v/compiler/scanner.v

742 lines
16 KiB
Go
Raw Normal View History

2019-06-23 04:21:30 +02:00
// Copyright (c) 2019 Alexander Medvednikov. All rights reserved.
// Use of this source code is governed by an MIT license
// that can be found in the LICENSE file.
2019-06-22 20:20:28 +02:00
module main
struct Scanner {
mut:
file_path string
text string
pos int
line_nr int
inside_string bool
dollar_start bool // for hacky string interpolation TODO simplify
dollar_end bool
debug bool
line_comment string
started bool
is_fmt bool
// vfmt fields
fmt_out strings.Builder
2019-06-22 20:20:28 +02:00
fmt_indent int
fmt_line_empty bool
}
const (
SINGLE_QUOTE = `\'`
2019-06-30 14:35:26 +02:00
//QUOTE = `"`
2019-06-22 20:20:28 +02:00
)
fn new_scanner(file_path string) *Scanner {
if !os.file_exists(file_path) {
2019-06-23 02:02:33 +02:00
panic('"$file_path" doesn\'t exist')
2019-06-22 20:20:28 +02:00
}
//text := os.read_file(file_path)
text := os.read_file(file_path) or {
panic('scanner: failed to open "$file_path"')
return &Scanner{}
}
2019-06-22 20:20:28 +02:00
scanner := &Scanner {
file_path: file_path
text: text
fmt_out: strings.new_builder(1000)
2019-06-22 20:20:28 +02:00
}
// println('new scanner "$file_path" txt.len=$scanner.text.len')
return scanner
}
// TODO remove once multiple return values are implemented
struct ScanRes {
tok Token
lit string
}
fn scan_res(tok Token, lit string) ScanRes {
return ScanRes{tok, lit}
}
fn is_white(c byte) bool {
return c.is_white()
}
fn is_nl(c byte) bool {
2019-06-26 13:54:03 +02:00
return c == `\r` || c == `\n`
2019-06-22 20:20:28 +02:00
}
fn (s mut Scanner) ident_name() string {
start := s.pos
for {
s.pos++
if s.pos >= s.text.len {
break
}
2019-06-22 20:20:28 +02:00
c := s.text[s.pos]
if !is_name_char(c) && !c.is_digit() {
break
}
}
name := s.text.substr(start, s.pos)
s.pos--
return name
}
fn (s mut Scanner) ident_number() string {
start := s.pos
is_hex := s.pos + 1 < s.text.len && s.text[s.pos] == `0` && s.text[s.pos + 1] == `x`
2019-06-22 20:20:28 +02:00
is_oct := !is_hex && s.text[s.pos] == `0`
mut is_float := false
for {
s.pos++
if s.pos >= s.text.len {
break
}
2019-06-22 20:20:28 +02:00
c := s.text[s.pos]
if c == `.` {
is_float = true
}
2019-07-02 14:45:27 +02:00
is_good_hex := is_hex && (c == `x` || (c >= `a` && c <= `f`) || (c >= `A` && c <= `F`))
2019-06-22 20:20:28 +02:00
// 1e+3, 1e-3, 1e3
if !is_hex && c == `e` && s.pos + 1 < s.text.len {
2019-06-22 20:20:28 +02:00
next := s.text[s.pos + 1]
if next == `+` || next == `-` || next.is_digit() {
s.pos++
continue
}
}
if !c.is_digit() && c != `.` && !is_good_hex {
break
}
// 1..9
if c == `.` && s.pos + 1 < s.text.len && s.text[s.pos + 1] == `.` {
2019-06-22 20:20:28 +02:00
break
}
if is_oct && c >= `8` && !is_float {
s.error('malformed octal constant')
}
}
