1
0
mirror of https://github.com/vlang/v.git synced 2023-08-10 21:13:21 +03:00
v/vlib/regex/regex_test.v

274 lines
7.5 KiB
V
Raw Normal View History

2020-01-13 15:30:41 +03:00
import regex
/******************************************************************************
*
* Test section
*
******************************************************************************/
2020-01-13 15:30:41 +03:00
struct TestItem {
src string
q string
s int = 0
e int = 0
2020-01-13 15:30:41 +03:00
}
const(
match_test_suite = [
// positive
TestItem{"this is a good.",r"this",0,4},
TestItem{"this is a good.",r"good",10,14},
TestItem{"this is a good.",r"go+d",10,14},
TestItem{"this is a good.",r"g[oae]+d",10,14},
TestItem{"this is a goed.",r"g[oae]+d",10,14},
TestItem{"this is a good.",r"g[oae]*d",10,14},
TestItem{"this is a goaezd.",r"g[ea-cm-z]*d",10,16},
TestItem{"this is a good.",r"this (\w+) a",0,9},
TestItem{"this is a good.",r"this( \w+){2} g",0,11},
TestItem{"this is a good.",r"( ?\w+){,1}",0,4},
TestItem{"this is a good.",r"( ?\w+)+",0,14},
TestItem{"this is a good.",r"this( \w+)+",0,14},
TestItem{"this is a good sample.",r"( ?\w+){,2}",0,7},
TestItem{"this is a good sample.",r"( ?\w+){,3}",0,9},
TestItem{"this is a good sample.",r"( ?\w+){,4}",0,14},
TestItem{"this is a good sample.",r"( ?\w+){,5}",0,21},
TestItem{"this is a good sample.",r"( ?\w+){2,3}",0,9},
2020-01-18 09:38:00 +03:00
TestItem{"this is a good sample.",r"(\s?\w+){2,3}",0,9},
2020-01-13 15:30:41 +03:00
TestItem{"this these those.",r"(th[ei]se?\s|\.)+",0,11},
TestItem{"this these those ",r"(th[eio]se? ?)+",0,17},
TestItem{"this these those ",r"(th[eio]se? )+",0,17},
TestItem{"this,these,those. over",r"(th[eio]se?[,. ])+",0,17},
TestItem{"soday,this,these,those. over",r"(th[eio]se?[,. ])+",6,23},
2020-01-18 09:38:00 +03:00
2020-01-13 15:30:41 +03:00
TestItem{"cpapaz",r"(c(pa)+z)",0,6},
TestItem{"this is a cpapaz over",r"(c(pa)+z)",10,16},
TestItem{"this is a cpapapez over",r"(c(p[ae])+z)",10,18},
TestItem{"test@post.pip.com",r"[a-z0-9_]+@([a-z0-9_]+\.?)+",0,17},
TestItem{"test1@post.pip.com, pera",r"[\w]+@([\w]+\.)+\w+",0,18},
TestItem{"pippo@pera.com ",r"[a-z0-9_]+@([a-z0-9_]+\.?)+",0,14},
TestItem{"adce aabe",r"(a(ab)+)|(a(dc)+)e",0,4},
TestItem{"zadce aabe",r"(a(ab)+)|(a(dc)+)e",1,5},
TestItem{"abbz accz addz.",r"c|(d)|e|(ab+)",0,3},
TestItem{"this those these ciao",r"((t[hieo]+se?)\s*)+",0,17},
TestItem{"this ciao",r"((t[hieo]+se?)\s*)+",0,5},
TestItem{"this cpapaz adce aabe",r"(c(pa)+z)(\s[\a]+){2}",5,21},
TestItem{"1234this cpapaz adce aabe",r"(c(pa)+z)(\s[\a]+){2}$",9,25},
TestItem{"this cpapaz adce aabe third",r"(c(pa)+z)(\s[\a]+){2}",5,21},
2020-01-18 09:38:00 +03:00
TestItem{"123cpapaz ole. pippo",r"(c(pa)+z)(\s+\a+[\.,]?)+",3,20},
TestItem{"this is a good sample.",r".*i(\w)+",0,4},
TestItem{"soday,this,these,those. over",r".*,(th[eio]se?[,. ])+",0,23},
TestItem{"soday,this,these,thesa.thesi over",r".*,(th[ei]se?[,. ])+(thes[ai][,. ])+",0,29},
2020-01-13 15:30:41 +03:00
TestItem{"cpapaz ole. pippo,",r".*(c(pa)+z)(\s+\a+[\.,]?)+",0,18},
TestItem{"cpapaz ole. pippo",r".*(c(pa)+z)(\s+\a+[\.,]?)+",0,17},
TestItem{"cpapaz ole. pippo, 852",r".*(c(pa)+z)(\s+\a+[\.,]?)+",0,18},
TestItem{"123cpapaz ole. pippo",r".*(c(pa)+z)(\s+\a+[\.,]?)+",0,20},
TestItem{"...cpapaz ole. pippo",r".*(c(pa)+z)(\s+\a+[\.,]?)+",0,20},
2020-01-18 09:38:00 +03:00
2020-01-13 15:30:41 +03:00
TestItem{"cpapaz ole. pippo,",r".*c.+ole.*pi",0,14},
TestItem{"cpapaz ole. pipipo,",r".*c.+ole.*p([ip])+o",0,18},
TestItem{"cpapaz ole. pipipo",r"^.*c.+ol?e.*p([ip])+o$",0,18},
2020-01-16 02:39:33 +03:00
TestItem{"abbb",r"ab{2,3}?",0,3},
2020-01-18 09:38:00 +03:00
TestItem{" pippo pera",r"\s(.*)pe(.*)",0,11},
TestItem{" abb",r"\s(.*)",0,4},
