerner at $real_name skal på trening.");
@@ -599,7 +606,7 @@ sub run {
 			$slack_userid_to_slack_name{$userid} = $slack_name;
 		}
 
-		if (exists($seen_names{lc $slack_name})) {
+		if (exists($seen_names{sort_key($slack_name)})) {
 			# The name exists exactly, once or more, so it's a direct match and we ignore any fuzz.
 			$slack_userid_to_real_name{$userid} = $slack_name;
 			push @slack_mapping_updates, {
@@ -664,11 +671,11 @@ sub run {
 		my $real_name = $slack_userid_to_real_name{$userid};
 
 		# See if we can find them in the spreadsheet.
-		if (!exists($seen_names{lc $real_name})) {
+		if (!exists($seen_names{sort_key($real_name)})) {
 			# TODO: Perhaps move this logic further down, for consistency?
 			skv_log("$slack_name ($userid) er påmeldt på Slack, og er mappet til $real_name, men var ikke i noen gruppe.");
 		} else {
-			my $seen = $seen_names{lc $real_name};
+			my $seen = $seen_names{sort_key($real_name)};
 			if (scalar @$seen >= 2) {
 				skv_log("$slack_name ($userid) er påmeldt på Slack, men står flere steder (se over); vet ikke hvilken celle som skal brukes.");
 			} else {
@@ -695,7 +702,7 @@ sub run {
 		for my $diff (@diffs) {
 			my $real_name = $diff->[0];
 
-			my $seen = $seen_names{lc $real_name};
+			my $seen = $seen_names{sort_key($real_name)};
 
 			# We've already complained about these earlier, so just skip them silently.
 			next if (scalar @$seen > 1);
-- 
2.39.2