Should the optional members of WriteParams be nullable?

[weinig]

The current IDL definition of WritableParams is:

dictionary WriteParams {
  required WriteCommandType type;
  unsigned long long? size;
  unsigned long long? position;
  (BufferSource or Blob or USVString)? data;
};

It's a bit weird that size, position and data are both nullable (by virtue of the ?) and optional (by virtue of having no required keyword).

Is this intentional? If so, what is the purpose of being both optional and nullable. Do they indicate separate things?

[saschanaz]

For the background, it's added in WICG/file-system-access#62.

[saschanaz]

https://fs.spec.whatwg.org/#write-a-chunk does not expect any of those member to be null, and rather it only checks whether those members exist. E.g. Step 6 - 8 only assumes it can ever be BufferSource/Blob/USVString and never null.

[saschanaz]

Given all implementations somehow dealt with null without filing a spec issue, I think we probably need to just modify the spec 😕

• https://searchfox.org/firefox-main/rev/0fc599421fc27ac61c501f277cb440777f85b18a/dom/fs/api/FileSystemWritableFileStream.cpp#614-616
• https://searchfox.org/wubkat/rev/2288ce31211e8b9fa825e5af8cb1afd1aa58cb6e/Source/WebCore/Modules/filesystem/FileSystemWritableFileStreamSink.cpp#73-74
• https://source.chromium.org/chromium/chromium/src/+/main:third_party/blink/renderer/modules/file_system_access/file_system_underlying_sink.cc;l=126-137;drc=14124cfccffd291611a9062ec583940c27895941