Go: fix flow through string concatenation in go/incomplete-hostname-regexp

go/ql/src/Security/CWE-020/IncompleteHostnameRegexp.ql

@@ -81,14 +81,12 @@ predicate regexpGuardsError(RegexpPattern regexp) {
 
 module IncompleteHostNameRegexpConfig implements DataFlow::ConfigSig {
   additional predicate isSourceString(DataFlow::Node source, string hostPart) {
-    exists(Expr e |
-      e = source.asExpr() and
-      isIncompleteHostNameRegexpPattern(e.getStringValue(), hostPart)
-    |
-      e instanceof StringLit
-      or
-      e instanceof AddExpr and
-      not isIncompleteHostNameRegexpPattern(e.(AddExpr).getAnOperand().getStringValue(), _)
+    exists(Expr e | e = source.asExpr() |
+      isIncompleteHostNameRegexpPattern(e.getStringValue(), hostPart) and
+      // Exclude constant names to avoid duplicate results, because the string
+      // literals which they are initialised with are also considered as
+      // sources.
+      not e instanceof ConstantName
     )
   }
 
@@ -101,6 +99,10 @@ module IncompleteHostNameRegexpConfig implements DataFlow::ConfigSig {
     ) and
     not regexpGuardsError(sink)
   }
+
+  predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
+    StringOps::Concatenation::taintStep(node1, node2)
+  }
 }
 
 module Flow = DataFlow::Global<IncompleteHostNameRegexpConfig>;

go/ql/src/Security/CWE-020/IncompleteHostnameRegexpGood2.go

@@ -6,7 +6,7 @@ import (
 	"regexp"
 )
 
-func checkRedirectGood(req *http.Request, via []*http.Request) error {
+func checkRedirectGood2(req *http.Request, via []*http.Request) error {
 	// GOOD: the host of `req.URL` must be `example.com`, `www.example.com` or `beta.example.com`
 	re := `^((www|beta)\.)?example\.com/`
 	if matched, _ := regexp.MatchString(re, req.URL.Host); matched {

go/ql/src/change-notes/2024-04-24-incomplete-hostname-regexp.md

@@ -0,0 +1,4 @@
+---
+category: minorAnalysis
+---
+* The query `go/incomplete-hostname-regexp` now recognizes more sources involving concatenation of string literals and also follows flow through string concatenation. This may lead to more alerts.

go/ql/test/query-tests/Security/CWE-020/IncompleteHostnameRegexp/IncompleteHostnameRegexp.expected

