hidden field named _charset_ now appears in FormData as UTF-8 to fix #25150

components/script/dom/formdata.rs

@@ -55,7 +55,7 @@ impl FormData {
         form: Option<&HTMLFormElement>,
     ) -> Fallible<DomRoot<FormData>> {
         if let Some(opt_form) = form {
-            return match opt_form.get_form_dataset(None) {
+            return match opt_form.get_form_dataset(None, None) {
                 Some(form_datums) => Ok(FormData::new(Some(form_datums), global)),
                 None => Err(Error::InvalidState),
             };

components/script/dom/htmlformelement.rs

@@ -357,7 +357,7 @@ impl HTMLFormElement {
         let encoding = self.pick_encoding();
 
         // Step 9
-        let mut form_data = match self.get_form_dataset(Some(submitter)) {
+        let mut form_data = match self.get_form_dataset(Some(submitter), Some(encoding)) {
             Some(form_data) => form_data,
             None => return,
         };
@@ -645,8 +645,14 @@ impl HTMLFormElement {
     }
 
     /// <https://html.spec.whatwg.org/multipage/#constructing-the-form-data-set>
+    /// terminology note:  "form data set" = "entry list"
     /// Steps range from 3 to 5
-    fn get_unclean_dataset(&self, submitter: Option<FormSubmitter>) -> Vec<FormDatum> {
+    /// 5.x substeps are mostly handled inside element-specific methods
+    fn get_unclean_dataset(
+        &self,
+        submitter: Option<FormSubmitter>,
+        encoding: Option<&'static Encoding>,
+    ) -> Vec<FormDatum> {
         let controls = self.controls.borrow();
         let mut data_set = Vec::new();
         for child in controls.iter() {
@@ -668,7 +674,7 @@ impl HTMLFormElement {
                     HTMLElementTypeId::HTMLInputElement => {
                         let input = child.downcast::<HTMLInputElement>().unwrap();
 
-                        data_set.append(&mut input.form_datums(submitter));
+                        data_set.append(&mut input.form_datums(submitter, encoding));
                     },
                     HTMLElementTypeId::HTMLButtonElement => {
                         let button = child.downcast::<HTMLButtonElement>().unwrap();
@@ -705,7 +711,11 @@ impl HTMLFormElement {
     }
 
     /// <https://html.spec.whatwg.org/multipage/#constructing-the-form-data-set>
-    pub fn get_form_dataset(&self, submitter: Option<FormSubmitter>) -> Option<Vec<FormDatum>> {
+    pub fn get_form_dataset(
+        &self,
+        submitter: Option<FormSubmitter>,
+        encoding: Option<&'static Encoding>,
+    ) -> Option<Vec<FormDatum>> {
         fn clean_crlf(s: &str) -> DOMString {
             // Step 4
             let mut buf = "".to_owned();
@@ -746,7 +756,7 @@ impl HTMLFormElement {
         self.constructing_entry_list.set(true);
 
         // Step 3-6
-        let mut ret = self.get_unclean_dataset(submitter);
+        let mut ret = self.get_unclean_dataset(submitter, encoding);
         for datum in &mut ret {
             match &*datum.ty {
                 "file" | "textarea" => (), // TODO

components/script/dom/htmlinputelement.rs

@@ -49,6 +49,7 @@ use crate::textinput::{Direction, SelectionDirection, TextInput, UTF16CodeUnits,
 use caseless::compatibility_caseless_match_str;
 use dom_struct::dom_struct;
 use embedder_traits::FilterPattern;
+use encoding_rs::Encoding;
 use html5ever::{LocalName, Prefix};
 use msg::constellation_msg::InputMethodType;
 use net_traits::blob_url_store::get_blob_origin;
@@ -929,14 +930,18 @@ impl HTMLInputElement {
     }
 
     /// <https://html.spec.whatwg.org/multipage/#constructing-the-form-data-set>
-    /// Steps range from 3.1 to 3.7 (specific to HTMLInputElement)
-    pub fn form_datums(&self, submitter: Option<FormSubmitter>) -> Vec<FormDatum> {
+    /// Steps range from 5.1 to 5.10 (specific to HTMLInputElement)
+    pub fn form_datums(
+        &self,
+        submitter: Option<FormSubmitter>,
+        encoding: Option<&'static Encoding>,
+    ) -> Vec<FormDatum> {
         // 3.1: disabled state check is in get_unclean_dataset
 
-        // Step 3.2
+        // Step 5.2
         let ty = self.Type();
 
-        // Step 3.4
+        // Step 5.4
         let name = self.Name();
         let is_submitter = match submitter {
             Some(FormSubmitter::InputElement(s)) => self == s,
@@ -944,12 +949,12 @@ impl HTMLInputElement {
         };
 
         match self.input_type() {
-            // Step 3.1: it's a button but it is not submitter.
+            // Step 5.1: it's a button but it is not submitter.
             InputType::Submit | InputType::Button | InputType::Reset if !is_submitter => {
                 return vec![];
             },
 
-            // Step 3.1: it's the "Checkbox" or "Radio Button" and whose checkedness is false.
+            // Step 5.1: it's the "Checkbox" or "Radio Button" and whose checkedness is false.
             InputType::Radio | InputType::Checkbox => {
                 if !self.Checked() || name.is_empty() {
                     return vec![];
@@ -959,7 +964,7 @@ impl HTMLInputElement {
             InputType::File => {
                 let mut datums = vec![];
 
-                // Step 3.2-3.7
+                // Step 5.2-5.7
                 let name = self.Name();
 
                 match self.GetFiles() {
@@ -988,7 +993,21 @@ impl HTMLInputElement {
 
             InputType::Image => return vec![], // Unimplemented
 
-            // Step 3.1: it's not the "Image Button" and doesn't have a name attribute.
+            // Step 5.10: it's a hidden field named _charset_
+            InputType::Hidden => {
+                if name == "_charset_" {
+                    return vec![FormDatum {
+                        ty: ty.clone(),
+                        name: name,
+                        value: FormDatumValue::String(match encoding {
+                            None => DOMString::from("UTF-8"),
+                            Some(enc) => DOMString::from(enc.name()),
+                        }),
+                    }];
+                }
+            },
+
+            // Step 5.1: it's not the "Image Button" and doesn't have a name attribute.
             _ => {
                 if name.is_empty() {
                     return vec![];
@@ -996,7 +1015,7 @@ impl HTMLInputElement {
             },
         }
 
-        // Step 3.9
+        // Step 5.12
         vec![FormDatum {
             ty: ty.clone(),
             name: name,

tests/wpt/metadata/html/semantics/forms/form-submission-0/constructing-form-data-set.html.ini

@@ -1,7 +1,4 @@
 [constructing-form-data-set.html]
-  [FormData constructor always produces UTF-8 _charset_ value.]
-    expected: FAIL
-
   [_charset_ control sets the expected encoding name.]
     expected: FAIL