number := s.text.substr(start, s.pos)
s.pos--
return number
}
fn (s Scanner) has_gone_over_line_end() bool {
mut i := s.pos-1
for i >= 0 && !is_white(s.text[i]) {
i--
}
for i >= 0 && is_white(s.text[i]) {
if is_nl(s.text[i]) {
return true
}
i--
}
return false
}
2019-06-22 20:20:28 +02:00
fn (s mut Scanner) skip_whitespace() {
for s.pos < s.text.len && is_white(s.text[s.pos]) {
if is_nl(s.text[s.pos]) {
2019-07-05 22:07:44 +02:00
// Count \r\n as one line
2019-07-05 22:12:06 +02:00
if !(s.text[s.pos] == `\n` && s.pos > 0 && s.text[s.pos-1] == `\r`) {
2019-07-05 22:07:44 +02:00
s.line_nr++
}
2019-06-22 20:20:28 +02:00
if s.is_fmt {
return
}
}
s.pos++
}
// if s.pos == s.text.len {
// return scan_res(.eof, '')
2019-06-22 20:20:28 +02:00
// }
}
fn (s mut Scanner) get_var_name(pos int) string {
mut pos_start := pos
for ; pos_start >= 0 && s.text[pos_start] != `\n` && s.text[pos_start] != `;`; pos_start-- {}
pos_start++
return s.text.substr(pos_start, pos)
}
// CAO stands for Compound Assignment Operators (e.g '+=' )
fn (s mut Scanner) cao_change(operator string) {
s.text = s.text.substr(0, s.pos - operator.len) + ' = ' + s.get_var_name(s.pos - operator.len) + ' ' + operator + ' ' + s.text.substr(s.pos + 1, s.text.len)
}
2019-07-06 15:33:15 +02:00
fn (s mut Scanner) scan() ScanRes {
2019-06-22 20:20:28 +02:00
// if s.file_path == 'd.v' {
// println('\nscan()')
// }
2019-07-01 18:07:22 +02:00
if s.started {
2019-06-22 20:20:28 +02:00
// || (s.pos == 0 && s.text.len > 0 && s.text[s.pos] == `\n`) {
s.pos++
}
s.started = true
if s.pos >= s.text.len {
return scan_res(.eof, '')
2019-06-22 20:20:28 +02:00
}
// skip whitespace
if !s.inside_string {
s.skip_whitespace()
}
if s.is_fmt && s.text[s.pos] == `\n` {
return scan_res(.nl, '')
2019-06-22 20:20:28 +02:00
}
// End of $var, start next string
if !s.is_fmt && s.dollar_end {
// fmt.Println("end of $var, get string", s.pos, string(s.text[s.pos]))
if s.text[s.pos] == SINGLE_QUOTE {
// fmt.Println("ENDDD")
s.dollar_end = false
return scan_res(.strtoken, '')
2019-06-22 20:20:28 +02:00
}
s.dollar_end = false
return scan_res(.strtoken, s.ident_string())
2019-06-22 20:20:28 +02:00
}
s.skip_whitespace()
// end of file
if s.pos >= s.text.len {
// println('scan(): returning .eof (pos >= len)')
return scan_res(.eof, '')
2019-06-22 20:20:28 +02:00
}
// handle each char
c := s.text[s.pos]
mut nextc := `\0`
if s.pos + 1 < s.text.len {
nextc = s.text[s.pos + 1]
}
// name or keyword
if is_name_char(c) {
name := s.ident_name()
// tmp hack to detect . in ${}
// Check if not .eof to prevent panic
next_char := if s.pos + 1 < s.text.len { s.text[s.pos + 1] } else { `\0` }
2019-06-22 20:20:28 +02:00
// println('!!! got name=$name next_char=$next_char')
if is_key(name) {
// println('IS KEY')
// tok := (key_to_token(name))
// println(tok.str())