2020-01-13 15:30:41 +03:00
// 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 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$",-1,0},
// check unicode
TestItem{"this is a test",r".*a [-Ⅵ ]+",0,34},
TestItem{"123 test",r"[-\s]+",3,23},
]
)
struct TestItemFa {
src string
q string
r []int
}
const (
match_test_suite_fa = [
// find_all tests
TestItemFa{
"oggi pippo è andato a casa di pluto ed ha trovato pippo",
r"p[iplut]+o",
[5, 10, 31, 36, 51, 56]
},
TestItemFa{
"oggi pibao è andato a casa di pbababao ed ha trovato pibabababao",
r"(pi?(ba)+o)",
[5, 10, 31, 39, 54, 65]
},
]
)
struct TestItemRe {
src string
q string
rep string
r string
}
const (
match_test_suite_re = [
// replace tests
TestItemRe{
"oggi pibao è andato a casa di pbababao ed ha trovato pibabababao",
r"(pi?(ba)+o)",
"CIAO",
"oggi CIAO è andato a casa di CIAO ed ha trovato CIAO"
},
TestItemRe{
"Today is a good day and tomorrow will be for sure.",
r"[Tt]o\w+",
"CIAO",
"CIAO is a good day and CIAO will be for sure."
}
]
)
2020-01-13 15:30:41 +03:00
fn test_regex(){
// check find_all
for c,to in match_test_suite_fa{
// debug print
//println("#$c [$to.src] q[$to.q] $to.r")
mut re, re_err, err_pos := regex.regex(to.q)
if re_err == regex.COMPILE_OK {
res := re.find_all(to.src)
if res.len != to.r.len {
println("ERROR: find_all, array of different size.")
assert false
}
for c1,i in res {
if i != to.r[c1] {
println("ERROR: find_all, different indexes.")
assert false
}
}
} else {
println("query: $to.q")
lc := "-".repeat(err_pos-1)
println("err : $lc^")
err_str := re.get_parse_error_string(re_err)
println("ERROR: $err_str")
assert false
}
}
// check replace
for c,to in match_test_suite_re{
// debug print
//println("#$c [$to.src] q[$to.q] $to.r")
mut re, re_err, err_pos := regex.regex(to.q)
if re_err == regex.COMPILE_OK {
res := re.replace(to.src,to.rep)
if res != to.r {
println("ERROR: replace.")
assert false
}
} else {
println("query: $to.q")
lc := "-".repeat(err_pos-1)
println("err : $lc^")
err_str := re.get_parse_error_string(re_err)
println("ERROR: $err_str")
assert false
}
}
// check match and find
2020-01-13 15:30:41 +03:00
for c,to in match_test_suite {
// debug print
//println("#$c [$to.src] q[$to.q] $to.s")
// test the find
if to.s > 0 {
mut re, re_err, err_pos := regex.regex(to.q)
if re_err == regex.COMPILE_OK {
//q_str := re.get_query()
//println("Query: $q_str")
start,end := re.find(to.src)
if start != to.s || end != to.e {
err_str := re.get_parse_error_string(start)
println("ERROR : $err_str")
assert false
} else {
//tmp_str := text[start..end]
//println("found in [$start, $end] => [$tmp_str]")
assert true
}
} else {
println("query: $to.q")
lc := "-".repeat(err_pos-1)
println("err : $lc^")
err_str := re.get_parse_error_string(re_err)
println("ERROR: $err_str")
assert false
}
continue
}
// test the match
mut re := regex.new_regex()
//re.debug = true
re_err,err_pos := re.compile(to.q)
if re_err == regex.COMPILE_OK {
//println("#$c [$to.src] q[$to.q]")
start, end := re.match_string(to.src)
mut tmp_str := ""
if start >= 0 && end > start{
tmp_str = to.src[start..end]
}
if start != to.s || end != to.e {
println("#$c [$to.src] q[$to.q] res[$tmp_str] $start, $end")
println("ERROR!")
//C.printf("ERROR!! res:(%d, %d) refh:(%d, %d)\n",start, end, to.s, to.e)
2020-01-13 15:30:41 +03:00
assert false
break
}
// rerun to test consistency
tmp_str1 := to.src.clone()
start1, end1 := re.match_string(tmp_str1)
if start1 != start || end1 != end {
println("two run ERROR!!")
assert false
break
}
} else {
println("query: $to.q")
lc := "-".repeat(err_pos-1)
println("err : $lc^")
2020-01-13 15:30:41 +03:00
err_str := re.get_parse_error_string(re_err)
println("ERROR: $err_str")
assert false
break
}
}
}