regex: bug fixes, improved tests

pull/7203/head
penguindark 2020-12-08 19:38:25 +01:00 committed by GitHub
parent a1c0bb3585
commit 2bc1076921
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 66 additions and 21 deletions

View File

@ -1814,6 +1814,12 @@ pub fn (mut re RE) match_base(in_txt byteptr, in_txt_len int ) (int,int) {
re.groups[g_index] = 0
}
re.groups[g_index+1] = i
// if a group end with a dot, manage the not increased char index
if i == re.groups[g_index] {
re.groups[g_index+1] = i+1
}
//println("GROUP ${re.prog[pc].group_id} END [${re.groups[g_index]}, ${re.groups[g_index+1]}]")
// continuous save, save until we have space
@ -2092,6 +2098,13 @@ pub fn (mut re RE) match_base(in_txt byteptr, in_txt_len int ) (int,int) {
re.prog[tmp_pc].group_rep = 0 // clear the repetitions
group_index--
m_state = .ist_next
// if dot char manage advance of the group
if l_ist == u32(ist_dot_char) {
//print("dot char next char")
i+=char_len
}
continue
}
else if rep >= re.prog[tmp_pc].rep_min {

View File

@ -190,6 +190,18 @@ cgroups_test_suite = [
[3, 0, 0, 4, 1, 7, 11, 1, 11, 16],
{'format':int(0)}
},
TestItemCGroup{
"acc +13 pippo",
r"(\w+)\s(.)([0-9]+) \w+",0,13,
[0, 3, 4, 5, 5, 7],
map[string]int{}
},
TestItemCGroup{
"acc +13",
r"(\w+)\s(.)([0-9]+)",0,7,
[0, 3, 4, 5, 5, 7],
map[string]int{}
},
]
)
@ -210,7 +222,12 @@ fn test_regex(){
continue
}
if to.cgn.len > 0 {
re.group_csave = [-1].repeat(3*20+1)
if debug { println("continuous save")}
} else {
if debug { println("NO continuous save")}
}
start, end := re.match_string(to.src)
@ -228,6 +245,7 @@ fn test_regex(){
}
// check cgroups
if to.cgn.len > 0 {
if re.group_csave.len == 0 || re.group_csave[0] != to.cg[0] {
println("Capturing group len error! ${re.group_csave[0]}")
assert false
@ -251,6 +269,20 @@ fn test_regex(){
continue
}
}
} else {
// check normal captured groups
if re.groups.len != to.cg.len {
assert false
}
for ln:=0; ln < re.groups.len; ln++ {
if re.groups[ln] != to.cg[ln] {
println("Capture group doesn't match:")
println("true ground: [${to.cg}]")
println("elaborated : [${re.groups}]")
assert false
}
}
}
}
// check find_all