return scan_res(key_to_token(name), '')
}
// 'asdf $b' => "b" is the last name in the string, dont start parsing string
// at the next ', skip it
if s.inside_string {
// println('is_letter inside string! nextc=${nextc.str()}')
if next_char == SINGLE_QUOTE {
2019-06-22 20:20:28 +02:00
// println('var is last before QUOTE')
s.pos++
s.dollar_start = false
s.inside_string = false
}
}
if s.dollar_start && next_char != `.` {
// println('INSIDE .strtoken .dollar var=$name')
2019-06-22 20:20:28 +02:00
s.dollar_end = true
s.dollar_start = false
}
if s.pos == 0 && next_char == ` ` {
s.pos++
//If a single letter name at the start of the file, increment
//Otherwise the scanner would be stuck at s.pos = 0
}
return scan_res(.name, name)
2019-06-22 20:20:28 +02:00
}
// number, `.123`
else if c.is_digit() || c == `.` && nextc.is_digit() {
num := s.ident_number()
return scan_res(.integer, num)
2019-06-22 20:20:28 +02:00
}
// all other tokens
switch c {
case `+`:
if nextc == `+` {
s.pos++
return scan_res(.inc, '')
2019-06-22 20:20:28 +02:00
}
else if nextc == `=` {
s.pos++
return scan_res(.plus_assign, '')
2019-06-22 20:20:28 +02:00
}
return scan_res(.plus, '')
2019-06-22 20:20:28 +02:00
case `-`:
if nextc == `-` {
s.pos++
return scan_res(.dec, '')
2019-06-22 20:20:28 +02:00
}
else if nextc == `=` {
s.pos++
return scan_res(.minus_assign, '')
2019-06-22 20:20:28 +02:00
}
return scan_res(.minus, '')
2019-06-22 20:20:28 +02:00
case `*`:
if nextc == `=` {
s.pos++
return scan_res(.mult_assign, '')
2019-06-22 20:20:28 +02:00
}
return scan_res(.mul, '')
2019-06-22 20:20:28 +02:00
case `^`:
if nextc == `=` {
s.pos++
return scan_res(.xor_assign, '')
2019-06-22 20:20:28 +02:00
}
return scan_res(.xor, '')
2019-06-22 20:20:28 +02:00
case `%`:
if nextc == `=` {
s.pos++
return scan_res(.mod_assign, '')
2019-06-22 20:20:28 +02:00
}
return scan_res(.mod, '')
2019-06-22 20:20:28 +02:00
case `?`:
return scan_res(.question, '')
2019-06-22 20:20:28 +02:00
case SINGLE_QUOTE:
return scan_res(.strtoken, s.ident_string())
2019-06-22 20:20:28 +02:00
// TODO allow double quotes
// case QUOTE:
// return scan_res(.strtoken, s.ident_string())
2019-06-22 20:20:28 +02:00
case `\``:
return scan_res(.chartoken, s.ident_char())
2019-06-22 20:20:28 +02:00
case `(`:
return scan_res(.lpar, '')
2019-06-22 20:20:28 +02:00
case `)`:
return scan_res(.rpar, '')
2019-06-22 20:20:28 +02:00
case `[`:
return scan_res(.lsbr, '')
2019-06-22 20:20:28 +02:00
case `]`:
return scan_res(.rsbr, '')
2019-06-22 20:20:28 +02:00
case `{`:
// Skip { in ${ in strings
if s.inside_string {
return s.scan()
}
return scan_res(.lcbr, '')
2019-06-22 20:20:28 +02:00
case `$`:
return scan_res(.dollar, '')
2019-06-22 20:20:28 +02:00
case `}`:
// s = `hello $name kek`
// s = `hello ${name} kek`
if s.inside_string {
s.pos++
// TODO UN.neEDED?
