fix: parse failure when template files are > 8KB in size

a-h · Jan 30, 2022 · 17fe039 · 17fe039
1 parent 0a04450
commit 17fe039
Show file tree

Hide file tree

Showing 6 changed files with 40 additions and 9 deletions.
diff --git a/go.mod b/go.mod
@@ -4,7 +4,7 @@ go 1.16
 
 require (
 	github.com/PuerkitoBio/goquery v1.7.1
-	github.com/a-h/lexical v0.0.47
+	github.com/a-h/lexical v0.0.50
 	github.com/a-h/pathvars v0.0.0-20200320143331-78b263b728e2
 	github.com/google/go-cmp v0.5.6
 	github.com/hashicorp/errwrap v1.1.0 // indirect
@@ -16,7 +16,7 @@ require (
 	go.uber.org/atomic v1.8.0 // indirect
 	go.uber.org/multierr v1.7.0 // indirect
 	go.uber.org/zap v1.17.0
-	golang.org/x/mod v0.5.1 // indirect
+	golang.org/x/mod v0.5.1
 )
 
 //replace github.com/a-h/lexical => /Users/adrian/github.com/a-h/lexical
diff --git a/go.sum b/go.sum
@@ -1,7 +1,11 @@
 github.com/PuerkitoBio/goquery v1.7.1 h1:oE+T06D+1T7LNrn91B4aERsRIeCLJ/oPSa6xB9FPnz4=
 github.com/PuerkitoBio/goquery v1.7.1/go.mod h1:XY0pP4kfraEmmV1O7Uf6XyjoslwsneBbgeDjLYuN8xY=
-github.com/a-h/lexical v0.0.47 h1:8VshFdJrb1wnDiA1bFqoxPhXsxpMqPFAsLcwLNQ8rq4=
-github.com/a-h/lexical v0.0.47/go.mod h1:d73jw5cgKXuYypRozNBuxRNFrTWQ3y5hVMG7rUjh1Qw=
+github.com/a-h/lexical v0.0.48 h1:vliyXGKMmEUP2hg8m+zq2vyE04iYNWvq1V3hH2QSdcE=
+github.com/a-h/lexical v0.0.48/go.mod h1:d73jw5cgKXuYypRozNBuxRNFrTWQ3y5hVMG7rUjh1Qw=
+github.com/a-h/lexical v0.0.49 h1:TRnSxSI7M7zAz/R5DFjL9kbX3tQgQ6BsvXe3VwpMtw0=
+github.com/a-h/lexical v0.0.49/go.mod h1:d73jw5cgKXuYypRozNBuxRNFrTWQ3y5hVMG7rUjh1Qw=
+github.com/a-h/lexical v0.0.50 h1:QGyK1vn51J3RyWgoMGdoObcCHcyqtEVBaGmQHCzhLog=
+github.com/a-h/lexical v0.0.50/go.mod h1:d73jw5cgKXuYypRozNBuxRNFrTWQ3y5hVMG7rUjh1Qw=
 github.com/a-h/pathvars v0.0.0-20200320143331-78b263b728e2 h1:OObPWjEG3FZsjTNQPATyh8IpAvUfB6WpIO4mSzRSbRc=
 github.com/a-h/pathvars v0.0.0-20200320143331-78b263b728e2/go.mod h1:QulNSBbUWcOdt7MTypwParV0/gDtr8qRBmQAiDdOnS4=
 github.com/andybalholm/cascadia v1.2.0 h1:vuRCkM5Ozh/BfmsaTm26kbjm0mIOM3yS5Ek/F5h18aE=

diff --git a/parser/doctypeparser_test.go b/parser/doctypeparser_test.go
@@ -76,9 +76,9 @@ func TestDocTypeParserErrors(t *testing.T) {
 					Col:   0,
 				},
 				Position{
-					Index: 17,
+					Index: 15,
 					Line:  1,
-					Col:   17,
+					Col:   14,
 				}),
 		},
 		{

diff --git a/parser/elementparser_test.go b/parser/elementparser_test.go
@@ -1,6 +1,7 @@
 package parser
 
 import (
+	"strings"
 	"testing"
 
 	"github.com/a-h/lexical/input"
@@ -575,3 +576,19 @@ func TestElementParserErrors(t *testing.T) {
 		})
 	}
 }
+
+func TestBigElement(t *testing.T) {
+	var sb strings.Builder
+	sb.WriteString("<div>")
+	for i := 0; i < 4096*4; i++ {
+		sb.WriteString("a")
+	}
+	sb.WriteString("</div>")
+	result := newElementParser().Parse(input.NewFromString(sb.String()))
+	if result.Error != nil {
+		t.Fatalf("unexpected error: %v", result.Error)
+	}
+	if !result.Success {
+		t.Errorf("unexpected failure to parse")
+	}
+}
diff --git a/parser/packageparser_test.go b/parser/packageparser_test.go
@@ -25,9 +25,9 @@ func TestPackageParserErrors(t *testing.T) {
 					Col:   11,
 				},
 				Position{
-					Index: 15,
+					Index: 12,
 					Line:  1,
-					Col:   15,
+					Col:   11,
 				},
 			),
 		},

diff --git a/parser/parser.go b/parser/parser.go
@@ -186,13 +186,23 @@ func (p TemplateFileParser) Parse(pi parse.Input) parse.Result {
 	return parse.Success("template file", tf, nil)
 }
 
+const maxBufferSize = 1024 * 1024 * 10 // 10MB
+
 func Parse(fileName string) (TemplateFile, error) {
 	f, err := os.Open(fileName)
 	if err != nil {
 		return TemplateFile{}, err
 	}
+	fi, err := f.Stat()
+	if err != nil {
+		return TemplateFile{}, err
+	}
+	bufferSize := maxBufferSize
+	if fi.Size() < int64(bufferSize) {
+		bufferSize = int(fi.Size())
+	}
 	reader := bufio.NewReader(f)
-	tfr := NewTemplateFileParser().Parse(input.New(reader))
+	tfr := NewTemplateFileParser().Parse(input.NewWithBufferSize(reader, bufferSize))
 	if tfr.Error != nil {
 		return TemplateFile{}, tfr.Error
 	}