Use noexcept in CharacterCategoryMap.

2 files changed, 18 insertions, 13 deletions

diff --git a/src/CharacterCategoryMap.cxx b/src/CharacterCategoryMap.cxx
index b8b20fee7..3ef5bb748 100644
--- a/src/CharacterCategoryMap.cxx
+++ b/src/CharacterCategoryMap.cxx
@@ -4050,12 +4050,17 @@ constexpr int maskCategory = 0x1F;
 // possibly for 0..0xff for most Western European text or 0..0xfff for most
 // alphabetic languages.
 
-CharacterCategory CategoriseCharacter(int character) {
+CharacterCategory CategoriseCharacter(int character) noexcept {
 	if (character < 0 || character > maxUnicode)
 		return ccCn;
 	const int baseValue = character * (maskCategory+1) + maskCategory;
-	const int *placeAfter = std::lower_bound(catRanges, std::end(catRanges), baseValue);
-	return static_cast<CharacterCategory>(*(placeAfter-1) & maskCategory);
+	try {
+		// lower_bound will never throw with these args but its not marked noexcept so add catch to pretend.
+		const int *placeAfter = std::lower_bound(catRanges, std::end(catRanges), baseValue);
+		return static_cast<CharacterCategory>(*(placeAfter - 1) & maskCategory);
+	} catch (...) {
+		return ccCn;
+	}
 }
 
 // Implementation of character sets recommended for identifiers in Unicode Standard Annex #31.
@@ -4158,7 +4163,7 @@ bool OmitXidContinue(int character) noexcept {
 
 // UAX #31 defines ID_Start as
 // [[:L:][:Nl:][:Other_ID_Start:]--[:Pattern_Syntax:]--[:Pattern_White_Space:]]
-bool IsIdStart(int character) {
+bool IsIdStart(int character) noexcept {
 	if (IsIdPattern(character)) {
 		return false;
 	}
@@ -4173,7 +4178,7 @@ bool IsIdStart(int character) {
 
 // UAX #31 defines ID_Continue as
 // [[:ID_Start:][:Mn:][:Mc:][:Nd:][:Pc:][:Other_ID_Continue:]--[:Pattern_Syntax:]--[:Pattern_White_Space:]]
-bool IsIdContinue(int character) {
+bool IsIdContinue(int character) noexcept {
 	if (IsIdPattern(character)) {
 		return false;
 	}
@@ -4187,7 +4192,7 @@ bool IsIdContinue(int character) {
 }
 
 // XID_Start is ID_Start modified for Normalization Form KC in UAX #31
-bool IsXidStart(int character) {
+bool IsXidStart(int character) noexcept {
 	if (OmitXidStart(character)) {
 		return false;
 	} else {
@@ -4196,7 +4201,7 @@ bool IsXidStart(int character) {
 }
 
 // XID_Continue is ID_Continue modified for Normalization Form KC in UAX #31
-bool IsXidContinue(int character) {
+bool IsXidContinue(int character) noexcept {
 	if (OmitXidContinue(character)) {
 		return false;
 	} else {
diff --git a/src/CharacterCategoryMap.h b/src/CharacterCategoryMap.h
index b3b03eb5a..cd75ec29c 100644
--- a/src/CharacterCategoryMap.h
+++ b/src/CharacterCategoryMap.h
@@ -22,20 +22,20 @@ enum CharacterCategory {
 	ccCc, ccCf, ccCs, ccCo, ccCn
 };
 
-CharacterCategory CategoriseCharacter(int character);
+CharacterCategory CategoriseCharacter(int character) noexcept;
 
 // Common definitions of allowable characters in identifiers from UAX #31.
-bool IsIdStart(int character);
-bool IsIdContinue(int character);
-bool IsXidStart(int character);
-bool IsXidContinue(int character);
+bool IsIdStart(int character) noexcept;
+bool IsIdContinue(int character) noexcept;
+bool IsXidStart(int character) noexcept;
+bool IsXidContinue(int character) noexcept;
 
 class CharacterCategoryMap {
 private:
 	std::vector<unsigned char> dense;
 public:
 	CharacterCategoryMap();
-	CharacterCategory CategoryFor(int character) const {
+	CharacterCategory CategoryFor(int character) const noexcept {
 		if (static_cast<size_t>(character) < dense.size()) {
 			return static_cast<CharacterCategory>(dense[character]);
 		} else {