4 Commits

Author SHA1 Message Date
amery cbe503c807 lexer: introduce Error{}
Signed-off-by: Alejandro Mery <amery@jpi.io>
2023-08-29 13:33:30 +00:00
amery 0fb67bebff lexer: introduce a Position (Line, Column) handler
Signed-off-by: Alejandro Mery <amery@jpi.io>
2023-08-29 13:33:30 +00:00
amery 2fd2a0f987 build-sys: import build system from darvaza.org/core
Signed-off-by: Alejandro Mery <amery@jpi.io>
2023-08-29 13:32:08 +00:00
amery 15938090f6 lexer: introduce StateFn and the basic state machine loop
Signed-off-by: Alejandro Mery <amery@jpi.io>
2023-08-29 13:26:49 +00:00
11 changed files with 24 additions and 457 deletions
-77
View File
@@ -1,78 +1 @@
# asciigoat's core library
[![Go Reference][godoc-badge]][godoc]
[![Go Report Card][goreport-badge]][goreport]
This package contains the basics for writing simple parsers of
text languages heavily inspired by
[Rob Pike](https://en.wikipedia.org/wiki/Rob_Pike)'s talk on
[Lexical Scanning in Go](https://go.dev/talks/2011/lex.slide#1) in 2011 which
you can [watch online](https://www.youtube.com/watch?v=HxaD_trXwRE) to get
better understanding of the ideas behind **asciigoat**.
**asciigoat** is [MIT](https://opensource.org/license/mit/) licensed.
[godoc]: https://pkg.go.dev/asciigoat.org/core
[godoc-badge]: https://pkg.go.dev/badge/asciigoat.org/core.svg
[goreport]: https://goreportcard.com/report/asciigoat.org/core
[goreport-badge]: https://goreportcard.com/badge/asciigoat.org/core
[godoc-lexer-reader]: https://pkg.go.dev/asciigoat.org/core/lexer#Reader
[godoc-readcloser]: https://pkg.go.dev/asciigoat.org/core#ReadCloser
## Lexer
### lexer.Reader
The lexer package provides [`lexer.Reader`][godoc-lexer-reader] which is
actually an [`io.RuneScanner`](https://pkg.go.dev/io#RuneScanner)
that buffers accepted runes until you are ready to
[emit](https://pkg.go.dev/asciigoat.org/core/lexer#Reader.Emit) or
[discard](https://pkg.go.dev/asciigoat.org/core/lexer#Reader.Discard).
### lexer.Position
[`lexer.Position`](https://pkg.go.dev/asciigoat.org/core/lexer#Position)
is a `(Line, Column)` pair with methods to facilitate tracking
your position on the source [Reader](https://pkg.go.dev/io#Reader).
### lexer.Error
[`lexer.Error`](https://pkg.go.dev/asciigoat.org/core/lexer#Error)
is an [unwrappable](https://pkg.go.dev/errors#Unwrap) error with a
token position and hint attached.
### lexer.StateFn
At the heart of **asciigoat** we have _state functions_ as proposed on [Rob Pike's famous talk](https://www.youtube.com/watch?v=HxaD_trXwRE) which return the next _state function_ parsing is done.
Additionally there is a [`Run()`](https://pkg.go.dev/asciigoat.org/lexer#Run) helper that implements the loop.
### rune checkers
_Rune checkers_ are simple functions that tell if a rune is of a class or it's not.
Fundamental checkers are provided by the [`unicode` package](https://pkg.go.dev/unicode).
Our [`lexer.Reader`][godoc-lexer-reader] uses them on its `Accept()` and `AcceptAll()` methods to
make it easier to consume the _source_ document.
To facilitate the declaration of _rune classes_ in the context of **asciigoat** powered parsers we include
a series of rune checker factories.
