From 257e35ee2e21659207412d7602673e2fc9044730 Mon Sep 17 00:00:00 2001
From: Daniel Stenberg <daniel@haxx.se>
Date: Mon, 13 May 2024 13:30:48 +0200
Subject: [PATCH] trurl: support ?= for set, to only do it if not already set

Idea-by: @emanuele6 in #296

Due to missing support in the libcurl URL API, this is not really
working for the scheme component right now, as that will always be
considered set once a URL has been parsed. The scheme guessing the
parser does sets it internally and there is currently no way provided by
the API to figure that out.
---
 tests.json | 28 ++++++++++++++++++++++++++++
 trurl.1    |  5 +++++
 trurl.c    | 39 ++++++++++++++++++++++++++++++++-------
 3 files changed, 65 insertions(+), 7 deletions(-)

diff --git a/tests.json b/tests.json
index 4eb55de8..4d81c814 100644
--- a/tests.json
+++ b/tests.json
@@ -2484,5 +2484,33 @@
           "stderr": "",
           "returncode": 0
       }
+  },
+  {
+      "input": {
+          "arguments": [
+              "example.com:88",
+              "--set",
+              "port?=99"
+          ]
+      },
+      "expected": {
+          "stdout": "http://example.com:88/\n",
+          "stderr": "",
+          "returncode": 0
+      }
+  },
+  {
+      "input": {
+          "arguments": [
+              "example.com",
+              "--set",
+              "port?=99"
+          ]
+      },
+      "expected": {
+          "stdout": "http://example.com:99/\n",
+          "stderr": "",
+          "returncode": 0
+      }
   }
 ]
diff --git a/trurl.1 b/trurl.1
index a492885d..85b1525d 100644
--- a/trurl.1
+++ b/trurl.1
@@ -217,6 +217,11 @@ options, host, port, path, query, fragment and zoneid.
 If a simple "="-assignment is used, the data is URL encoded when applied. If
 ":=" is used, the data is assumed to already be URL encoded and stored as-is.
 
+If "?=" is used, the set is only performed if the component is not already
+set. It avoids overwriting any already set data.
+
+You can also combine : and ? into "?:=" if desired.
+
 If no URL or \fI--url-file\fP argument is provided, trurl tries to create
 a URL using the components provided by the \fI--set\fP options. If not enough
 components are specified, this fails.
diff --git a/trurl.c b/trurl.c
index 43142fcf..b9bd3ab4 100644
--- a/trurl.c
+++ b/trurl.c
@@ -899,20 +899,45 @@ static const struct var *setone(CURLU *uh, const char *setline,
   if(ptr && (ptr > setline)) {
     size_t vlen = ptr - setline;
     bool urlencode = true;
+    bool conditional = false;
     bool found = false;
-    if(ptr[-1] == ':') {
-      urlencode = false;
-      vlen--;
+    if(vlen) {
+      int back = -1;
+      size_t reqlen = 1;
+      while(vlen > reqlen) {
+        if(ptr[back] == ':') {
+          urlencode = false;
+          vlen--;
+        }
+        else if(ptr[back] == '?') {
+          conditional = true;
+          vlen--;
+        }
+        else
+          break;
+        reqlen++;
+        back--;
+      }
     }
     v = comp2var(setline, vlen);
     if(v) {
-      CURLUcode rc;
+      CURLUcode rc = CURLUE_OK;
+      bool skip = false;
       if((v->part == CURLUPART_HOST) && ('[' == ptr[1]))
         /* when setting an IPv6 numerical address, disable URL encoding */
         urlencode = false;
-      rc = curl_url_set(uh, v->part, ptr[1] ? &ptr[1] : NULL,
-                        (o->curl ? 0 : CURLU_NON_SUPPORT_SCHEME)|
-                        (urlencode ? CURLU_URLENCODE : 0) );
+
+      if(conditional) {
+        char *piece;
+        rc = curl_url_get(uh, v->part, &piece, 0);
+        if(!rc)
+          skip = true;
+      }
+
+      if(!skip)
+        rc = curl_url_set(uh, v->part, ptr[1] ? &ptr[1] : NULL,
+                          (o->curl ? 0 : CURLU_NON_SUPPORT_SCHEME)|
+                          (urlencode ? CURLU_URLENCODE : 0) );
       if(rc)
         warnf("Error setting %s: %s", v->name, curl_url_strerror(rc));
       found = true;