Improve debugging output

This commit is contained in:
Adam Scarr 2017-08-13 19:50:41 +10:00
parent a4677a5834
commit a733d0ae13
5 changed files with 133 additions and 75 deletions

View File

@ -44,16 +44,13 @@ func (dp *debugParser) logf(ps *State, result *Result, format string, args ...in
buf.WriteString(fmt.Sprintf("%"+strconv.Itoa(longestLocation)+"s | ", dp.Location))
buf.WriteString(fmt.Sprintf("%-15s", ps.Preview(15)))
buf.WriteString(" | ")
output := ""
if ps.Errored() {
output = "fail"
} else if result != nil {
output = result.Token
}
buf.WriteString(fmt.Sprintf("%-10s | ", output))
buf.WriteString(strings.Repeat(" ", len(activeParsers)-1))
buf.WriteString(fmt.Sprintf(format, args...))
buf.WriteString(fmt.Sprintf(" > %#v", result))
if ps.Errored() {
buf.WriteString(fmt.Sprintf(" did not find %s", ps.Error.expected))
} else if result != nil {
buf.WriteString(fmt.Sprintf(" found %s", result.String()))
}
buf.WriteRune('\n')
return buf.String()
}
@ -64,7 +61,7 @@ func (dp *debugParser) logStart(ps *State) {
fmt.Fprint(log, pendingOpenLog)
pendingOpenLog = ""
}
pendingOpenLog = dp.logf(ps, nil, dp.Name())
pendingOpenLog = dp.logf(ps, nil, dp.Name()+" {")
}
}
@ -73,6 +70,8 @@ func (dp *debugParser) logEnd(ps *State, result *Result) {
if pendingOpenLog != "" {
fmt.Fprintf(log, dp.logf(ps, result, dp.Name()))
pendingOpenLog = ""
} else {
fmt.Fprintf(log, dp.logf(ps, result, "}"))
}
}
}

View File

@ -7,17 +7,6 @@ import (
"unicode/utf8"
)
var TrashResult = &Result{}
// Result is the output of a parser. Usually only one of its fields will be set and should be though of
// more as a union type. having it avoids interface{} littered all through the parsing code and makes
// the it easy to do the two most common operations, getting a token and finding a child.
type Result struct {
Token string
Child []Result
Result interface{}
}
// Parser is the workhorse of parsify. A parser takes a State and returns a result, consuming some
// of the State in the process.
// Given state is shared there are a few rules that should be followed:
@ -64,10 +53,6 @@ func Parsify(p Parserish) Parser {
}
case string:
return Exact(p)
case VoidParser:
return func(ptr *State, node *Result) {
p(ptr)
}
case func(*State):
return func(ptr *State, node *Result) {
p(ptr)

