475 lines
12 KiB
Odin
475 lines
12 KiB
Odin
package ltx
|
|
|
|
import "base:runtime"
|
|
import "core:fmt"
|
|
import "core:os"
|
|
import "core:path/filepath"
|
|
import "core:strings"
|
|
import "core:unicode"
|
|
// TODO(Paul): change the key-value format to allow for true arbitrary string values with the use of quotes (maybe?) and make the rules clean and easy to remember.
|
|
|
|
main :: proc() {
|
|
args := parse_args()
|
|
if args.type == .None do return
|
|
|
|
ltx: Ltx
|
|
|
|
parse_err := parse_file(<x, args.file_path)
|
|
defer free_file(<x)
|
|
if parse_err != .None {
|
|
fmt.eprintln(get_error(<x, parse_err, context.temp_allocator))
|
|
return
|
|
}
|
|
|
|
sb := strings.builder_make()
|
|
defer strings.builder_destroy(&sb)
|
|
|
|
switch args.type {
|
|
case .None:
|
|
return
|
|
case .XML:
|
|
to_xml(&sb, ltx.nodes)
|
|
case .PlainText:
|
|
to_plain_text(&sb, ltx.nodes)
|
|
case .SmartIndentedText:
|
|
to_text_with_smart_indent(&sb, ltx.nodes)
|
|
}
|
|
|
|
fmt.println(strings.to_string(sb))
|
|
|
|
// for node in ltx.nodes {
|
|
// print_node(node)
|
|
// }
|
|
|
|
// sb := strings.builder_make()
|
|
// defer strings.builder_destroy(&sb)
|
|
|
|
// fmt.println("--- BEGIN XML ---")
|
|
// strings.builder_reset(&sb)
|
|
// to_xml(&sb, ltx.nodes)
|
|
// fmt.println(strings.to_string(sb))
|
|
// fmt.println("--- END XML ---")
|
|
|
|
// fmt.println("--- BEGIN PLAIN TEXT ---")
|
|
// strings.builder_reset(&sb)
|
|
// to_text_with_smart_indent(&sb, ltx.nodes)
|
|
// fmt.println(strings.to_string(sb))
|
|
// fmt.println("--- END PLAIN TEXT ---")
|
|
}
|
|
|
|
print_help :: proc() {
|
|
fmt.eprintln(
|
|
"A naive implementation of TeX-based arbitrary markup. Prints syntax tree and space-preserving XML substitution",
|
|
)
|
|
fmt.eprintln("Usage:")
|
|
fmt.eprintf("\t%s <filepath>\n", os.args[0])
|
|
}
|
|
|
|
parse_args :: proc() -> Cli_Args {
|
|
args := Cli_Args{}
|
|
|
|
if len(os.args) < 2 {
|
|
print_help()
|
|
return args
|
|
}
|
|
|
|
output_flags_used := 0
|
|
output_type := Cli_Output_Type.None
|
|
file_path: string
|
|
|
|
for arg in os.args[1:] {
|
|
if strings.starts_with(arg, "--") {
|
|
output_flags_used += 1
|
|
switch arg[2:] {
|
|
case "xml":
|
|
output_type = .XML
|
|
case "text":
|
|
output_type = .PlainText
|
|
case "smart":
|
|
output_type = .SmartIndentedText
|
|
case:
|
|
fmt.eprintln("Unknown flag:", arg)
|
|
print_help()
|
|
return args
|
|
}
|
|
} else {
|
|
if file_path != "" {
|
|
fmt.eprintln("Unknown argument:", arg)
|
|
print_help()
|
|
return args
|
|
}
|
|
file_path = arg
|
|
}
|
|
}
|
|
|
|
if output_flags_used > 1 {
|
|
fmt.eprintln("Too many flags passed")
|
|
print_help()
|
|
} else {
|
|
args.file_path = file_path
|
|
args.type = output_flags_used == 0 ? .PlainText : output_type
|
|
}
|
|
|
|
return args
|
|
}
|
|
|
|
seek :: proc(ltx: ^Ltx) -> (rune, Ltx_Error) {
|
|
next_idx := ltx.idx + 1
|
|
if next_idx >= len(ltx.source) do return 0, .EOF
|
|
if current_char(ltx) == '\n' {
|
|
ltx.pos = {
|
|
col = 0,
|
|
line = ltx.pos.line + 1,
|
|
}
|
|
} else {
|
|
ltx.pos.col += 1
|
|
}
|
|
ltx.idx = next_idx
|
|
c := current_char(ltx)
|
|
return c, .None
|
|
}
|
|
|
|
has_next :: proc(ltx: ^Ltx) -> b32 {
|
|