2019-06-22 20:20:28 +02:00
if s.text[s.pos] == SINGLE_QUOTE {
s.inside_string = false
return scan_res(.strtoken, '')
2019-06-22 20:20:28 +02:00
}
return scan_res(.strtoken, s.ident_string())
2019-06-22 20:20:28 +02:00
}
else {
return scan_res(.rcbr, '')
2019-06-22 20:20:28 +02:00
}
case `&`:
if nextc == `=` {
s.pos++
return scan_res(.and_assign, '')
2019-06-22 20:20:28 +02:00
}
if nextc == `&` {
2019-06-22 20:20:28 +02:00
s.pos++
return scan_res(.and, '')
2019-06-22 20:20:28 +02:00
}
return scan_res(.amp, '')
2019-06-22 20:20:28 +02:00
case `|`:
if nextc == `|` {
2019-06-22 20:20:28 +02:00
s.pos++
return scan_res(.ortok, '')
2019-06-22 20:20:28 +02:00
}
if nextc == `=` {
s.pos++
return scan_res(.or_assign, '')
2019-06-22 20:20:28 +02:00
}
return scan_res(.pipe, '')
2019-06-22 20:20:28 +02:00
case `,`:
return scan_res(.comma, '')
2019-06-23 05:21:22 +02:00
case `\r`:
if nextc == `\n` {
s.pos++
return scan_res(.nl, '')
2019-06-23 05:21:22 +02:00
}
2019-06-22 20:20:28 +02:00
case `\n`:
return scan_res(.nl, '')
2019-06-22 20:20:28 +02:00
case `.`:
if nextc == `.` {
2019-06-22 20:20:28 +02:00
s.pos++
return scan_res(.dotdot, '')
2019-06-22 20:20:28 +02:00
}
return scan_res(.dot, '')
2019-06-22 20:20:28 +02:00
case `#`:
start := s.pos + 1
for s.pos < s.text.len && s.text[s.pos] != `\n` {
2019-06-22 20:20:28 +02:00
s.pos++
}
s.line_nr++
hash := s.text.substr(start, s.pos)
if s.is_fmt {
// fmt needs NL after #
s.pos--
}
return scan_res(.hash, hash.trim_space())
2019-06-22 20:20:28 +02:00
case `>`:
if nextc == `=` {
2019-06-22 20:20:28 +02:00
s.pos++
return scan_res(.ge, '')
2019-06-22 20:20:28 +02:00
}
else if nextc == `>` {
if s.pos + 2 < s.text.len && s.text[s.pos + 2] == `=` {
2019-06-22 20:20:28 +02:00
s.pos += 2
return scan_res(.righ_shift_assign, '')
2019-06-22 20:20:28 +02:00
}
s.pos++
return scan_res(.righ_shift, '')
2019-06-22 20:20:28 +02:00
}
else {
return scan_res(.gt, '')
2019-06-22 20:20:28 +02:00
}
case `<`:
if nextc == `=` {
2019-06-22 20:20:28 +02:00
s.pos++
return scan_res(.le, '')
2019-06-22 20:20:28 +02:00
}
else if nextc == `<` {
if s.pos + 2 < s.text.len && s.text[s.pos + 2] == `=` {
2019-06-22 20:20:28 +02:00
s.pos += 2
return scan_res(.left_shift_assign, '')
2019-06-22 20:20:28 +02:00
}
s.pos++
return scan_res(.left_shift, '')
2019-06-22 20:20:28 +02:00
}
else {
return scan_res(.lt, '')
2019-06-22 20:20:28 +02:00
}
case `=`:
if nextc == `=` {
2019-06-22 20:20:28 +02:00
s.pos++
return scan_res(.eq, '')
2019-06-22 20:20:28 +02:00
}
else {
return scan_res(.assign, '')
2019-06-22 20:20:28 +02:00
}
case `:`:
if nextc == `=` {
2019-06-22 20:20:28 +02:00
s.pos++
return scan_res(.decl_assign, '')
2019-06-22 20:20:28 +02:00
}
else {
return scan_res(.colon, '')
2019-06-22 20:20:28 +02:00
}
case `;`:
return scan_res(.semicolon, '')
2019-06-22 20:20:28 +02:00
case `!`:
if nextc == `=` {
2019-06-22 20:20:28 +02:00
s.pos++
return scan_res(.ne, '')
2019-06-22 20:20:28 +02:00
}
else {
return scan_res(.not, '')
2019-06-22 20:20:28 +02:00
}
case `~`:
return scan_res(.bit_not, '')
2019-06-22 20:20:28 +02:00
case `/`:
if nextc == `=` {
s.pos++
return scan_res(.div_assign, '')
2019-06-22 20:20:28 +02:00
}
if nextc == `/` {
// debug("!!!!!!.key_goT LI.ne COM")
2019-06-22 20:20:28 +02:00
start := s.pos + 1
for s.pos < s.text.len && s.text[s.pos] != `\n`{
2019-06-22 20:20:28 +02:00
s.pos++
}
s.line_nr++
s.line_comment = s.text.substr(start + 1, s.pos)
s.line_comment = s.line_comment.trim_space()
s.fgenln('// $s.line_comment')
if s.is_fmt {
// fmt needs NL after comment