* `NewIsIn(string)`
* `NewIsInRunes(...rune)`
* `NewIsNot(checker)`
* `NewIsOneOf(...checker)`
## Others
### ReadCloser
[ReadCloser][godoc-readcloser] assists in providing a
[io.Closer](https://pkg.go.dev/io#Closer) to Readers or buffers without on,
or unearthing one if available so
[io.ReadCloser](https://pkg.go.dev/io#ReadCloser) can be fulfilled.
## See also
* [asciigoat.org/ini](https://asciigoat.org/ini)
* [oss.jpi.io](https://oss.jpi.io)
+3 -5
View File
@@ -2,22 +2,20 @@ module asciigoat.org/core
go 1.19
require (
github.com/fatih/structtag v1.2.0
github.com/mgechev/revive v1.3.3
github.com/pkg/errors v0.9.1
)
require github.com/mgechev/revive v1.3.3
require (
github.com/BurntSushi/toml v1.3.2 // indirect
github.com/chavacava/garif v0.0.0-20230608123814-4bd63c2919ab // indirect
github.com/fatih/color v1.15.0 // indirect
github.com/fatih/structtag v1.2.0 // indirect
github.com/mattn/go-colorable v0.1.13 // indirect
github.com/mattn/go-isatty v0.0.17 // indirect
github.com/mattn/go-runewidth v0.0.9 // indirect
github.com/mgechev/dots v0.0.0-20210922191527-e955255bf517 // indirect
github.com/mitchellh/go-homedir v1.1.0 // indirect
github.com/olekukonko/tablewriter v0.0.5 // indirect
github.com/pkg/errors v0.9.1 // indirect
golang.org/x/sys v0.11.0 // indirect
golang.org/x/tools v0.12.0 // indirect
)
+5 -30
View File
@@ -1,7 +1,6 @@
package lexer
import (
"errors"
"fmt"
"strings"
)
@@ -10,14 +9,6 @@ var (
_ error = (*Error)(nil)
)
var (
// ErrUnacceptableRune indicates the read rune isn't acceptable in the context
ErrUnacceptableRune = errors.New("rune not acceptable in context")
// ErrNotImplemented indicates something hasn't been implemented yet
ErrNotImplemented = errors.New("not implemented")
)
// Error represents a generic parsing error
type Error struct {
Filename string
@@ -25,29 +16,17 @@ type Error struct {
Column int
Content string
Hint string
Err error
}
func (err Error) prefix() string {
switch {
case err.Line > 0 || err.Column > 0:
if err.Filename != "" {
return fmt.Sprintf("%s:%v:%v", err.Filename, err.Line, err.Column)
}
return fmt.Sprintf("%v:%v", err.Line, err.Column)
default:
return err.Filename
}
}
func (err Error) Error() string {
var s []string
prefix := err.prefix()
if prefix != "" {
s = append(s, prefix)
switch {
case err.Line > 0 || err.Column > 0:
s = append(s, fmt.Sprintf("%s:%v:%v", err.Filename, err.Line, err.Column))
case err.Filename != "":
s = append(s, err.Filename)
}
if err.Err != nil {
@@ -58,10 +37,6 @@ func (err Error) Error() string {
s = append(s, fmt.Sprintf("%q", err.Content))
}
if err.Hint != "" {
s = append(s, err.Hint)
}
return strings.Join(s, ": ")
}
+3 -17
View File
@@ -1,31 +1,17 @@
// Package lexer provides basic helpers to implement parsers
package lexer
import (
"errors"
"io"
)
// StateFn is a State Function of the parser
type StateFn func() (StateFn, error)
// Run runs a state machine until the state function either
// returns nil or an error
func Run(fn StateFn) error {
for fn != nil {
var err error
var err error
for fn != nil && err == nil {
fn, err = fn()
switch {
case errors.Is(err, io.EOF):
// EOF
return nil
case err != nil:
// failed
return err
}
}
// ended
return nil
return err
}
+13 -11
View File
@@ -41,26 +41,28 @@ func (p *Position) Step() {
p.Column++
}
// StepN moves the column N places forward
func (p *Position) StepN(n int) {
// Next returns a new Position one rune forward
// on the line
func (p Position) Next() Position {
if p.Line == 0 {
p.Reset()
}
switch {
case n > 0:
p.Column += n
default:
panic(fmt.Errorf("invalid %v increment", n))
return Position{
Line: p.Line,
Column: p.Column + 1,
}
}
// StepLine moves position to the start of the next line
func (p *Position) StepLine() {
// NextLine returns a new Position at the begining of the next
// line.