return len(ltx.source) > ltx.idx + 1
|
|
}
|
|
|
|
peek :: proc(ltx: ^Ltx) -> (rune, Ltx_Error) {
|
|
if !has_next(ltx) do return 0, .EOF
|
|
return rune(ltx.source[ltx.idx + 1]), .None
|
|
}
|
|
|
|
current_char :: proc(ltx: ^Ltx) -> rune {
|
|
assert(ltx.idx < len(ltx.source), "index cannot be greater than string length")
|
|
return rune(ltx.source[ltx.idx])
|
|
}
|
|
|
|
consume_whitespace :: proc(ltx: ^Ltx) -> Ltx_Error {
|
|
for unicode.is_white_space(current_char(ltx)) do seek(ltx) or_return // TODO: do prop error handling
|
|
return .None
|
|
}
|
|
|
|
validate_key :: proc(key: string) -> Ltx_Error {
|
|
assert(len(key) > 0, "expected non-empty key")
|
|
if !unicode.is_alpha(rune(key[0])) do return .InvalidKey
|
|
for c in key {
|
|
if !unicode.is_alpha(c) && !unicode.is_number(c) && c != '_' do return .InvalidKey
|
|
}
|
|
return .None
|
|
}
|
|
|
|
parse :: proc(ltx: ^Ltx) -> (err: Ltx_Error) {
|
|
ltx.idx = 0
|
|
stack := make([dynamic]Node)
|
|
defer delete(stack)
|
|
|
|
text_node_start_idx := 0
|
|
for has_next(ltx) {
|
|
if current_char(ltx) == TokenArray[.Backslash] {
|
|
// TODO: cehck if the char after slash is a token
|
|
if text_node_start_idx < ltx.idx {
|
|
node := Node {
|
|
kind = .Text,
|
|
text = ltx.source[text_node_start_idx:ltx.idx],
|
|
}
|
|
if len(stack) == 0 do append(<x.nodes, node)
|
|
else do append(&stack[len(stack) - 1].children, node)
|
|
}
|
|
seek(ltx) or_return // skip \
|
|
|
|
tag_start_idx := ltx.idx
|
|
for unicode.is_letter(current_char(ltx)) do seek(ltx) or_return
|
|
assert(tag_start_idx < ltx.idx)
|
|
tag_name := ltx.source[tag_start_idx:ltx.idx]
|
|
node := Node {
|
|
kind = .Tag,
|
|
name = tag_name,
|
|
}
|
|
end_pos: struct {
|
|
tag_name: int,
|
|
fields: int,
|
|
content: int,
|
|
} = {
|
|
tag_name = ltx.idx,
|
|
fields = 0,
|
|
content = 0,
|
|
}
|
|
|
|
consume_whitespace(ltx) or_return
|
|
for current_char(ltx) == TokenArray[.LeftBracket] {
|
|
c := seek(ltx) or_return
|
|
field_start := ltx.idx
|
|
key: string
|
|
for has_next(ltx) {
|
|
if c == TokenArray[.Assign] {
|
|
if field_start >= ltx.idx do return .KeyExpected
|
|
key = strings.trim(ltx.source[field_start:ltx.idx], " \t")
|
|
if len(key) <= 0 do return .KeyExpected
|
|
validate_key(key) or_return
|
|
seek(ltx) or_return
|
|
field_start = ltx.idx
|
|
} else if c == TokenArray[.RightBracket] {
|
|
if field_start >= ltx.idx do return .ValueExpected
|
|
raw_value := strings.trim(ltx.source[field_start:ltx.idx], " \t")
|
|
if len(raw_value) <= 0 do return .ValueExpected
|
|
value := Field{}
|
|
if key != "" {
|
|
value.type = .Attribute
|
|
value.value = raw_value
|
|
} else {
|
|
key = raw_value
|
|
value.type = .Flag
|
|
}
|
|
_, _, found := map_upsert(&node.attributes, key, value)
|
|
if found do return .KeyAlreadyExists
|
|
seek(ltx) or_return
|
|
end_pos.fields = ltx.idx
|
|
break
|
|
}
|
|
c = seek(ltx) or_return
|
|
}
|
|
consume_whitespace(ltx) or_return
|
|
}
|
|
consume_whitespace(ltx) or_return
|
|
if current_char(ltx) == TokenArray[.LeftBrace] {
|
|
seek(ltx) or_return // consume {
|
|
append(&stack, node)
|
|
end_pos.content = ltx.idx
|
|
} else {
|
|
append(<x.nodes, node)
|
|
}
|
|
if end_pos.content != 0 do text_node_start_idx = end_pos.content
|
|
