miloignis/kraken
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Optimizations, regex character ranges

Browse Source
This commit is contained in:
Nathan Braswell
2016-05-05 04:51:10 -04:00
parent 02c77899b8
commit 9d7a65294f
8 changed files with 100 additions and 76 deletions
Download Patch File Download Diff File Expand all files Collapse all files

41
stdlib/regex.krak
View File

@@ -1,4 +1,6 @@
import io
import string
import ast_transformation
import vector
import string
import mem
@@ -15,10 +17,16 @@ fun regex(in: string::string):regex {
}
obj regexState (Object) {
var character: char
// if only one character, both are the same
var characterBegin: char
var characterEnd: char
var next_states: set::set<*regexState>
fun construct(charIn:char): *regexState {
character = charIn
return construct(charIn, charIn)
}
fun construct(charFirst:char, charSecond:char): *regexState {
characterBegin = charFirst
characterEnd = charSecond
next_states.construct()
return this
}
@@ -26,17 +34,18 @@ obj regexState (Object) {
return construct((0) cast char)
}
fun copy_construct(old:*regexState): void {
character = old->character
characterBegin = old->characterBegin
characterEnd = old->characterEnd
next_states.copy_construct(&old->next_states)
}
fun destruct():void {
next_states.destruct()
}
fun match_char(input: char): set::set<*regexState> {
return next_states.filter(fun(it:*regexState):bool { return it->character == input; })
return next_states.filter(fun(it:*regexState):bool { return it->characterBegin <= input && input <= it->characterEnd; })
}
fun is_end():bool {
return next_states.any_true(fun(state: *regexState):bool { return state->character == 1; })
return next_states.any_true(fun(state: *regexState):bool { return state->characterBegin == 1; })
}
}
@@ -67,14 +76,6 @@ obj regex (Object, Serializable) {
begin = old->begin
referenceCounter = old->referenceCounter
*referenceCounter += 1
/*construct(old->regexString)*/
/*begin = mem::safe_recursive_clone(old->begin, fun(it: *regexState, cloner: fun(*regexState):*regexState, register: fun(*regexState):void): void {*/
/*var newOne = mem::new<regexState>()->construct(it->character)*/
/*register(newOne)*/
/*it->next_states.for_each(fun(next_state: *regexState) {*/
/*newOne->next_states.add(cloner(next_state))*/
/*})*/
/*})*/
}
fun destruct():void {
@@ -105,6 +106,7 @@ obj regex (Object, Serializable) {
}
fun compile(regex_string: string::string): util::pair<*regexState, set::set<*regexState>> {
/*io::println(regex_string)*/
var first = mem::new<regexState>()->construct()
var previous_begin = set::set<*regexState>()
var previous_end = set::set<*regexState>()
@@ -132,11 +134,14 @@ obj regex (Object, Serializable) {
} else if (regex_string[i] == '(' && !escapeing) {
// note that we don't have a ')' case, as we skip past it with our indicies
var perenEnd = i + 1
for (var depth = 1; depth > 0; perenEnd++;)
for (var depth = 1; depth > 0; perenEnd++;) {
if (perenEnd >= regex_string.length())
ast_transformation::error(string::string("can't find matching peren in: ") + regex_string)
if (regex_string[perenEnd] == '(')
depth++
else if (regex_string[perenEnd] == ')')
depth--
}
var innerBeginEnd = compile(regex_string.slice(i+1, perenEnd-1))
// NOTE: perenEnd is one past the close peren
i = perenEnd-1
@@ -158,7 +163,13 @@ obj regex (Object, Serializable) {
escapeing = true
} else {
var next = mem::new<regexState>()->construct(regex_string[i])
var next: *regexState
if (regex_string[i] == '[' && !escapeing) {
next = mem::new<regexState>()->construct(regex_string[i+1], regex_string[i+3])
i += 4 // [a-b] is 5, i++ adds one
} else {
next = mem::new<regexState>()->construct(regex_string[i])
}
if (alternating) {
previous_end.for_each(fun(it: *regexState):void { it->next_states.add(next); })
current_begin.add(next)