Bug 1697421 - improve parsing of content-type headers for fetch r=necko-reviewers,valentin

Differential Revision: https://phabricator.services.mozilla.com/D157115

dom/fetch/Fetch.cpp

@@ -21,6 +21,7 @@
 #include "nsProxyRelease.h"
 
 #include "mozilla/ErrorResult.h"
+#include "mozilla/dom/MimeType.h"
 #include "mozilla/dom/BindingDeclarations.h"
 #include "mozilla/dom/BodyConsumer.h"
 #include "mozilla/dom/Exceptions.h"
@@ -1259,20 +1260,40 @@ template already_AddRefed<Promise> FetchBody<EmptyBody>::ConsumeBody(
 template <class Derived>
 void FetchBody<Derived>::GetMimeType(nsACString& aMimeType,
                                      nsACString& aMixedCaseMimeType) {
-  // Extract mime type.
   ErrorResult result;
-  nsCString contentTypeValues;
+  nsAutoCString contentTypeValues, contentTypeUtf8;
   MOZ_ASSERT(DerivedClass()->GetInternalHeaders());
   DerivedClass()->GetInternalHeaders()->Get("Content-Type"_ns,
                                             contentTypeValues, result);
   MOZ_ALWAYS_TRUE(!result.Failed());
 
-  // HTTP ABNF states Content-Type may have only one value.
+  nsCCharSeparatedTokenizer contentTypeTokens(contentTypeValues, ',');
-  // This is from the "parse a header value" of the fetch spec.
+
-  if (!contentTypeValues.IsVoid() && contentTypeValues.Find(",") == -1) {
+  // currently we extract the MIME-type from the first content type header
-    // Convert from a bytestring to a UTF8 CString.
+  // In order to fully comply with the fetch spec we must be able to
-    CopyLatin1toUTF8(contentTypeValues, aMimeType);
+  // parse multiple content-type headers and extrace headers from it
-    aMixedCaseMimeType = aMimeType;
+  // To achieve this we need to modify CMimeType::Parse and implement the
+  // algorithm https://fetch.spec.whatwg.org/#concept-header-extract-mime-type
+  // This issue is tracked by
+  // https://bugzilla.mozilla.org/show_bug.cgi?id=1510180
+  auto contentStr = contentTypeTokens.nextToken();
+
+  CopyLatin1toUTF8(contentStr, contentTypeUtf8);
+  UniquePtr<CMimeType> contentTypeRecord = CMimeType::Parse(contentTypeUtf8);
+  if (contentTypeRecord) {
+    contentTypeRecord->Serialize(aMixedCaseMimeType);
+    // validate invalid/empty parameter checks
+    // according to https://mimesniff.spec.whatwg.org/#parsing-a-mime-type
+    // CMimeType::Parse parses the parameters and discards ill-formed parameters
+    // If we have invalid/illformed paramters we need to discard the parsed
+    // mime-type
+    if (contentStr.Contains(';') && !aMixedCaseMimeType.Contains(';')) {
+      // parameters were discarded after parsing.
+      // This should result in invalid MimeType
+      aMixedCaseMimeType = "";
+    }
+
+    aMimeType = aMixedCaseMimeType;
     ToLowerCase(aMimeType);
   }
 }

testing/web-platform/meta/fetch/content-type/response.window.js.ini

@@ -5,9 +5,6 @@
   [Request: combined response Content-Type: */* text/html]
     expected: FAIL
 
-  [Response: combined response Content-Type: text/html */*]
-    expected: FAIL
-
   [fetch(): separate response Content-Type: */* text/html]
     expected: FAIL
 
@@ -23,9 +20,6 @@
   [fetch(): combined response Content-Type: */* text/html]
     expected: FAIL
 
-  [Request: combined response Content-Type: text/html */*]
-    expected: FAIL
-
   [Request: combined response Content-Type: text/plain;charset=gbk text/plain;charset=windows-1252]
     expected: FAIL
 
@@ -50,24 +44,12 @@
   [<iframe>: separate response Content-Type: text/plain;charset=gbk text/html]
     expected: FAIL
 
-  [fetch(): combined response Content-Type: text/html */*;charset=gbk]
-    expected: FAIL
-
-  [fetch(): combined response Content-Type: text/plain;charset=gbk text/plain]
-    expected: FAIL
-
   [Request: combined response Content-Type: text/html text/plain]
     expected: FAIL
 
-  [Request: combined response Content-Type: text/plain */*]
-    expected: FAIL
-
   [fetch(): separate response Content-Type: text/html;charset=gbk text/plain text/html]
     expected: FAIL
 
-  [Request: combined response Content-Type: text/plain;charset=gbk text/plain]
-    expected: FAIL
-
   [fetch(): combined response Content-Type: text/plain;charset=gbk text/html]
     expected: FAIL
 
@@ -86,18 +68,9 @@
   [Request: combined response Content-Type: text/html;x=" text/plain]
     expected: FAIL
 
-  [Response: combined response Content-Type: text/plain */*]
-    expected: FAIL
-
-  [Request: combined response Content-Type: text/html */*;charset=gbk]
-    expected: FAIL
-
   [Response: combined response Content-Type: text/html;x=" text/plain]
     expected: FAIL
 
-  [fetch(): combined response Content-Type: text/html */*]
-    expected: FAIL
-
   [Response: combined response Content-Type: text/html;" text/plain]
     expected: FAIL
 
@@ -131,45 +104,21 @@
   [fetch(): combined response Content-Type: text/html;charset=gbk text/html;x=",text/plain]
     expected: FAIL
 
-  [fetch(): combined response Content-Type: text/plain */*]
-    expected: FAIL
-
-  [Request: combined response Content-Type: text/plain */*;charset=gbk]
-    expected: FAIL
-
   [fetch(): separate response Content-Type: text/plain;charset=gbk text/html;charset=windows-1254]
     expected: FAIL
 
-  [fetch(): combined response Content-Type: text/plain */*;charset=gbk]
-    expected: FAIL
-
   [fetch(): combined response Content-Type: text/plain;charset=gbk text/plain;charset=windows-1252]
     expected: FAIL
 
-  [Response: combined response Content-Type: text/plain;charset=gbk text/plain]
-    expected: FAIL
-
   [fetch(): separate response Content-Type: text/html text/plain]
     expected: FAIL
 
   [Request: combined response Content-Type: text/html;" text/plain]
     expected: FAIL
 
-  [Response: combined response Content-Type: text/plain */*;charset=gbk]
-    expected: FAIL
-
   [Request: combined response Content-Type: text/html;charset=gbk text/html;x=",text/plain]
     expected: FAIL
 
-  [Response: combined response Content-Type: text/html */*;charset=gbk]
-    expected: FAIL
-
-  [fetch(): combined response Content-Type: text/plain ]
-    expected: FAIL
-
-  [Response: combined response Content-Type: text/plain ]
-    expected: FAIL
-
   [fetch(): separate response Content-Type:  text/plain]
     expected: FAIL
 
@@ -182,9 +131,6 @@
   [fetch(): combined response Content-Type:  text/plain]
     expected: FAIL
 
-  [Request: combined response Content-Type: text/plain ]
-    expected: FAIL
-
   [fetch(): separate response Content-Type: text/html;" \\" text/plain]
     expected: FAIL