From 908296cdfb7cb4bf407c1a3d7283119c57e456ab Mon Sep 17 00:00:00 2001
From: penguindark <57967770+penguindark@users.noreply.github.com>
Date: Sat, 1 Jan 2022 08:21:27 +0100
Subject: [PATCH] regex: improve errors for edge cases (#13008)

* code cleaning, added more clear errors for dots and ORs

* added failed match index for better find functions, updated tests

* added index in match failed, updated tests

* test cleaning

* test check
---
 vlib/regex/regex.v      | 53 ++++++++++++-----------------------------
 vlib/regex/regex_test.v | 26 ++++++++++----------
 vlib/regex/regex_util.v | 14 ++++++++++-
 3 files changed, 41 insertions(+), 52 deletions(-)

diff --git a/vlib/regex/regex.v b/vlib/regex/regex.v
index 36b8b8f2a3..a0c8f51514 100644
--- a/vlib/regex/regex.v
+++ b/vlib/regex/regex.v
@@ -42,6 +42,7 @@ pub const (
 	err_group_qm_notation    = -10 // group invalid notation
 	err_invalid_or_with_cc   = -11 // invalid or on two consecutive char class
 	err_neg_group_quantifier = -12 // negation groups can not have quantifier
+	err_consecutive_dots     = -13 // two consecutive dots is an error
 )
 
 const (
@@ -200,6 +201,7 @@ pub fn (re RE) get_parse_error_string(err int) string {
 		regex.err_group_qm_notation { return 'err_group_qm_notation' }
 		regex.err_invalid_or_with_cc { return 'err_invalid_or_with_cc' }
 		regex.err_neg_group_quantifier { return 'err_neg_group_quantifier' }
+		regex.err_consecutive_dots { return 'err_consecutive_dots' }
 		else { return 'err_unknown' }
 	}
 }
@@ -283,14 +285,7 @@ pub const (
 	f_src = 0x00020000 // search mode enabled
 )
 
-struct StateDotObj {
-mut:
-	i                 int = -1 // char index in the input buffer
-	pc                int = -1 // program counter saved
-	mi                int = -1 // match_index saved
-	group_stack_index int = -1 // continuous save on capturing groups
-}
-
+// Log function prototype
 pub type FnLog = fn (string)
 
 pub struct RE {
@@ -1042,6 +1037,7 @@ fn (mut re RE) impl_compile(in_txt string) (int, int) {
 			re.prog[goto_pc].goto_pc = pc // start goto point to the end group pc
 			// re.prog[goto_pc].group_id = group_count         // id of this group, used for storing data
 
+			// duplicate the negation group info and settings
 			if re.prog[goto_pc].group_neg == true {
 				re.prog[pc].group_neg = re.prog[goto_pc].group_neg
 				re.prog[pc].rep_min = re.prog[goto_pc].rep_min
@@ -1054,6 +1050,11 @@ fn (mut re RE) impl_compile(in_txt string) (int, int) {
 
 		// ist_dot_char match any char except the following token
 		if char_len == 1 && pc >= 0 && byte(char_tmp) == `.` {
+			// consecutive ist_dot_char is a syntax error
+			if pc > 0 && re.prog[pc - 1].ist == regex.ist_dot_char {
+				return regex.err_consecutive_dots, i
+			}
+
 			re.prog[pc].ist = u32(0) | regex.ist_dot_char
 			re.prog[pc].rep_min = 1
 			re.prog[pc].rep_max = 1
@@ -1228,7 +1229,7 @@ fn (mut re RE) impl_compile(in_txt string) (int, int) {
 
 	// check for OR at the end of the program
 	if pc > 0 && re.prog[pc - 1].ist == regex.ist_or_branch {
-		return regex.err_syntax_error, in_txt.len
+		return regex.err_syntax_error, in_txt.len - 1
 	}
 
 	// store the number of groups in the query
@@ -1873,7 +1874,7 @@ pub fn (mut re RE) match_base(in_txt &byte, in_txt_len int) (int, int) {
 			}
 
 			// print("No good exit!!")
-			return regex.no_match_found, 0
+			return regex.no_match_found, state.i
 		}
 
 		// starting and init
@@ -1959,7 +1960,7 @@ pub fn (mut re RE) match_base(in_txt &byte, in_txt_len int) (int, int) {
 			}
 
