From 3a6b6562d10edbb5c21088c3f86a3aa4073750ff Mon Sep 17 00:00:00 2001
From: Laurent Goderre <laurent.goderre@docker.com>
Date: Fri, 22 Dec 2023 09:45:20 -0500
Subject: [PATCH] Add more functionality to the ErLang parser (#2390)

* ERLang parser support for empty lists
* ERLang add support for single quote strings
* ERLang parser support for comments
---------
Signed-off-by: Laurent Goderre <laurent.goderre@docker.com>
---
 syft/pkg/cataloger/erlang/erlang_parser.go    | 47 ++++++++++++++++++-
 .../cataloger/erlang/erlang_parser_test.go    | 42 +++++++++++++++++
 2 files changed, 88 insertions(+), 1 deletion(-)

diff --git a/syft/pkg/cataloger/erlang/erlang_parser.go b/syft/pkg/cataloger/erlang/erlang_parser.go
index 048768cb5..829b0c143 100644
--- a/syft/pkg/cataloger/erlang/erlang_parser.go
+++ b/syft/pkg/cataloger/erlang/erlang_parser.go
@@ -2,6 +2,7 @@ package erlang
 
 import (
 	"bytes"
+	"errors"
 	"fmt"
 	"io"
 	"strings"
@@ -12,6 +13,8 @@ type erlangNode struct {
 	value interface{}
 }
 
+var errSkipComments = errors.New("")
+
 func (e erlangNode) Slice() []erlangNode {
 	out, ok := e.value.([]erlangNode)
 	if ok {
@@ -56,6 +59,10 @@ func parseErlang(reader io.Reader) (erlangNode, error) {
 	i := 0
 	for i < len(data) {
 		item, err := parseErlangBlock(data, &i)
+		if err == errSkipComments {
+			skipWhitespace(data, &i)
+			continue
+		}
 		if err != nil {
 			return node(nil), fmt.Errorf("%w\n%s", err, printError(data, i))
 		}
@@ -141,11 +148,26 @@ func parseErlangNode(data []byte, i *int) (erlangNode, error) {
 	c := data[*i]
 	switch c {
 	case '[', '{':
+		offset := *i + 1
+		skipWhitespace(data, &offset)
+		c2 := data[offset]
+
+		// Add support for empty lists
+		if (c == '[' && c2 == ']') || (c == '{' && c2 == '}') {
+			*i = offset + 1
+			return node(nil), nil
+		}
+
 		return parseErlangList(data, i)
 	case '"':
+		fallthrough
+	case '\'':
 		return parseErlangString(data, i)
 	case '<':
 		return parseErlangAngleString(data, i)
+	case '%':
+		parseErlangComment(data, i)
+		return node(nil), errSkipComments
 	}
 
 	if isLiteral(c) {
@@ -205,7 +227,7 @@ func parseErlangString(data []byte, i *int) (erlangNode, error) {
 		buf.WriteByte(c)
 		*i++
 	}
-	return node(buf.String()), nil
+	return node(nil), fmt.Errorf("unterminated string at %d", *i)
 }
 
 func parseErlangList(data []byte, i *int) (erlangNode, error) {
@@ -216,6 +238,10 @@ func parseErlangList(data []byte, i *int) (erlangNode, error) {
 	for *i < len(data) {
 		item, err := parseErlangNode(data, i)
 		if err != nil {
+			if err == errSkipComments {
+				skipWhitespace(data, i)
+				continue
+			}
 			return node(nil), err
 		}
 		out.value = append(out.value.([]erlangNode), item)
@@ -225,6 +251,9 @@ func parseErlangList(data []byte, i *int) (erlangNode, error) {
 		case ',':
 			*i++
 			continue
+		case '%':
+			// Starts a new comment node
+			continue
 		case ']', '}':
 			*i++
 			return out, nil
@@ -234,3 +263,19 @@ func parseErlangList(data []byte, i *int) (erlangNode, error) {
 	}
 	return out, nil
 }
+
+func parseErlangComment(data []byte, i *int) {
+	for *i < len(data) {
+		c := data[*i]
+
+		*i++
+
+		// Rest of a line is a comment. Deals with CR, LF and CR/LF
+		if c == '\n' {
+			break
+		} else if c == '\r' && data[*i] == '\n' {
+			*i++
+			break
+		}
+	}
+}
diff --git a/syft/pkg/cataloger/erlang/erlang_parser_test.go b/syft/pkg/cataloger/erlang/erlang_parser_test.go
index 68225fb9b..adb031ddb 100644
--- a/syft/pkg/cataloger/erlang/erlang_parser_test.go
+++ b/syft/pkg/cataloger/erlang/erlang_parser_test.go
@@ -38,6 +38,21 @@ func Test_parseErlang(t *testing.T) {
  {<<"bcrypt">>, <<"3418821BC17CE6E96A4A77D1A88D7485BF783E212069FACFC79510AFBFF95352">>},
  {<<"unicode_util_compat">>, <<"25EEE6D67DF61960CF6A794239566599B09E17E668D3700247BC498638152521">>}]}
 ].`,
+		},
+		{
+			name: "empty list",
+			content: `
+{test, [
+ {with_space, [ ]},
+ {without_space, []}
+]}`,
+		},
+		{
+			name: "valid strings",
+			content: `
+{strings, [
+ "foo", 'bar'
+]}`,
 		},
 		{
 			name:    "invalid string content",
@@ -46,6 +61,14 @@ func Test_parseErlang(t *testing.T) {
 {"1.2.0
 ">>},
 ].`,
+		},
+		{
+			name:    "string mismach",
+			wantErr: require.Error,
+			content: `
+{bad_string, [
+ 'foo"
+ ]}`,
 		},
 		{
 			name:    "invalid content",
@@ -54,6 +77,25 @@ func Test_parseErlang(t *testing.T) {
 {"1.2.0"}.
 ].`,
 		},
+		{
+			name: "valid comments",
+			content: `
+{ comments, [
+	{ foo, bar },
+	%% this is a comment
+	% this is also a comment
+	{ hello, 'bar' }, %%inline comment
+	{ baz }
+]}`,
+		},
+		{
+			name: "starts with a comments",
+			content: `
+%% starts with comment
+{ comments, [
+	{ foo, bar }
+]}`,
+		},
 	}
 
 	for _, test := range tests {