Fix error offset bug introduced at 1176.

src/pcre2_match.c

@@ -6184,6 +6184,10 @@ if ((match_data->flags & PCRE2_MD_COPIED_SUBJECT) != 0)
   }
 match_data->subject = NULL;
 
+/* Zero the error offset in case the first code unit is invalid UTF. */
+
+match_data->startchar = 0;
+
 
 /* ============================= JIT matching ============================== */
 

testdata/testinput10

@@ -579,4 +579,10 @@
 
 /(?:\x{ff}|\x{3000})/I,utf
 
+/x/utf
+    abxyz
+    \x80\=startchar
+    abc\x80\=startchar
+    abc\x80\=startchar,offset=3
+
 # End of testinput10

testdata/testoutput10

@@ -1803,4 +1803,14 @@ Options: utf
 Starting code units: \xc3 \xe3 
 Subject length lower bound = 1
 
+/x/utf
+    abxyz
+ 0: x
+    \x80\=startchar
+Failed: error -22: UTF-8 error: isolated byte with 0x80 bit set at offset 0
+    abc\x80\=startchar
+Failed: error -22: UTF-8 error: isolated byte with 0x80 bit set at offset 3
+    abc\x80\=startchar,offset=3
+Error -36 (bad UTF-8 offset)
+
 # End of testinput10