 			// exit on no match
-			return result, 0
+			return result, state.i
 		}
 		// ist_load
 		else if m_state == .ist_load {
@@ -2164,30 +2165,6 @@ pub fn (mut re RE) match_base(in_txt &byte, in_txt_len int) (int, int) {
 				continue
 			}
 			// check bsls
-			/*
-			else if ist == regex.ist_bsls_char {
-				state.match_flag = false
-				tmp_res := re.prog[state.pc].validator(byte(ch))
-				// println("BSLS in_ch: ${ch:c} res: $tmp_res")
-				if tmp_res {
-					state.match_flag = true
-					l_ist = u32(regex.ist_bsls_char)
-
-					if state.first_match < 0 {
-						state.first_match = state.i
-					}
-
-					state.match_index = state.i
-
-					re.prog[state.pc].rep++ // increase repetitions
-					state.i += char_len // next char
-					m_state = .ist_quant_p
-					continue
-				}
-				m_state = .ist_quant_n
-				continue
-			}
-			*/
 			else if ist == regex.ist_bsls_char {
 				// println("ist_bsls_char rep: ${re.prog[state.pc].rep}")
 
@@ -2541,14 +2518,14 @@ pub fn (mut re RE) match_base(in_txt &byte, in_txt_len int) (int, int) {
 					return state.first_match, state.i
 				}
 				// println("Program not finished! ")
-				return regex.no_match_found, 0
+				return regex.no_match_found, state.i
 			}
 			if src_end {
 				// println("program end")
 				return state.first_match, state.i
 			}
 			// print("No match found!!")
-			return regex.no_match_found, 0
+			return regex.no_match_found, state.i
 		} else {
 			// println("Group match! OK")
 			// println("first_match: $state.first_match, i: $state.i")
@@ -2559,5 +2536,5 @@ pub fn (mut re RE) match_base(in_txt &byte, in_txt_len int) (int, int) {
 		}
 	}
 	// println("no_match_found, natural end")
-	return regex.no_match_found, 0
+	return regex.no_match_found, state.i
 }
diff --git a/vlib/regex/regex_test.v b/vlib/regex/regex_test.v
index a37e9afb31..8f479b757b 100644
--- a/vlib/regex/regex_test.v
+++ b/vlib/regex/regex_test.v
@@ -18,7 +18,7 @@ const(
 match_test_suite = [
 	// minus in CC
 	TestItem{"d.def",r"abc.\.[\w\-]{,100}",-1,0},
-	TestItem{"abc12345.asd",r"abc.\.[\w\-]{,100}",-1,0},
+	TestItem{"abc12345.asd",r"abc.\.[\w\-]{,100}",-1,4},
 	TestItem{"abca.exe",r"abc.\.[\w\-]{,100}",0,8},
 	TestItem{"abc2.exe-test_12",r"abc.\.[\w\-]{,100}",0,16},
 	TestItem{"abcdefGHK",r"[a-f]+\A+",0,9},
@@ -96,30 +96,30 @@ match_test_suite = [
 
 	// negative
 	TestItem{"zthis ciao",r"((t[hieo]+se?)\s*)+",-1,0},
-	TestItem{"this is a good.",r"thes",-1,0},
-	TestItem{"test1post.pip.com, pera",r"[\w]+@([\w]+\.)+\w+",-1,0},
+	TestItem{"this is a good.",r"thes",-1,2},
+	TestItem{"test1post.pip.com, pera",r"[\w]+@([\w]+\.)+\w+",-1,9},
 	TestItem{"this cpapaz adce",r"(c(pa)+z)(\s[\a]+){2}",-1,0},
 	TestItem{"this cpapaz adce aabe third",r"(c(pa)+z)(\s[\a]+){2}$",-1,0},
 	TestItem{"1234this cpapaz adce aabe ter",r"(c(pa)+z)(\s[\a]+){2}$",-1,0},
 	TestItem{"cpapaz ole. pipipo,",r"^.*c.+ol?e.*p([ip])+o$",-1,0},
-	TestItem{"/home/us_er/pippo/info-01.jpeg", r"(/?[-\w_]+)*\.txt$",-1,0}
+	TestItem{"/home/us_er/pippo/info-01.jpeg", r"(/?[-\w_]+)*\.txt$",-1,26}
 