func (p Position) NextLine() Position {
if p.Line == 0 {
p.Reset()
}
p.Line++
p.Column = 1
return Position{
Line: p.Line + 1,
Column: 1,
}
}
-47
View File
@@ -1,47 +0,0 @@
package lexer
import (
"strings"
"unicode"
)
// NewIsNot generates a rune condition checker that reverses the
// decision of the given checker.
func NewIsNot(cond func(rune) bool) func(rune) bool {
return func(r rune) bool {
return !cond(r)
}
}
// NewIsIn generates a rune condition checker that accepts runes
// contained on the provided string
func NewIsIn(s string) func(rune) bool {
return func(r rune) bool {
return strings.ContainsRune(s, r)
}
}
// NewIsInRunes generates a rune condition checker that accepts
// the runes specified
func NewIsInRunes(s ...rune) func(rune) bool {
return NewIsIn(string(s))
}
// NewIsOneOf generates a run condition checker that accepts runes
// accepted by any of the given checkers
func NewIsOneOf(s ...func(rune) bool) func(rune) bool {
return func(r rune) bool {
for _, cond := range s {
if cond(r) {
return true
}
}
return false
}
}
// IsSpace reports whether the rune is a space character as
// defined by Unicode's White Space property
func IsSpace(r rune) bool {
return unicode.IsSpace(r)
}
-69
View File
@@ -1,69 +0,0 @@
package reflection
import (
"bytes"
"fmt"
"reflect"
)
// An InvalidUnmarshalError describes an invalid argument passed to New.
// (It must be a non-nil pointer.)
type InvalidUnmarshalError struct {
Method string
Prefix string
reflect.Type
}
func (e *InvalidUnmarshalError) Error() string {
var buf bytes.Buffer
if e.Prefix != "" {
_, _ = fmt.Fprintf(&buf, "%s: ", e.Prefix)
}
method := e.Method
if method == "" {
method = "New"
}
_, _ = fmt.Fprintf(&buf, "%s(", method)
switch {
case e.Type == nil:
_, _ = buf.WriteString("nil")
case e.Type.Kind() != reflect.Pointer:
_, _ = fmt.Fprintf(&buf, "%s %s", "non-pointer", e.Type.String())
default:
_, _ = fmt.Fprintf(&buf, "%s %s", "nil", e.Type.String())
}
_, _ = buf.WriteString(")")
return buf.String()
}
// An UnmarshalTypeError tells something went wrong while processing
// a type.
