From a12ed0fa7bca1806ad7664d45f2b4e5ce92e1d94 Mon Sep 17 00:00:00 2001
From: Daniel Stenberg <daniel@haxx.se>
Date: Tue, 27 Aug 2024 10:35:52 +0200
Subject: [PATCH] trurl: allow "control bytes" in JSON output

For example when doing

    trurl 'http://example.org/%18' --json

Adapted and extended test cases to verify.

Fixes #262
Reported-by: Emanuele Torre
---
 tests.json | 64 +++++++++++++++++++++++++++++++++++++-----------------
 trurl.c    | 32 +++++++++++++++++++++++++--
 2 files changed, 74 insertions(+), 22 deletions(-)

diff --git a/tests.json b/tests.json
index 4e2c8e98..0ca0552b 100644
--- a/tests.json
+++ b/tests.json
@@ -2644,27 +2644,51 @@
           "returncode": 0
       }
   },
-  {
-    "input": {
-        "arguments": [
-            "http://example.org/%18",
-            "--json"
-        ]
-    },
-    "expected": {
-        "stdout":[
-            {
-                "url": "http://example.org/%18",
-                "parts": {
-                  "scheme": "http",
-                  "host": "example.org"
+    {
+        "input": {
+            "arguments": [
+                "http://example.org/%18",
+                "--json"
+            ]
+        },
+        "expected": {
+            "stdout":[
+                {
+                    "url": "http://example.org/%18",
+                    "parts": {
+                        "scheme": "http",
+                        "host": "example.org",
+                        "path": "/\u0018"
+                    }
                 }
-              }
-        ],
-        "stderr": "trurl note: URL decode error, most likely because of rubbish in the input (path)\n",
-        "returncode": 0
-    }
-  },
+            ],
+            "stderr": "",
+            "returncode": 0
+        }
+    },
+    {
+        "input": {
+            "arguments": [
+                "http://example.org/%18",
+                "--json",
+                "--urlencode"
+            ]
+        },
+        "expected": {
+            "stdout":[
+                {
+                    "url": "http://example.org/%18",
+                    "parts": {
+                        "scheme": "http",
+                        "host": "example.org",
+                        "path": "/%18"
+                    }
+                }
+            ],
+            "stderr": "",
+            "returncode": 0
+        }
+    },
   {
       "input": {
           "arguments": [
diff --git a/trurl.c b/trurl.c
index f907baed..0fe44444 100644
--- a/trurl.c
+++ b/trurl.c
@@ -1120,14 +1120,42 @@ static void json(struct option *o, CURLU *uh)
   bool params_errors = false;
   for(i = 0; variables[i].name; i++) {
     char *part;
-    rc = geturlpart(o, 0, uh, variables[i].part, &part);
+    /* ask for the URL encoded version so that weird control characters do not
+       cause problems. URL decode it when push to json. */
+    rc = geturlpart(o, VARMODIFIER_URLENCODED, uh, variables[i].part, &part);
     if(!rc) {
+      int olen;
+      char *dec = NULL;
+
+      if(!o->urlencode) {
+        if(variables[i].part == CURLUPART_QUERY) {
+          /* query parts have '+' for space */
+          char *n;
+          char *p = part;
+          do {
+            n = strchr(p, '+');
+            if(n) {
+              *n = ' ';
+              p = n + 1;
+            }
+          } while(n);
+        }
+
+        dec = curl_easy_unescape(NULL, part, 0, &olen);
+        if(!dec)
+          errorf(o, ERROR_MEM, "out of memory");
+      }
+
       if(!first)
         fputs(",\n", stdout);
       first = false;
       printf("      \"%s\": ", variables[i].name);
-      jsonString(stdout, part, strlen(part), false);
+      if(dec)
+        jsonString(stdout, dec, (size_t)olen, false);
+      else
+        jsonString(stdout, part, strlen(part), false);
       curl_free(part);
+      curl_free(dec);
     }
     else if(is_valid_trurl_error(rc)) {
         trurl_warnf(o, "%s (%s)", curl_url_strerror(rc), variables[i].name);