else if end_pos.fields != 0 do text_node_start_idx = end_pos.fields
|
|
else do text_node_start_idx = end_pos.tag_name
|
|
} else if current_char(ltx) == TokenArray[.RightBrace] {
|
|
if len(stack) <= 0 do return .UnexpectedRightBrace
|
|
node := pop(&stack)
|
|
if text_node_start_idx < ltx.idx {
|
|
text_node := Node {
|
|
kind = .Text,
|
|
text = ltx.source[text_node_start_idx:ltx.idx],
|
|
}
|
|
append(&node.children, text_node)
|
|
}
|
|
if len(stack) > 0 {
|
|
append(&stack[len(stack) - 1].children, node)
|
|
} else {
|
|
append(<x.nodes, node)
|
|
}
|
|
seek(ltx) or_return // consume }
|
|
text_node_start_idx = ltx.idx
|
|
} else {
|
|
seek(ltx) or_return
|
|
}
|
|
}
|
|
if text_node_start_idx < ltx.idx {
|
|
append(<x.nodes, Node{kind = .Text, text = ltx.source[text_node_start_idx:ltx.idx]})
|
|
}
|
|
if len(stack) > 0 do return .ClosingBraceExpected
|
|
return .None
|
|
}
|
|
|
|
parse_file :: proc(ltx: ^Ltx, file_path: string, allocator := context.allocator) -> Ltx_Error {
|
|
source, err := os.read_entire_file_from_path(file_path, allocator)
|
|
if err != nil do return .CannotReadFile
|
|
abs_path, abs_err := filepath.abs(file_path)
|
|
if abs_err != nil do return .CannotReadFile
|
|
ltx.source_path = abs_path
|
|
ltx.source = string(source)
|
|
return parse(ltx)
|
|
}
|
|
|
|
free_file :: proc(ltx: ^Ltx) -> runtime.Allocator_Error {
|
|
delete(ltx.source) or_return
|
|
delete(ltx.source_path) or_return
|
|
return .None
|
|
}
|
|
|
|
print_indent :: proc(level: int) {
|
|
for i in 0 ..< level {
|
|
fmt.print(" ")
|
|
}
|
|
}
|
|
|
|
print_node :: proc(node: Node, indent_level := 0) {
|
|
print_indent(indent_level)
|
|
switch node.kind {
|
|
case .Text:
|
|
fmt.printf("TEXT: \"%s\"\n", escape_white_space(node.text, context.temp_allocator))
|
|
case .Tag:
|
|
fmt.print("TAG:", node.name)
|
|
if len(node.attributes) > 0 {
|
|
for k, v in node.attributes {
|
|
if v.type == .Flag {
|
|
fmt.printf(" (%s)", k)
|
|
} else {
|
|
fmt.printf(" {{%s: %s}}", k, v.value)
|
|
}
|
|
}
|
|
}
|
|
fmt.println()
|
|
if len(node.children) > 0 {
|
|
for child in node.children {
|
|
print_node(child, indent_level + 1)
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
to_text_with_smart_indent :: proc(sb: ^strings.Builder, nodes: [dynamic]Node, indent_level := 0) {
|
|
should_inc_indent_level := true
|
|
for node in nodes {
|
|
if node.kind == .Text {
|
|
should_inc_indent_level = true
|
|
break
|
|
}
|
|
}
|
|
|
|
for node in nodes {
|
|
switch node.kind {
|
|
case .Text:
|
|
for i in 0 ..< indent_level - 1 do strings.write_rune(sb, ' ')
|
|
text := strings.trim(
|
|
process_white_space(node.text, context.temp_allocator),
|
|
" \t\n\r\v",
|
|
)
|
|
// TODO(Paul): do something about newlines; have them be limited to 2 max if they appear consecutively perhaps
|
|
if text != "" {
|
|
fmt.sbprintln(sb, text)
|
|
}
|
|
case .Tag:
|
|
to_text_with_smart_indent(
|
|
sb,
|
|
node.children,
|
|
should_inc_indent_level ? indent_level + 1 : indent_level,
|
|
)
|
|
}
|
|
}
|
|
}
|
|
|
|
// Ltx to plain text; strip ltx tags
|
|
to_plain_text :: proc(sb: ^strings.Builder, nodes: [dynamic]Node) {
|
|
for node in nodes {
|
|
switch node.kind {
|
|
case .Text:
|
|
fmt.sbprint(sb, node.text)
