From 43d1b3b49eb7fe6957fb75e7d1e9e576e6b401ad Mon Sep 17 00:00:00 2001
From: brharrington <brharrington@gmail.com>
Date: Thu, 22 Jun 2023 10:34:27 -0500
Subject: [PATCH] support allow set for PathSanitizer (#1065)

This allows an explicit set of allowed segment strings
for some use-cases where there is a small set of known
strings that are desirable to include, but do not match
with the normal rules.
---
 .../spectator/ipc/http/PathSanitizer.java     | 30 +++++++++++++++----
 .../spectator/ipc/http/PathSanitizerTest.java | 14 ++++++++-
 2 files changed, 37 insertions(+), 7 deletions(-)

diff --git a/spectator-ext-ipc/src/main/java/com/netflix/spectator/ipc/http/PathSanitizer.java b/spectator-ext-ipc/src/main/java/com/netflix/spectator/ipc/http/PathSanitizer.java
index 723de7adb..50cdac5c1 100644
--- a/spectator-ext-ipc/src/main/java/com/netflix/spectator/ipc/http/PathSanitizer.java
+++ b/spectator-ext-ipc/src/main/java/com/netflix/spectator/ipc/http/PathSanitizer.java
@@ -1,5 +1,5 @@
 /*
- * Copyright 2014-2021 Netflix, Inc.
+ * Copyright 2014-2023 Netflix, Inc.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -17,6 +17,9 @@
 
 import com.netflix.spectator.impl.AsciiSet;
 
+import java.util.Collections;
+import java.util.Set;
+
 /**
  * Helper for sanitizing a URL path for including as the {@code ipc.endpoint} value. Makes
  * a best effort to try and remove segments that tend to be variable like numbers, UUIDs,
@@ -37,7 +40,22 @@ private PathSanitizer() {
 
   /** Returns a sanitized path string for use as an endpoint tag value. */
   public static String sanitize(String path) {
-    return sanitizeSegments(removeParameters(path));
+    return sanitizeSegments(removeParameters(path), Collections.emptySet());
+  }
+
+  /**
+   * Returns a sanitized path string for use as an endpoint tag value.
+   *
+   * @param path
+   *     Path from a URI that should be sanitized.
+   * @param allowed
+   *     Set of allowed segment strings. This can be used to override the default rules for
+   *     a set of known good values.
+   * @return
+   *     Sanitized path that can be used as an endpoint tag value.
+   */
+  public static String sanitize(String path, Set<String> allowed) {
+    return sanitizeSegments(removeParameters(path), allowed);
   }
 
   private static String removeParameters(String path) {
@@ -49,7 +67,7 @@ private static String removeParameters(String path, char c) {
     return i > 0 ? path.substring(0, i) : path;
   }
 
-  private static String sanitizeSegments(String path) {
+  private static String sanitizeSegments(String path, Set<String> allowed) {
     StringBuilder builder = new StringBuilder();
     int length = path.length();
     int pos = path.charAt(0) == '/' ? 1 : 0;
@@ -67,7 +85,7 @@ private static String sanitizeSegments(String path) {
       }
 
       if (!segment.isEmpty()) {
-        if (shouldSuppressSegment(segment))
+        if (shouldSuppressSegment(segment, allowed))
           appendIfSpaceAvailable(builder, "-");
         else
           appendIfSpaceAvailable(builder, segment);
@@ -78,10 +96,10 @@ private static String sanitizeSegments(String path) {
     return builder.toString();
   }
 
-  private static boolean shouldSuppressSegment(String segment) {
+  private static boolean shouldSuppressSegment(String segment, Set<String> allowed) {
     // GraphQL is a common endpoint, but hits the consonants check to detect strings
     // that are likely to be random. Special case it for now.
-    if ("graphql".equals(segment)) {
+    if ("graphql".equals(segment) || allowed.contains(segment)) {
       return false;
     }
 
diff --git a/spectator-ext-ipc/src/test/java/com/netflix/spectator/ipc/http/PathSanitizerTest.java b/spectator-ext-ipc/src/test/java/com/netflix/spectator/ipc/http/PathSanitizerTest.java
index bc3bbcf16..faf704262 100644
--- a/spectator-ext-ipc/src/test/java/com/netflix/spectator/ipc/http/PathSanitizerTest.java
+++ b/spectator-ext-ipc/src/test/java/com/netflix/spectator/ipc/http/PathSanitizerTest.java
@@ -1,5 +1,5 @@
 /*
- * Copyright 2014-2021 Netflix, Inc.
+ * Copyright 2014-2023 Netflix, Inc.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -22,9 +22,11 @@
 import java.nio.file.Files;
 import java.nio.file.Path;
 import java.nio.file.Paths;
+import java.util.Collections;
 import java.util.List;
 import java.util.Locale;
 import java.util.Random;
+import java.util.Set;
 import java.util.UUID;
 
 public class PathSanitizerTest {
@@ -107,6 +109,16 @@ public void graphql() {
     Assertions.assertEquals("_graphql", sanitize(path));
   }
 
+  @Test
+  public void allowSet() {
+    String path = "/foo-bar";
+    Assertions.assertEquals("_-", PathSanitizer.sanitize(path));
+
+    Set<String> allowed = Collections.singleton("foo-bar");
+    Assertions.assertEquals("_foo-bar", PathSanitizer.sanitize(path, allowed));
+    Assertions.assertEquals("_api_v1_foo-bar", PathSanitizer.sanitize("/api/v1" + path, allowed));
+  }
+
   @Test
   public void randomLookingString() {
     String path = "/random/AOTV16LMT2";