bug#70312: [PATCH v2] Avoid unnecessary escaping in url-build-query-string

["Dagfinn Ilmari Mannsåker" <ilmari@ilmari.org>]

[-- Attachment #1: Type: text/plain, Size: 403 bytes --]

Hi again,

I realised I'd forgotten to add tests, and that made me realise that
url-query-allowed-chars is not correct for this, since that also
contains '=', '&', and ';'. So here's an updated patch, which creates a
new url-query-key-value-allowed-chars constant, which is
url-query-allowed-chars minus the aforementioned three chars, and adds
tests covering that, for both keys and values.

- ilmari


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: v2-0001-Avoid-unnecessary-escaping-in-url-build-query-str.patch --]
[-- Type: text/x-diff, Size: 2898 bytes --]

From 89db0a1226d8d7cca1846e9c737d4a67c971ec75 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dagfinn=20Ilmari=20Manns=C3=A5ker?= <ilmari@ilmari.org>
Date: Tue, 9 Apr 2024 15:02:45 +0100
Subject: [PATCH v2] Avoid unnecessary escaping in url-build-query-string

* lisp/url/url-util.el (url-build-query-string):
Create a new url-query-key-value-allowed-chars constant and pass that to
url-hexify-string to avoid unnecessarily escaping characters that don't
need to be escaped in query string keys and values.
* test/lisp/url/url-util-tests.el (url-util-tests):
Add test cases.
---
 lisp/url/url-util.el            | 12 +++++++++++-
 test/lisp/url/url-util-tests.el |  6 +++++-
 2 files changed, 16 insertions(+), 2 deletions(-)

diff --git a/lisp/url/url-util.el b/lisp/url/url-util.el
index 5f45b98c7a5..f063efe18a6 100644
--- a/lisp/url/url-util.el
+++ b/lisp/url/url-util.el
@@ -268,7 +268,8 @@ url-build-query-string
    (lambda (key-vals)
      (let ((escaped
             (mapcar (lambda (sym)
-                      (url-hexify-string (format "%s" sym))) key-vals)))
+                      (url-hexify-string (format "%s" sym) url-query-key-value-allowed-chars))
+                    key-vals)))
        (mapconcat (lambda (val)
                     (let ((vprint (format "%s" val))
                           (eprint (format "%s" (car escaped))))
@@ -410,6 +411,15 @@ url-query-allowed-chars
   "Allowed-character byte mask for the query segment of a URI.
 These characters are specified in RFC 3986, Appendix A.")
 
+(defconst url-query-key-value-allowed-chars
+  (let ((vec (copy-sequence url-query-allowed-chars)))
+    (aset vec ?= nil)
+    (aset vec ?& nil)
+    (aset vec ?\; nil)
+    vec)
+  "Allowed-charcter byte mask for keys and values in the query segment of a URI.
+url-query-allowed-chars minus '=', '&', and ';'.")
+
 ;;;###autoload
 (defun url-encode-url (url)
   "Return a properly URI-encoded version of URL.
diff --git a/test/lisp/url/url-util-tests.el b/test/lisp/url/url-util-tests.el
index 133aa0ffd88..c6246d69a2a 100644
--- a/test/lisp/url/url-util-tests.el
+++ b/test/lisp/url/url-util-tests.el
@@ -32,7 +32,11 @@ url-util-tests
            ("key1=val1;key2=val2;key3=val1;key3=val2;key4;key5"
             ((key1 "val1") (key2 val2) (key3 val1 val2) ("key4") (key5 "")) t)
            ("key1=val1;key2=val2;key3=val1;key3=val2;key4=;key5="
-            ((key1 val1) (key2 val2) ("key3" val1 val2) (key4) (key5 "")) t t)))
+            ((key1 val1) (key2 val2) ("key3" val1 val2) (key4) (key5 "")) t t)
+           ("key1=val/slash;key2=val%3Bsemi;key3=val%26amp;key4=val%3Deq"
+            ((key1 "val/slash") (key2 "val;semi") (key3 "val&amp") (key4 "val=eq")) t)
+           ("key%3Deq=val1;key%3Bsemi=val2;key%26amp=val3"
+            (("key=eq" val1) ("key;semi" val2) ("key&amp" val3)) t)))
         test)
     (while tests
       (setq test (car tests)
-- 
2.39.2