From 8b72cb64b2d556cc27c2e6bc91d3884bd328dbba Mon Sep 17 00:00:00 2001
From: Max Kellermann <max.kellermann@gmail.com>
Date: Tue, 15 Nov 2022 17:09:02 +0100
Subject: [PATCH] lib/icu/CaseFold: rename to IcuCanonicalize() and add flag
parameter
Prepare for adding more transformations.
---
.../icu/{CaseFold.cxx => Canonicalize.cxx} | 17 +++++++++--------
.../icu/{CaseFold.hxx => Canonicalize.hxx} | 16 ++++++++++------
src/lib/icu/Compare.cxx | 19 ++++++++++---------
src/lib/icu/meson.build | 2 +-
4 files changed, 30 insertions(+), 24 deletions(-)
rename src/lib/icu/{CaseFold.cxx => Canonicalize.cxx} (78%)
rename src/lib/icu/{CaseFold.hxx => Canonicalize.hxx} (76%)
diff --git a/src/lib/icu/CaseFold.cxx b/src/lib/icu/Canonicalize.cxx
similarity index 78%
rename from src/lib/icu/CaseFold.cxx
rename to src/lib/icu/Canonicalize.cxx
index fd435cb06..8f97cb9f5 100644
--- a/src/lib/icu/CaseFold.cxx
+++ b/src/lib/icu/Canonicalize.cxx
@@ -17,10 +17,10 @@
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
*/
-#include "CaseFold.hxx"
+#include "Canonicalize.hxx"
#include "config.h"
-#ifdef HAVE_ICU_CASE_FOLD
+#ifdef HAVE_ICU_CANONICALIZE
#include "util/AllocatedString.hxx"
@@ -32,18 +32,19 @@
#endif
AllocatedString
-IcuCaseFold(std::string_view src) noexcept
+IcuCanonicalize(std::string_view src, bool fold_case) noexcept
try {
#ifdef HAVE_ICU
auto u = UCharFromUTF8(src);
if (u.data() == nullptr)
return {src};
- auto folded = IcuFoldCase(ToStringView(std::span{u}));
- if (folded == nullptr)
- return {src};
+ if (fold_case)
+ if (auto folded = IcuFoldCase(ToStringView(std::span{u}));
+ folded != nullptr)
+ u = std::move(folded);
- return UCharToUTF8(ToStringView(std::span{folded}));
+ return UCharToUTF8(ToStringView(std::span{u}));
#else
#error not implemented
#endif
@@ -51,4 +52,4 @@ try {
return {src};
}
-#endif /* HAVE_ICU_CASE_FOLD */
+#endif /* HAVE_ICU_CANONICALIZE */
diff --git a/src/lib/icu/CaseFold.hxx b/src/lib/icu/Canonicalize.hxx
similarity index 76%
rename from src/lib/icu/CaseFold.hxx
rename to src/lib/icu/Canonicalize.hxx
index 5cfe77717..d2d33c004 100644
--- a/src/lib/icu/CaseFold.hxx
+++ b/src/lib/icu/Canonicalize.hxx
@@ -17,21 +17,25 @@
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
*/
-#ifndef MPD_ICU_CASE_FOLD_HXX
-#define MPD_ICU_CASE_FOLD_HXX
+#pragma once
#include "config.h"
#ifdef HAVE_ICU
-#define HAVE_ICU_CASE_FOLD
+#define HAVE_ICU_CANONICALIZE
#include <string_view>
class AllocatedString;
+/**
+ * Transform the given string to "canonical" form to allow fuzzy
+ * string comparisons. The full set of features (if ICU is being
+ * used):
+ *
+ * - case folding (optional)
+ */
AllocatedString
-IcuCaseFold(std::string_view src) noexcept;
-
-#endif
+IcuCanonicalize(std::string_view src, bool fold_case) noexcept;
#endif
diff --git a/src/lib/icu/Compare.cxx b/src/lib/icu/Compare.cxx
index 109b167fc..d8b147c45 100644
--- a/src/lib/icu/Compare.cxx
+++ b/src/lib/icu/Compare.cxx
@@ -18,7 +18,7 @@
*/
#include "Compare.hxx"
-#include "CaseFold.hxx"
+#include "Canonicalize.hxx"
#include "util/StringAPI.hxx"
#include "util/StringCompare.hxx"
#include "config.h"
@@ -28,10 +28,10 @@
#include <windows.h>
#endif
-#ifdef HAVE_ICU_CASE_FOLD
+#ifdef HAVE_ICU_CANONICALIZE
IcuCompare::IcuCompare(std::string_view _needle) noexcept
- :needle(IcuCaseFold(_needle)) {}
+ :needle(IcuCanonicalize(_needle, true)) {}
#elif defined(_WIN32)
@@ -54,8 +54,8 @@ IcuCompare::IcuCompare(std::string_view _needle) noexcept
bool
IcuCompare::operator==(const char *haystack) const noexcept
{
-#ifdef HAVE_ICU_CASE_FOLD
- return StringIsEqual(IcuCaseFold(haystack).c_str(), needle.c_str());
+#ifdef HAVE_ICU_CANONICALIZE
+ return StringIsEqual(IcuCanonicalize(haystack, true).c_str(), needle.c_str());
#elif defined(_WIN32)
if (needle == nullptr)
/* the MultiByteToWideChar() call in the constructor
@@ -80,8 +80,8 @@ IcuCompare::operator==(const char *haystack) const noexcept
bool
IcuCompare::IsIn(const char *haystack) const noexcept
{
-#ifdef HAVE_ICU_CASE_FOLD
- return StringFind(IcuCaseFold(haystack).c_str(),
+#ifdef HAVE_ICU_CANONICALIZE
+ return StringFind(IcuCanonicalize(haystack, true).c_str(),
needle.c_str()) != nullptr;
#elif defined(_WIN32)
if (needle == nullptr)
@@ -117,8 +117,9 @@ IcuCompare::IsIn(const char *haystack) const noexcept
bool
IcuCompare::StartsWith(const char *haystack) const noexcept
{
-#ifdef HAVE_ICU_CASE_FOLD
- return StringStartsWith(IcuCaseFold(haystack).c_str(), needle);
+#ifdef HAVE_ICU_CANONICALIZE
+ return StringStartsWith(IcuCanonicalize(haystack, true).c_str(),
+ needle);
#elif defined(_WIN32)
if (needle == nullptr)
/* the MultiByteToWideChar() call in the constructor
diff --git a/src/lib/icu/meson.build b/src/lib/icu/meson.build
index 65d90b199..2eb5ce446 100644
--- a/src/lib/icu/meson.build
+++ b/src/lib/icu/meson.build
@@ -2,7 +2,7 @@ icu_dep = dependency('icu-i18n', version: '>= 50', required: get_option('icu'))
conf.set('HAVE_ICU', icu_dep.found())
icu_sources = [
- 'CaseFold.cxx',
+ 'Canonicalize.cxx',
'Compare.cxx',
'Collate.cxx',
'Converter.cxx',