s.pos--
}
else {
// Skip comment
return s.scan()
}
return scan_res(.line_com, s.line_comment)
2019-06-22 20:20:28 +02:00
}
// Multiline comments
if nextc == `*` {
2019-06-22 20:20:28 +02:00
start := s.pos
mut nest_count := 1
2019-06-22 20:20:28 +02:00
// Skip comment
for nest_count > 0 {
2019-06-22 20:20:28 +02:00
s.pos++
if s.pos >= s.text.len {
s.line_nr--
s.error('comment not terminated')
}
if s.text[s.pos] == `\n` {
s.line_nr++
continue
}
if s.text[s.pos] == `/` && s.text[s.pos + 1] == `*` {
nest_count++
continue
}
if s.text[s.pos] == `*` && s.text[s.pos + 1] == `/` {
nest_count--
2019-06-22 20:20:28 +02:00
}
}
s.pos++
end := s.pos + 1
comm := s.text.substr(start, end)
s.fgenln(comm)
if s.is_fmt {
return scan_res(.mline_com, comm)
2019-06-22 20:20:28 +02:00
}
// Skip if not in fmt mode
return s.scan()
}
return scan_res(.div, '')
2019-06-22 20:20:28 +02:00
}
2019-06-29 17:58:20 +02:00
$if windows {
if c == `\0` {
return scan_res(.eof, '')
2019-06-29 17:58:20 +02:00
}
}
2019-06-23 10:01:55 +02:00
println('(char code=$c) pos=$s.pos len=$s.text.len')
2019-07-03 13:20:43 +02:00
mut msg := 'invalid character `${c.str()}`'
if c == `"` {
msg += ', use \' to denote strings'
}
s.error(msg)
return scan_res(.eof, '')
2019-06-22 20:20:28 +02:00
}
fn (s &Scanner) error(msg string) {
file := s.file_path.all_after('/')
2019-06-23 10:01:55 +02:00
println('panic: $file:${s.line_nr + 1}')
println(msg)
2019-06-22 20:20:28 +02:00
// os.print_backtrace()
// println(file)
// println(s.file_path)
exit(1)
2019-06-22 20:20:28 +02:00
}
2019-06-23 10:01:55 +02:00
// println('array out of bounds $idx len=$a.len')
2019-06-22 20:20:28 +02:00
// This is really bad. It needs a major clean up
fn (s mut Scanner) ident_string() string {
// println("\nidentString() at char=", string(s.text[s.pos]),
// "chard=", s.text[s.pos], " pos=", s.pos, "txt=", s.text[s.pos:s.pos+7])
debug := s.file_path.contains('test_test')
if debug {
println('identStr() $s.file_path line=$s.line_nr pos=$s.pos')
}
mut start := s.pos
s.inside_string = false
slash := `\\`
for {
s.pos++
if s.pos >= s.text.len {
break
}
c := s.text[s.pos]
if debug {
println(c.str())
}
prevc := s.text[s.pos - 1]
// end of string
if c == SINGLE_QUOTE && (prevc != slash || (prevc == slash && s.text[s.pos - 2] == slash)) {
// handle '123\\' slash at the end
break
}
if c == `\n` {
s.line_nr++
}
// Don't allow \0
if c == `0` && s.pos > 2 && s.text[s.pos - 1] == `\\` {
s.error('0 character in a string literal')
}
// Don't allow \x00
if c == `0` && s.pos > 5 && s.text[s.pos - 1] == `0` && s.text[s.pos - 2] == `x` &&
s.text[s.pos - 3] == `\\` {
s.error('0 character in a string literal')
}
// ${var}
if !s.is_fmt && c == `{` && prevc == `$` {
s.inside_string = true
// fmt.Println("breaking out of is()")
// so that s.pos points to $ at the next step
s.pos -= 2
// fmt.Println("break pos=", s.pos, "c=", string(s.text[s.pos]), "d=", s.text[s.pos])
break
}
// $var
// if !s.is_fmt && c != `{` && c != ` ` && ! (c >= `0` && c <= `9`) && prevc == `$` {
if !s.is_fmt && (c.is_letter() || c == `_`) && prevc == `$` {
s.inside_string = true
s.dollar_start = true
// println('setting s.dollar=true pos=$s.pos')
s.pos -= 2
break
}
}
mut lit := ''
if s.text[start] == SINGLE_QUOTE {
start++
}
mut end := s.pos
if s.inside_string {
end++
}
if start > s.pos{}
else {
lit = s.text.substr(start, end)
}
// if lit.contains('\n') {
// println('\nstring lit="$lit" pos=$s.pos line=$s.line_nr')
// }
/*
for c in lit {
if s.file_path.contains('range_test') {
println('!')
