Change 579272908 by alexsav@alexsav:fig-export-clean-1452-change-1:1454:citc on 2023/11/03 12:42:34

	Adds check for u16_u8 xFF
	
	PRESUBMIT=passed
	BUG=308240685
	FIXED=308240685
	R=mghiware
	APPROVED=mghiware
	REQUIRED_REVIEW=1
	DELTA=8 (7 added, 0 deleted, 1 changed)
	DELTA_BY_EXTENSION=cxx=8
	OCL=578306755
	FIG_CHANGESET=ea4fd90ff399df9c012782dcd2c45439c6e9fcaf
	FIG_WORKSPACE=alexsav/1452:clean

Affected files ...

... //depot//src/hunspell/csutil.cxx#4 edit

==== //depot//src/hunspell/csutil.cxx#3 - /google/src/files/579272908/depot//src/hunspell/csutil.cxx ====
--- /google/src/files/537916729/depot//src/hunspell/csutil.cxx	2023-06-05 13:32:42.000000000 -0400
+++ /google/src/files/579272908/depot//src/hunspell/csutil.cxx	2023-11-03 15:42:34.000000000 -0400
@@ -48,7 +48,14 @@
     while ((u2 < u2_max) && (u8 < u8_max)) {
         if (u2->h) { // > 0xFF
             // XXX 4-byte haven't implemented yet.
-            if (u2->h >= 0x08) {   // >= 0x800 (3-byte UTF-8 character)
+            if (u2->h == 0xFF) {
+              // Error case. Handling this with the below else case will insert
+              // \U0000FFFD, shifting the string and pushing it into other
+              // memory locations, which might be later read as garbage and
+              // cause memory errors.
+              *u8 = 0xFF;
+              u8++;
+            } else if (u2->h >= 0x08) {   // >= 0x800 (3-byte UTF-8 character)
                 *u8 = 0xe0 + (u2->h >> 4);
                 u8++;
                 if (u8 < u8_max) {