|
|
case .Tag:
|
|
to_plain_text(sb, node.children)
|
|
}
|
|
}
|
|
}
|
|
|
|
// Ltx to XML
|
|
// TODO: CDATA garbage
|
|
to_xml :: proc(sb: ^strings.Builder, nodes: [dynamic]Node, depth := 0) {
|
|
if len(nodes) <= 0 do return
|
|
if depth == 0 do fmt.sbprintln(sb, "<ltx>")
|
|
for node in nodes {
|
|
switch node.kind {
|
|
case .Text:
|
|
fmt.sbprint(sb, node.text)
|
|
case .Tag:
|
|
fmt.sbprintf(sb, "<%s", node.name)
|
|
for k, v in node.attributes do fmt.sbprintf(sb, " %s=\"%s\"", k, v.value)
|
|
if len(node.children) > 0 {
|
|
fmt.sbprint(sb, ">")
|
|
to_xml(sb, node.children, depth + 1)
|
|
fmt.sbprintf(sb, "</%s>", node.name)
|
|
} else {
|
|
fmt.sbprint(sb, " />")
|
|
}
|
|
}
|
|
}
|
|
if depth == 0 do fmt.sbprint(sb, "\n</ltx>")
|
|
}
|
|
|
|
// Ltx_Error to english string
|
|
error_to_string :: proc(error: Ltx_Error) -> string {
|
|
switch error {
|
|
case .None:
|
|
return ""
|
|
case .EOF:
|
|
return "unexpected end of file"
|
|
case .KeyExpected:
|
|
return "key expected before '='"
|
|
case .ClosingBraceExpected:
|
|
return "closing brace expected"
|
|
case .ClosingBracketExpected:
|
|
return "closing bracket expected"
|
|
case .UnexpectedRightBrace:
|
|
return "unexpected '}'"
|
|
case .ValueExpected:
|
|
return "value expected after '='"
|
|
case .KeyAlreadyExists:
|
|
return "attribute key/flag already exists in attribute"
|
|
case .InvalidKey:
|
|
return "invalid key"
|
|
case .CannotReadFile:
|
|
return "cannot read file"
|
|
}
|
|
return "unknown error"
|
|
}
|
|
|
|
// Pretty format LTX error
|
|
get_error :: proc(ltx: ^Ltx, error: Ltx_Error, allocator := context.allocator) -> string {
|
|
if error == .None do return ""
|
|
file_path := len(ltx.source_path) > 0 ? ltx.source_path : ""
|
|
line := ltx.pos.line + 1
|
|
col := ltx.pos.col + 1
|
|
error_msg := error_to_string(error)
|
|
return fmt.aprintf(
|
|
"%s(%d,%d): error: %s",
|
|
file_path,
|
|
line,
|
|
col,
|
|
error_msg,
|
|
allocator = allocator,
|
|
)
|
|
}
|
|
|
|
// Escapes white space characters with in their c-string slash form
|
|
escape_white_space :: proc(s: string, allocator := context.allocator) -> string {
|
|
sb := strings.builder_make(allocator)
|
|
|
|
for c in s {
|
|
switch c {
|
|
case '\t':
|
|
strings.write_string(&sb, "\\t")
|
|
case '\n':
|
|
strings.write_string(&sb, "\\n")
|
|
case '\r':
|
|
strings.write_string(&sb, "\\r")
|
|
case '\\':
|
|
strings.write_string(&sb, "\\\\")
|
|
case:
|
|
strings.write_rune(&sb, c)
|
|
}
|
|
}
|
|
return strings.to_string(sb)
|
|
}
|
|
|
|
// Processes and normalizes whitespace in a multi-line string.
|
|
process_white_space :: proc(s: string, allocator := context.allocator) -> string {
|
|
sb: strings.Builder = strings.builder_make(allocator)
|
|
lines := strings.split(s, "\n")
|
|
for _line in lines {
|
|
line := strings.trim_right(_line, "\t ")
|
|
if len(line) <= 0 do continue
|
|
last_idx := 0
|
|
last_char: rune = 0
|
|
for r, i in line {
|
|
if r == '\r' do continue
|
|
if unicode.is_space(r) {
|
|
if last_char == r {
|
|
last_idx = i
|
|
} else {
|
|
fmt.sbprint(&sb, line[last_idx:i])
|
|
last_idx = i
|
|
}
|
|
}
|
|
last_char = r
|
|
}
|
|
fmt.sbprintln(&sb, line[last_idx:])
|
|
}
|
|
return strings.to_string(sb)
|
|
}
|