125
readme.md
View File

@ -28,58 +28,81 @@ When a parser isnt working as you intended you can build with debugging and enab
This works great with tests, eg in the goparsify source tree
```
$ cd html
$ go test -tags debug -parselogs
html.go:50 | <body>hello <p | | tag
html.go:45 | <body>hello <p | | tstart
html.go:45 | body>hello <p c | < | <
html.go:20 | >hello <p color | body | identifier
html.go:35 | >hello <p color | | attrs
html.go:34 | >hello <p color | | attr
html.go:20 | >hello <p color | fail | identifier
html.go:45 | hello <p color= | > | >
html.go:26 | hello <p color= | | elements
html.go:25 | hello <p color= | | element
html.go:21 | <p color="blue" | hello | text
html.go:25 | <p color="blue" | | element
html.go:21 | <p color="blue" | fail | text
html.go:50 | <p color="blue" | | tag
html.go:45 | <p color="blue" | | tstart
html.go:45 | p color="blue"> | < | <
html.go:20 | color="blue">w | p | identifier
html.go:35 | color="blue">w | | attrs
html.go:34 | color="blue">w | | attr
html.go:20 | ="blue">world</ | color | identifier
html.go:34 | "blue">world</p | = | =
html.go:34 | >world</p></bod | | string literal
html.go:34 | >world</p></bod | | attr
html.go:20 | >world</p></bod | fail | identifier
html.go:45 | world</p></body | > | >
html.go:26 | world</p></body | | elements
html.go:25 | world</p></body | | element
html.go:21 | </p></body> | world | text
html.go:25 | </p></body> | | element
html.go:21 | </p></body> | fail | text
html.go:50 | </p></body> | | tag
html.go:45 | </p></body> | | tstart
html.go:45 | /p></body> | < | <
html.go:20 | /p></body> | fail | identifier
html.go:46 | </p></body> | | tend
html.go:46 | p></body> | </ | </
html.go:20 | ></body> | p | identifier
html.go:46 | </body> | > | >
html.go:25 | </body> | | element
html.go:21 | </body> | fail | text
html.go:50 | </body> | | tag
html.go:45 | </body> | | tstart
html.go:45 | /body> | < | <
html.go:20 | /body> | fail | identifier
html.go:46 | </body> | | tend
html.go:46 | body> | </ | </
html.go:20 | > | body | identifier
html.go:46 | | > | >
adam:goparsify(master)$ go test -tags debug ./html -v
=== RUN TestParse
html.go:48 | <body>hello <p | tag {
html.go:43 | <body>hello <p | tstart {
html.go:43 | body>hello <p c | < found <
html.go:20 | >hello <p color | identifier found body
html.go:33 | >hello <p color | attrs {
html.go:32 | >hello <p color | attr {
html.go:20 | >hello <p color | identifier did not find [a-zA-Z][a-zA-Z0-9]*
html.go:32 | >hello <p color | } did not find [a-zA-Z][a-zA-Z0-9]*
html.go:33 | >hello <p color | } found
html.go:43 | hello <p color= | > found >
html.go:43 | hello <p color= | } found [<,body,,map[string]string{},>]
html.go:24 | hello <p color= | elements {
html.go:23 | hello <p color= | element {
html.go:21 | <p color="blue" | text found hello
html.go:23 | <p color="blue" | } found "hello "
html.go:23 | <p color="blue" | element {
html.go:21 | <p color="blue" | text did not find <>
html.go:48 | <p color="blue" | tag {
html.go:43 | <p color="blue" | tstart {
html.go:43 | p color="blue"> | < found <
html.go:20 | color="blue">w | identifier found p
html.go:33 | color="blue">w | attrs {
html.go:32 | color="blue">w | attr {
html.go:20 | ="blue">world</ | identifier found color
html.go:32 | "blue">world</p | = found =
html.go:32 | >world</p></bod | string literal found "blue"
html.go:32 | >world</p></bod | } found [color,=,"blue"]
html.go:32 | >world</p></bod | attr {
html.go:20 | >world</p></bod | identifier did not find [a-zA-Z][a-zA-Z0-9]*
html.go:32 | >world</p></bod | } did not find [a-zA-Z][a-zA-Z0-9]*
html.go:33 | >world</p></bod | } found [[color,=,"blue"]]
html.go:43 | world</p></body | > found >
html.go:43 | world</p></body | } found [<,p,,map[string]string{"color":"blue"},>]
html.go:24 | world</p></body | elements {
html.go:23 | world</p></body | element {
html.go:21 | </p></body> | text found world
html.go:23 | </p></body> | } found "world"
html.go:23 | </p></body> | element {
html.go:21 | </p></body> | text did not find <>
html.go:48 | </p></body> | tag {
html.go:43 | </p></body> | tstart {
html.go:43 | /p></body> | < found <
html.go:20 | /p></body> | identifier did not find [a-zA-Z][a-zA-Z0-9]*
html.go:43 | </p></body> | } did not find [a-zA-Z][a-zA-Z0-9]*
html.go:48 | </p></body> | } did not find [a-zA-Z][a-zA-Z0-9]*
html.go:23 | </p></body> | } did not find [a-zA-Z][a-zA-Z0-9]*
html.go:24 | </p></body> | } found ["world"]
html.go:44 | </p></body> | tend {
html.go:44 | p></body> | </ found </
html.go:20 | ></body> | identifier found p
html.go:44 | </body> | > found >
html.go:44 | </body> | } found [</,,p,>]
html.go:48 | </body> | } found "hello "
html.go:23 | </body> | } found html.htmlTag{Name:"p", Attributes:map[string]string{"color":"blue"}, Body:[]interface {}{"world"}}
html.go:23 | </body> | element {
html.go:48 | </body> | tag {
html.go:43 | </body> | tstart {
html.go:43 | /body> | < found <
html.go:20 | /body> | identifier did not find [a-zA-Z][a-zA-Z0-9]*
html.go:43 | </body> | } did not find [a-zA-Z][a-zA-Z0-9]*
html.go:48 | </body> | } did not find [a-zA-Z][a-zA-Z0-9]*
html.go:21 | </body> | text did not find <>
html.go:23 | </body> | } did not find [a-zA-Z][a-zA-Z0-9]*
html.go:24 | </body> | } found ["hello ",html.htmlTag{Name:"p", Attributes:map[string]string{"color":"blue"}, Body:[]interface {}{"world"}}]
html.go:44 | </body> | tend {
html.go:44 | body> | </ found </
html.go:20 | > | identifier found body
html.go:44 | | > found >
html.go:44 | | } found [</,,body,>]
html.go:48 | | } found [[<,body,,map[string]string{},>],,[]interface {}{"hello ", html.htmlTag{Name:"p", Attributes:map[string]string{"color":"blue"}, Body:[]interface {}{"world"}}},[</,,body,>]]
--- PASS: TestParse (0.00s)
PASS
ok github.com/vektah/goparsify/html 0.118s
ok github.com/vektah/goparsify/html 0.117s
```
### debugging performance

36
result.go Normal file
View File

@ -0,0 +1,36 @@
package goparsify
import (
"fmt"
"strings"
)
var TrashResult = &Result{}
// Result is the output of a parser. Usually only one of its fields will be set and should be though of
// more as a union type. having it avoids interface{} littered all through the parsing code and makes
// the it easy to do the two most common operations, getting a token and finding a child.
type Result struct {
Token string
Child []Result
Result interface{}
}
func (r Result) String() string {
if r.Result != nil {
if rs, ok := r.Result.(fmt.Stringer); ok {
return rs.String()
}
return fmt.Sprintf("%#v", r.Result)
}
if len(r.Child) > 0 {
children := []string{}
for _, child := range r.Child {
children = append(children, child.String())
}
return "[" + strings.Join(children, ",") + "]"
}
return r.Token
}

15
result_test.go Normal file
View File

@ -0,0 +1,15 @@
package goparsify
import (
"math/big"
"testing"
"github.com/stretchr/testify/require"
)
func TestResult_String(t *testing.T) {
require.Equal(t, "Hello", Result{Token: "Hello"}.String())
require.Equal(t, "[Hello,World]", Result{Child: []Result{{Token: "Hello"}, {Token: "World"}}}.String())
require.Equal(t, "10", Result{Result: 10}.String())
require.Equal(t, "10", Result{Result: big.NewInt(10)}.String())
}