Binary Xml format broken on big-endian machines

[uweigand]

The binary format used by XmlDictionaryReader and XmlDictionaryWriter is supposed to be always little-endian.

However, the current code does not implement this correctly when running on big-endian systems. These pre-existing problems were recently exposed by additional and updated test cases added here:
#73332 (XmlDictionaryReader)
#71478 (XmlDictionaryWriter)

See #73332 (comment) for more details.

@jkotas @Daniel-Svensson

I couldn't figure out the best area label to add to this issue. If you have write-permissions please help me learn by adding exactly one area label.

Tagging subscribers to this area: @dotnet/area-system-xml
See info in area-owners.md if you want to be subscribed.

Issue Details

---

The binary format used by XmlDictionaryReader and XmlDictionaryWriter is supposed to be always little-endian.

However, the current code does not implement this correctly when running on big-endian systems. These pre-existing problems were recently exposed by additional and updated test cases added here:
#73332 (XmlDictionaryReader)
#71478 (XmlDictionaryWriter)

See #73332 (comment) for more details.

@jkotas @Daniel-Svensson

Author:	uweigand
Assignees:	-
Labels:	area-System.Xml, untriaged
Milestone:	-

[StephenMolloy]

@uweigand - I think the fix for this should be pretty straightforward as we do account for endian-ness for most types. The problem is that I don't have any big-endian hardware to test with. We would welcome a PR if you want to create one and verify it on your s390x.

And as @jkotas said in the original comment thread that spurred this issue - serialized representation ("on the wire" if you will) should always be little-endian for DataContractSerializer.

[uweigand]

@uweigand - I think the fix for this should be pretty straightforward as we do account for endian-ness for most types. The problem is that I don't have any big-endian hardware to test with. We would welcome a PR if you want to create one and verify it on your s390x.

And as @jkotas said in the original comment thread that spurred this issue - serialized representation ("on the wire" if you will) should always be little-endian for DataContractSerializer.

@StephenMolloy OK, I'll work on coming up with a PR to fix this, hopefully within the next few days. Thanks!

[KalleOlaviNiemitalo]

The format is documented in [MC-NBFX]: .NET Binary Format: XML Data Structure, where §2.1 Common Definitions indeed specifies little-endian byte order.