type UnmarshalTypeError struct {
Prefix string
Err error
reflect.Type
}
func (e UnmarshalTypeError) Unwrap() error {
return e.Err
}
func (e UnmarshalTypeError) Error() string {
var buf bytes.Buffer
if e.Prefix != "" {
_, _ = fmt.Fprintf(&buf, "%s: ", e.Prefix)
}
_, _ = fmt.Fprintf(&buf, "%s: %s", e.Type.String(), e.Err)
return buf.String()
}
-16
View File
@@ -1,16 +0,0 @@
package reflection
import (
"fmt"
"reflect"
)
func structFieldName(sf reflect.StructField) string {
if sf.Anonymous {
idx := sf.Index[len(sf.Index)-1]
return fmt.Sprintf("%s-%v", "anonymous", idx)
}
return sf.Name
}
-40
View File
@@ -1,40 +0,0 @@
// Package reflection helps Marshalling/Unmarshalling data to structs
package reflection
import (
"reflect"
)
// Reflection provides Marshalling/Unmarshalling oriented view
// of a value
type Reflection struct {
v reflect.Value
types map[reflect.Type]TypeInfo
}
// Value returns the object it reflects
func (r *Reflection) Value() any {
return r.v.Interface()
}
// New creates a Reflection of the given pointer
func New(v any) (*Reflection, error) {
rv := reflect.ValueOf(v)
if rv.Kind() != reflect.Pointer || rv.IsNil() {
err := &InvalidUnmarshalError{Type: rv.Type()}
return nil, err
}
r := &Reflection{
v: rv,
types: make(map[reflect.Type]TypeInfo),
}
if err := r.scan(); err != nil {
return nil, err
}
return r, nil
}
-37
View File
@@ -1,37 +0,0 @@
package reflection
import (
"reflect"
)
func (r *Reflection) scanType(rt reflect.Type) error {
switch rt.Kind() {
case reflect.Pointer, reflect.Slice:
rt = rt.Elem()
}
if _, ok := r.types[rt]; ok {
// cached
return nil
}
ti, err := NewTypeInfo(rt)
if err != nil {
return &UnmarshalTypeError{Type: rt, Err: err}
}
r.types[rt] = ti
for i := range ti.fields {
err := r.scanType(ti.fields[i].sf.Type)
if err != nil {
return err
}
}
return nil
}
func (r *Reflection) scan() error {
return r.scanType(r.v.Type())
}
-108
View File
@@ -1,108 +0,0 @@
package reflection
import (
"log"
"reflect"
"github.com/fatih/structtag"
"github.com/pkg/errors"
)
// TypeInfo ...
type TypeInfo struct {
rt reflect.Type
fields []TypeInfoField
}
// Type ...
func (ti *TypeInfo) Type() reflect.Type {
return ti.rt
}
// Kind ...
func (ti *TypeInfo) Kind() reflect.Kind {
return ti.rt.Kind()
}
// TypeInfoField ...
type TypeInfoField struct {
sf reflect.StructField
tags *structtag.Tags
Name string
}
// Type ...
func (tif *TypeInfoField) Type() reflect.Type {
return tif.sf.Type
}
// Kind ...
func (tif *TypeInfoField) Kind() reflect.Kind {
return tif.sf.Type.Kind()
}
// Tag ...
func (tif *TypeInfoField) Tag(key string) (*structtag.Tag, bool) {
if tif.tags != nil {
tag, _ := tif.tags.Get(key)
if tag != nil {
return tag, true
}
}
return nil, false
}
// NewTypeInfo ...
func NewTypeInfo(rt reflect.Type) (TypeInfo, error) {
log.Printf("%s.%s: %s (%s)", "reflection", "NewTypeInfo", rt, rt.Kind())
ti := TypeInfo{
rt: rt,
}
err := ti.init()
return ti, err
}
func (ti *TypeInfo) init() error {
switch ti.Kind() {
case reflect.Struct:
return ti.initStruct()
default:
return nil
}
}
func (ti *TypeInfo) initStruct() error {
// load fields
n := ti.rt.NumField()
fields := make([]TypeInfoField, 0, n)
for i := 0; i < n; i++ {
sf := ti.rt.Field(i)
tags, err := structtag.Parse(string(sf.Tag))
if err != nil {
// tag parse error
err = &UnmarshalTypeError{
Type: ti.rt,
Err: errors.Wrap(err, structFieldName(sf)),
}
return err
}
if sf.IsExported() {
log.Printf("%s.%s: [%v]: %q %s (%s)",
"reflection", "NewTypeInfo",
i, structFieldName(sf), sf.Type, sf.Type.Kind())
fields = append(fields, TypeInfoField{
sf: sf,
tags: tags,
})
}
}
ti.fields = fields
return nil
}