From f50940531dd57135fe60aa393ac9d3281f352d88 Mon Sep 17 00:00:00 2001
From: Bram Moolenaar <Bram@vim.org>
Date: Fri, 29 Jul 2022 16:22:25 +0100
Subject: [PATCH 002/123] patch 9.0.0105: illegal memory access when pattern
 starts with illegal byte

Problem:    Illegal memory access when pattern starts with illegal byte.
Solution:   Do not match a character with an illegal byte.
---
 src/regexp.c                     |  6 +++++-
 src/testdir/test_regexp_utf8.vim | 15 +++++++++++++++
 2 files changed, 20 insertions(+), 1 deletion(-)

diff --git a/src/regexp.c b/src/regexp.c
index 1a5cfd0..bec0464 100644
--- a/src/regexp.c
+++ b/src/regexp.c
@@ -1641,7 +1641,11 @@ cstrchr(char_u *s, int c)
 	{
 	    if (enc_utf8 && c > 0x80)
 	    {
-		if (utf_fold(utf_ptr2char(p)) == cc)
+		int uc = utf_ptr2char(p);
+
+		// Do not match an illegal byte.  E.g. 0xff matches 0xc3 0xbf,
+		// not 0xff.
+		if ((uc < 0x80 || uc != *p) && utf_fold(uc) == cc)
 		    return p;
 	    }
 	    else if (*p == c || *p == cc)
diff --git a/src/testdir/test_regexp_utf8.vim b/src/testdir/test_regexp_utf8.vim
index d88e263..e7672dd 100644
--- a/src/testdir/test_regexp_utf8.vim
+++ b/src/testdir/test_regexp_utf8.vim
@@ -1,5 +1,7 @@
 " Tests for regexp in utf8 encoding
 
+source shared.vim
+
 func s:equivalence_test()
   let str = "AÀÁÂÃÄÅĀĂĄǍǞǠǺȂȦȺḀẠẢẤẦẨẪẬẮẰẲẴẶ BƁɃḂḄḆ CÇĆĈĊČƇȻḈꞒ DĎĐƊḊḌḎḐḒ EÈÉÊËĒĔĖĘĚȄȆȨɆḔḖḘḚḜẸẺẼẾỀỂỄỆ FƑḞꞘ GĜĞĠĢƓǤǦǴḠꞠ HĤĦȞḢḤḦḨḪⱧ IÌÍÎÏĨĪĬĮİƗǏȈȊḬḮỈỊ JĴɈ KĶƘǨḰḲḴⱩꝀ LĹĻĽĿŁȽḶḸḺḼⱠ MḾṀṂ NÑŃŅŇǸṄṆṈṊꞤ OÒÓÔÕÖØŌŎŐƟƠǑǪǬǾȌȎȪȬȮȰṌṎṐṒỌỎỐỒỔỖỘỚỜỞỠỢ PƤṔṖⱣ QɊ RŔŖŘȐȒɌṘṚṜṞⱤꞦ SŚŜŞŠȘṠṢṤṦṨⱾꞨ TŢŤŦƬƮȚȾṪṬṮṰ UÙÚÛÜŨŪŬŮŰƯǕǙǛǓǗȔȖɄṲṴṶṸṺỤỦỨỪỬỮỰ  VƲṼṾ WŴẀẂẄẆẈ XẊẌ YÝŶŸƳȲɎẎỲỴỶỸ ZŹŻŽƵẐẒẔⱫ aàáâãäåāăąǎǟǡǻȃȧᶏḁẚạảấầẩẫậắằẳẵặⱥ bƀɓᵬᶀḃḅḇ cçćĉċčƈȼḉꞓꞔ dďđɗᵭᶁᶑḋḍḏḑḓ eèéêëēĕėęěȅȇȩɇᶒḕḗḙḛḝẹẻẽếềểễệ fƒᵮᶂḟꞙ gĝğġģǥǧǵɠᶃḡꞡ hĥħȟḣḥḧḩḫẖⱨꞕ iìíîïĩīĭįǐȉȋɨᶖḭḯỉị jĵǰɉ kķƙǩᶄḱḳḵⱪꝁ lĺļľŀłƚḷḹḻḽⱡ mᵯḿṁṃ nñńņňʼnǹᵰᶇṅṇṉṋꞥ oòóôõöøōŏőơǒǫǭǿȍȏȫȭȯȱɵṍṏṑṓọỏốồổỗộớờởỡợ pƥᵱᵽᶈṕṗ qɋʠ rŕŗřȑȓɍɽᵲᵳᶉṛṝṟꞧ sśŝşšșȿᵴᶊṡṣṥṧṩꞩ tţťŧƫƭțʈᵵṫṭṯṱẗⱦ uùúûüũūŭůűųǚǖưǔǘǜȕȗʉᵾᶙṳṵṷṹṻụủứừửữự vʋᶌṽṿ wŵẁẃẅẇẉẘ xẋẍ yýÿŷƴȳɏẏẙỳỵỷỹ zźżžƶᵶᶎẑẓẕⱬ"
   let groups = split(str)
@@ -560,6 +562,19 @@ func Test_match_invalid_byte()
   call delete('Xinvalid')
 endfunc
 
+func Test_match_illegal_byte()
+  let lines =<< trim END
+      silent! buffer ÿ\c
+      next ÿ
+      0scriptnames
+      source
+  END
+  call writefile(lines, 'Xregexp')
+  call system(GetVimCommand() .. ' -X -Z -e -s -S Xregexp -c qa!')
+
+  call delete('Xregexp')
+endfunc
+
 func Test_match_too_complicated()
   set regexpengine=1
   exe "noswapfile vsplit \xeb\xdb\x99"
-- 
1.8.3.1