 	// check unicode
 	TestItem{"this is a Ⅰ Ⅱ Ⅲ Ⅳ Ⅴ Ⅵ test",r".*a [Ⅰ-Ⅵ ]+",0,34},
 	TestItem{"123Ⅰ Ⅱ Ⅲ Ⅳ Ⅴ Ⅵ test",r"[Ⅰ-Ⅴ\s]+",3,23},
 
 	// new edge cases
-	TestItem{"12345678", r"[0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9]",-1,0},
+	TestItem{"12345678", r"[0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9]",-1,8},
 	TestItem{"12345678", r"[0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9]",0,8},
 	TestItem{"123456789", r"^[0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9]$",0,9}
 	TestItem{"12345678", r"^\d{8}$",0,8},
 	TestItem{"12345678", r"^\d{7}$",-1,0},
-	TestItem{"12345678", r"^\d{9}$",-1,0},
+	TestItem{"12345678", r"^\d{9}$",-1,8},
 	
 	TestItem{"eth", r"(oth)|(eth)",0,3},
-	TestItem{"et", r"(oth)|(eth)",-1,0},
-	TestItem{"et", r".*(oth)|(eth)",-1,0},
-	TestItem{"peoth", r".*(ith)|(eth)",-1,0},
+	TestItem{"et", r"(oth)|(eth)",-1,2},
+	TestItem{"et", r".*(oth)|(eth)",-1,2},
+	TestItem{"peoth", r".*(ith)|(eth)",-1,5},
 
 	TestItem{"poth", r"(eth)|(oth)",1,4},
 	TestItem{"poth", r"(oth)|(eth)",1,4},
@@ -132,7 +132,7 @@ match_test_suite = [
 	TestItem{"accccb deer", r"^a(.*)b d(.+)r",0,11},
 	TestItem{"accccb deer", r"^a(.*)b d(.+)",0,11},
 	TestItem{"accccb deer", r"^(.*)$",0,11},
-	TestItem{"accccb deer", r"^a(.*)b d(.+)p",-1,0},
+	TestItem{"accccb deer", r"^a(.*)b d(.+)p",-1,11},
 	TestItem{"##.#....#.##.####...#.##", r".{18}[.#]",0,19},
 	TestItem{"#.#......##.#..#..##........##....###...##...######.......#.....#..#......#...#........###.#..#.", r'.*#[.#]{4}##[.#]{4}##[.#]{4}###',0,49},
 
@@ -328,19 +328,19 @@ find_all_test_suite = [
 		[29, 49],
 		['#....###...##...####']
 	},
-		Test_find_all{
+	Test_find_all{
 		"#.#......##.#..#..##........##....###...##...######.......#.....#..#......#...#........###.#..#.",
 		r".*#[.#]{4}##[.#]{4}##[.#]{4}###",
 		[0, 49],
 		['#.#......##.#..#..##........##....###...##...####']
 	},
-		Test_find_all{
+	Test_find_all{
 		"1234 Aa dddd Aaf 12334 Aa opopo Aaf",
 		r"Aa.+Aaf",
 		[5, 16, 23, 35],
 		['Aa dddd Aaf', 'Aa opopo Aaf']
 	},
-		Test_find_all{
+	Test_find_all{
 		"@for something @endfor @for something else @endfor altro testo @for body @endfor uno due @for senza dire più @endfor pippo",
 		r"@for.+@endfor",
 		[0, 22, 23, 50, 63, 80, 89, 117],
diff --git a/vlib/regex/regex_util.v b/vlib/regex/regex_util.v
index ed29b8acfd..02853ced95 100644
--- a/vlib/regex/regex_util.v
+++ b/vlib/regex/regex_util.v
@@ -272,8 +272,14 @@ pub fn (mut re RE) find_all(in_txt string) []int {
 				i += e
 				continue
 			}
+			/*
+			if e > 0 {
+				i += e
+				continue
+			}
+			*/
+			i++
 		}
-		i++
 	}
 	// re.flag = old_flag
 	return res
@@ -306,6 +312,12 @@ pub fn (mut re RE) find_all_str(in_txt string) []string {
 				continue
 			}
 		}
+		/*
+		if e > 0 {
+			i += e
+			continue
+		}
+		*/
 		i++
 	}
 	// re.flag = old_flag