println(c)
}
}
*/
return lit
}
fn (s mut Scanner) ident_char() string {
start := s.pos
slash := `\\`
mut len := 0
for {
s.pos++
if s.pos >= s.text.len {
break
}
if s.text[s.pos] != slash {
len++
}
double_slash := s.text[s.pos - 1] == slash && s.text[s.pos - 2] == slash
if s.text[s.pos] == `\`` && (s.text[s.pos - 1] != slash || double_slash) {
if double_slash {
len++
}
break
}
}
len--
c := s.text.substr(start + 1, s.pos)
if len != 1 {
2019-07-07 21:46:21 +02:00
u := c.ustring()
if u.len != 1 {
2019-06-22 20:20:28 +02:00
s.error('invalid character literal (more than one character: $len)')
2019-07-07 21:46:21 +02:00
}
2019-06-22 20:20:28 +02:00
}
return c
}
fn (p mut Parser) peek() Token {
for {
tok := p.scanner.peek()
if tok != .nl {
2019-06-22 20:20:28 +02:00
return tok
}
}
}
fn (s mut Scanner) peek() Token {
pos := s.pos
line := s.line_nr
inside_string := s.inside_string
dollar_start := s.dollar_start
dollar_end := s.dollar_end
// /////
res := s.scan()
tok := res.tok
s.pos = pos
s.line_nr = line
s.inside_string = inside_string
s.dollar_start = dollar_start
s.dollar_end = dollar_end
return tok
}
fn (s mut Scanner) debug_tokens() {
s.pos = 0
fname := s.file_path.all_after('/')
println('\n===DEBUG TOKENS $fname ============')
// allToks := ''
s.debug = true
for {
res := s.scan()
tok := res.tok
lit := res.lit
// printiln(tok)
print(tok.str())
// allToks += tok.String()
if lit != '' {
println(' `$lit`')
// allToks += " `" + lit + "`"
}
else {
println('')
}
// allToks += "\n"
if tok == .eof {
2019-06-22 20:20:28 +02:00
println('============ END OF DEBUG TOKENS ==================')
// fmt.Println("========"+s.file+"========\n", allToks)
break
}
}
}
fn is_name_char(c byte) bool {
return c.is_letter() || c == `_`
}
fn (s mut Scanner) get_opening_bracket() int {
mut pos := s.pos
mut parentheses := 0
mut inside_string := false
for pos > 0 && s.text[pos] != `\n` {
if s.text[pos] == `)` && !inside_string {
parentheses++
}
if s.text[pos] == `(` && !inside_string {
parentheses--
}
if s.text[pos] == `\'` && s.text[pos - 1] != `\\` && s.text[pos - 1] != `\`` {
inside_string = !inside_string
}
if parentheses == 0 {
break
}
pos--
}
return pos
}
2019-06-27 01:55:37 +02:00
// Foo { bar: 3, baz: 'hi' } => '{ bar: 3, baz: "hi" }'
fn (s mut Scanner) create_type_string(T Type, name string) {
line := s.line_nr
inside_string := s.inside_string
mut newtext := '\'{ '
start := s.get_opening_bracket() + 1
end := s.pos
for i, field in T.fields {
if i != 0 {
newtext += ', '
}
newtext += '$field.name: ' + '$${name}.${field.name}'
}
newtext += ' }\''
s.text = s.text.substr(0, start) + newtext + s.text.substr(end, s.text.len)
s.pos = start - 2
s.line_nr = line
s.inside_string = inside_string
}
fn (p mut Parser) create_type_string(T Type, name string) {
p.scanner.create_type_string(T, name)
2019-06-27 01:55:37 +02:00
}