From 6014908ab660840126c800edd47ae871e387f81a Mon Sep 17 00:00:00 2001
From: Ian Littman <iansltx@gmail.com>
Date: Thu, 2 Jan 2025 14:43:36 -0600
Subject: [PATCH] Correct fix for long MSI interned strings (#25104)

h/t https://github.com/binref/refinery/issues/72, for #24720. No changes
file as this is an unreleased bug.

Also added output for version in the custom package parser tool.

# Checklist for submitter

If some of the following don't apply, delete the relevant line.

<!-- Note that API documentation changes are now addressed by the
product design team. -->

- [x] Changes file added for user-visible changes in `changes/`,
`orbit/changes/` or `ee/fleetd-chrome/changes`.
See [Changes
files](https://github.com/fleetdm/fleet/blob/main/docs/Contributing/Committing-Changes.md#changes-files)
for more information.
- [x] Manual QA for all new/changed functionality
---
 pkg/file/msi.go                     | 18 +++++++-----------
 tools/custom-package-parser/main.go |  4 ++--
 2 files changed, 9 insertions(+), 13 deletions(-)

diff --git a/pkg/file/msi.go b/pkg/file/msi.go
index db549e9b4ea1..7271dd0d9117 100644
--- a/pkg/file/msi.go
+++ b/pkg/file/msi.go
@@ -242,24 +242,20 @@ func decodeStrings(dataReader, poolReader io.Reader) ([]string, error) {
 			}
 			return nil, fmt.Errorf("failed to read pool entry: %w", err)
 		}
-		stringEntrySize := int(stringEntry.Size)
+		stringEntrySize := uint32(stringEntry.Size)
 
-		// For string pool entries too long for the size to fit in a single uint16, entry size is 8 bytes instead of 4,
-		// with the first two bytes as zeroes, the next two are the two most-significant bytes of the size, shifted
-		// 17 (?!?) bits to the left, the following two are the less-significant bits of the size, and the last two are
-		// the reference count. Verified with the OpenVPN Connect v3 installer, which has a large string blob for
-		// licenses where a 17-bit shift captures the length properly.
+		// For string pool entries too long for the size to fit in a single uint16, the format sets the size as zero,
+		// maintains the reference count location in the structure, then uses the following four bytes (little-endian)
+		// to store the string size. See https://github.com/binref/refinery/issues/72.
 		if stringEntry.Size == 0 && stringEntry.RefCount != 0 {
-			stringEntrySize = int(stringEntry.RefCount) << 17
-			err := binary.Read(poolReader, binary.LittleEndian, &stringEntry)
+			err := binary.Read(poolReader, binary.LittleEndian, &stringEntrySize)
 			if err != nil {
-				return nil, fmt.Errorf("failed to read large string pool entry: %w", err)
+				return nil, fmt.Errorf("failed to read size of large string in string pool: %w", err)
 			}
-			stringEntrySize += int(stringEntry.Size)
 		}
 
 		buf.Reset()
-		buf.Grow(stringEntrySize)
+		buf.Grow(int(stringEntrySize))
 		_, err = io.CopyN(&buf, dataReader, int64(stringEntrySize))
 		if err != nil {
 			return nil, fmt.Errorf("failed to read string data: %w", err)
diff --git a/tools/custom-package-parser/main.go b/tools/custom-package-parser/main.go
index d64506625e2c..9820da7cb7d9 100644
--- a/tools/custom-package-parser/main.go
+++ b/tools/custom-package-parser/main.go
@@ -31,8 +31,8 @@ func main() {
 	}
 
 	fmt.Printf(
-		"- Name: '%s'\n- Bundle Identifier: '%s'\n- Package IDs: '%s'\n",
-		metadata.Name, metadata.BundleIdentifier, strings.Join(metadata.PackageIDs, ","),
+		"- Name: '%s'\n- Bundle Identifier: '%s'\n- Package IDs: '%s'\n- Version: %s\n\n",
+		metadata.Name, metadata.BundleIdentifier, strings.Join(metadata.PackageIDs, ","), metadata.Version,
 	)
 }