@@ -1,12 +1,22 @@
 edges
 | IncompleteHostnameRegexp.go:11:8:11:36 | "^((www\|beta).)?example.com/" | IncompleteHostnameRegexp.go:12:38:12:39 | re | provenance |  |
+| main.go:49:21:49:45 | `https://www.example.com` | main.go:62:15:62:25 | sourceConst | provenance |  |
+| main.go:62:15:62:25 | sourceConst | main.go:65:15:65:23 | localVar3 | provenance |  |
 nodes
 | IncompleteHostnameRegexp.go:11:8:11:36 | "^((www\|beta).)?example.com/" | semmle.label | "^((www\|beta).)?example.com/" |
 | IncompleteHostnameRegexp.go:12:38:12:39 | re | semmle.label | re |
-| main.go:39:60:39:79 | "^test2.github.com$" | semmle.label | "^test2.github.com$" |
-| main.go:44:15:44:39 | `https://www.example.com` | semmle.label | `https://www.example.com` |
+| main.go:40:60:40:79 | "^test2.github.com$" | semmle.label | "^test2.github.com$" |
+| main.go:45:15:45:39 | `https://www.example.com` | semmle.label | `https://www.example.com` |
+| main.go:49:21:49:45 | `https://www.example.com` | semmle.label | `https://www.example.com` |
+| main.go:56:15:56:34 | ...+... | semmle.label | ...+... |
+| main.go:58:15:58:42 | ...+... | semmle.label | ...+... |
+| main.go:62:15:62:25 | sourceConst | semmle.label | sourceConst |
+| main.go:65:15:65:23 | localVar3 | semmle.label | localVar3 |
 subpaths
 #select
 | IncompleteHostnameRegexp.go:11:8:11:36 | "^((www\|beta).)?example.com/" | IncompleteHostnameRegexp.go:11:8:11:36 | "^((www\|beta).)?example.com/" | IncompleteHostnameRegexp.go:12:38:12:39 | re | This regular expression has an unescaped dot before ')?example.com', so it might match more hosts than expected when $@. | IncompleteHostnameRegexp.go:12:38:12:39 | re | the regular expression is used |
-| main.go:39:60:39:79 | "^test2.github.com$" | main.go:39:60:39:79 | "^test2.github.com$" | main.go:39:60:39:79 | "^test2.github.com$" | This regular expression has an unescaped dot before 'github.com', so it might match more hosts than expected when $@. | main.go:39:60:39:79 | "^test2.github.com$" | the regular expression is used |
-| main.go:44:15:44:39 | `https://www.example.com` | main.go:44:15:44:39 | `https://www.example.com` | main.go:44:15:44:39 | `https://www.example.com` | This regular expression has an unescaped dot before 'example.com', so it might match more hosts than expected when $@. | main.go:44:15:44:39 | `https://www.example.com` | the regular expression is used |
+| main.go:40:60:40:79 | "^test2.github.com$" | main.go:40:60:40:79 | "^test2.github.com$" | main.go:40:60:40:79 | "^test2.github.com$" | This regular expression has an unescaped dot before 'github.com', so it might match more hosts than expected when $@. | main.go:40:60:40:79 | "^test2.github.com$" | the regular expression is used |
+| main.go:45:15:45:39 | `https://www.example.com` | main.go:45:15:45:39 | `https://www.example.com` | main.go:45:15:45:39 | `https://www.example.com` | This regular expression has an unescaped dot before 'example.com', so it might match more hosts than expected when $@. | main.go:45:15:45:39 | `https://www.example.com` | the regular expression is used |
+| main.go:49:21:49:45 | `https://www.example.com` | main.go:49:21:49:45 | `https://www.example.com` | main.go:65:15:65:23 | localVar3 | This regular expression has an unescaped dot before 'example.com', so it might match more hosts than expected when $@. | main.go:65:15:65:23 | localVar3 | the regular expression is used |
+| main.go:56:15:56:34 | ...+... | main.go:56:15:56:34 | ...+... | main.go:56:15:56:34 | ...+... | This regular expression has an unescaped dot before 'example.com', so it might match more hosts than expected when $@. | main.go:56:15:56:34 | ...+... | the regular expression is used |
+| main.go:58:15:58:42 | ...+... | main.go:58:15:58:42 | ...+... | main.go:58:15:58:42 | ...+... | This regular expression has an unescaped dot before 'example.com', so it might match more hosts than expected when $@. | main.go:58:15:58:42 | ...+... | the regular expression is used |

go/ql/test/query-tests/Security/CWE-020/IncompleteHostnameRegexp/IncompleteHostnameRegexpGood2.go

@@ -0,0 +1,16 @@
+package main
+
+import (
+	"errors"
+	"net/http"
+	"regexp"
+)
+
+func checkRedirectGood2(req *http.Request, via []*http.Request) error {
+	// GOOD: the host of `req.URL` must be `example.com`, `www.example.com` or `beta.example.com`
+	re := `^((www|beta)\.)?example\.com/`
+	if matched, _ := regexp.MatchString(re, req.URL.Host); matched {
+		return nil
+	}
+	return errors.New("Invalid redirect")
+}

go/ql/test/query-tests/Security/CWE-020/IncompleteHostnameRegexp/main.go

@@ -3,10 +3,11 @@
 package main
 
 import (
-	"github.com/elazarl/goproxy"
 	"net/http"
 	"regexp"
 	"time"
+
+	"github.com/elazarl/goproxy"
 )
 
 func Match(notARegex string) bool {
@@ -44,3 +45,22 @@ func main() {
 	regexp.Match(`https://www.example.com`, []byte(""))   // NOT OK
 	regexp.Match(`https://www\.example\.com`, []byte("")) // OK
 }
+
+const sourceConst = `https://www.example.com`
+const firstHalfConst = `https://www.example.`
+
+func concatenateStrings() {
+	firstHalf := `https://www.example.`
+	regexp.Match(firstHalf+`com`, []byte("")) // MISSING: NOT OK
+
+	regexp.Match(firstHalfConst+`com`, []byte("")) // NOT OK
+
+	regexp.Match(`https://www.example.`+`com`, []byte("")) // NOT OK
+}
+
+func avoidDuplicateResults() {
+	localVar1 := sourceConst
+	localVar2 := localVar1
+	localVar3 := localVar2
+	regexp.Match(localVar3, []byte("")) // NOT OK
+}