From 88c3c0a021132f8eee18302b785f2fe8eef63025 Mon Sep 17 00:00:00 2001
From: Ben Boyter <ben@boyter.org>
Date: Mon, 12 Feb 2018 14:03:55 +1100
Subject: [PATCH 1/3] Format readme slightly

---
 README.md | 56 ++++---------------------------------------------------
 1 file changed, 4 insertions(+), 52 deletions(-)

diff --git a/README.md b/README.md
index 1ecbe2d..5dea050 100644
--- a/README.md
+++ b/README.md
@@ -145,55 +145,7 @@ zip -r9 lc-1.0.0-x86_64-pc-windows.zip lc.exe && zip -r9 lc-1.0.0-x86_64-unknown
 
 ### TODO
 
-Add error handling for all the file operations and just in general. Most are currently ignored
-
-Add logic to guess the file type for SPDX value FileType
-
-Add addtional unit and integration tests
-
-Investigate using "github.com/gosuri/uitable" for formatting https://github.com/gosuri/uitable
-
-Investigate using zlib compression for databases as per the below
-
-```
-package main
-
-import (
-	"bytes"
-	"compress/zlib"
-	"fmt"
-	"io"
-	"io/ioutil"
-)
-
-func readFile(filepath string) []byte {
-	// TODO only read as deep into the file as we need
-	bytes, err := ioutil.ReadFile(filepath)
-
-	if err != nil {
-		fmt.Print(err)
-	}
-
-	return bytes
-}
-
-func main() {
-
-	contents := readFile("database_keywords.json")
-	fmt.Println(len(contents))
-
-	var in bytes.Buffer
-	b := []byte(contents)
-	w := zlib.NewWriter(&in)
-	w.Write(b)
-	w.Close()
-
-	fmt.Println(len(in.String()))
-
-	var out bytes.Buffer
-	r, _ := zlib.NewReader(&in)
-	io.Copy(&out, r)
-	fmt.Println(len(out.String()))
-	// fmt.Println(len(out.String()))
-}
-```
+* Add error handling for all the file operations and just in general. Most are currently ignored
+* Add logic to guess the file type for SPDX value FileType
+* Add addtional unit and integration tests
+* Investigate using "github.com/gosuri/uitable" for formatting https://github.com/gosuri/uitable
\ No newline at end of file

From 55dc7ffc1ea228c0ca74f5b2aa170f1b27027775 Mon Sep 17 00:00:00 2001
From: Ben Boyter <ben@boyter.org>
Date: Mon, 12 Feb 2018 14:36:21 +1100
Subject: [PATCH 2/3] Update to have some integration tests

---
 parsers/guesser.go      |  4 ++--
 parsers/guesser_test.go | 36 ++++++++++++++++++++++++++++++++++++
 scripts/include.go      |  2 +-
 3 files changed, 39 insertions(+), 3 deletions(-)

diff --git a/parsers/guesser.go b/parsers/guesser.go
index b016f80..3871208 100644
--- a/parsers/guesser.go
+++ b/parsers/guesser.go
@@ -23,7 +23,7 @@ var ToolName = "licensechecker"
 var ToolVersion = "1.0.2"
 
 // Set by user as command line arguments
-var confidence = 0.0
+var confidence = 0.85
 var Confidence = ""
 var PossibleLicenceFiles = ""
 var DirFilePaths = []string{}
@@ -33,7 +33,7 @@ var DeepGuess = ""
 var Format = ""
 var FileOutput = ""
 var ExtentionBlacklist = ""
-var maxSize = 0
+var maxSize = 50000
 var MaxSize = ""
 var DocumentName = ""
 var PackageName = ""
diff --git a/parsers/guesser_test.go b/parsers/guesser_test.go
index 824b16f..5e9ca51 100644
--- a/parsers/guesser_test.go
+++ b/parsers/guesser_test.go
@@ -31,3 +31,39 @@ func TestLoadDatabase(t *testing.T) {
 		t.Errorf("Expected database to not be empty")
 	}
 }
+
+func TestWalkDirectory(t *testing.T) {
+	actual := walkDirectory("../examples/identifier/", [][]LicenseMatch{})
+
+	if len(actual) != 3 {
+		t.Errorf("Expected 3 results for directory")
+	}
+}
+
+func TestProcessFile(t *testing.T) {
+	actual := processFile("../examples/identifier/", "has_identifier.py", []LicenseMatch{})
+
+	if actual.Md5Hash != "0ad2e6786423fa6933a49ae4f97ae79e" {
+		t.Errorf("Expected MD5 to match")
+	}
+
+	if actual.Sha1Hash != "64904ca8a945009f95734d19198aaacd5e2db959" {
+		t.Errorf("Expected SHA1 to match")
+	}
+
+	if actual.Sha256Hash != "bf201f35c6a8504b4d956b4403ff5a7fef490889d5166a34f01b653e4ce08a06" {
+		t.Errorf("Expected SHA256 to match")
+	}
+
+	if len(actual.LicenseIdentified) != 2 {
+		t.Errorf("Expected 2 identified licenses")
+	}
+
+	if actual.LicenseIdentified[0].LicenseId != "GPL-2.0" {
+		t.Errorf("Expected license not identified")
+	}
+
+	if actual.LicenseIdentified[1].LicenseId != "GPL-3.0+" {
+		t.Errorf("Expected license not identified")
+	}
+}
diff --git a/scripts/include.go b/scripts/include.go
index 45468e7..3d642a0 100644
--- a/scripts/include.go
+++ b/scripts/include.go
@@ -22,7 +22,7 @@ func readFile(filepath string) []byte {
 }
 
 // Reads all .json files in the current folder
-// and encodes them as strings literals in textfiles.go
+// and encodes them as strings literals in constants.go
 func main() {
 	files, _ := ioutil.ReadDir(".")
 	out, _ := os.Create("./parsers/constants.go")

From 7980a11816f098e493e8e889f98f4eb73f93a3e9 Mon Sep 17 00:00:00 2001
From: Ben Boyter <ben@boyter.org>
Date: Mon, 12 Feb 2018 14:42:04 +1100
Subject: [PATCH 3/3] Rename version to 1.1.1 so that it reflects in the next
 build

---
 parsers/guesser.go | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/parsers/guesser.go b/parsers/guesser.go
index 3871208..1f55738 100644
--- a/parsers/guesser.go
+++ b/parsers/guesser.go
@@ -20,7 +20,7 @@ import (
 
 // Shared all over the place
 var ToolName = "licensechecker"
-var ToolVersion = "1.0.2"
+var ToolVersion = "1.1.1"
 
 // Set by user as command line arguments
 var confidence = 0.85