Use BlobCallback in toBlob instead of FileCallback

Use BlobCallback in toBlob instead of FileCallback

Recently there's a spec change: https://github.com/whatwg/html/pull/391.
As FileCallback is used extensively in existing DOM File System, I try not to
simply rename it to BlobCallback. Adding a new BlobCallback.idl is a temporary
definition to suit the needs of toBlob, before crbug.com/569301 feature is
available.

BUG=568603

[xlai (Olivia), 2016-01-19 21:31:00]

Description was changed from

==========
Use BlobCallback in toBlob instead of FileCallback

Recently there's a spec change: https://github.com/whatwg/html/pull/391.
As FileCallback is used extensively in existing DOM File System, I try not to
simply rename it to BlobCallback. Adding a new BlobCallback.idl is a temporary
definition to suit the needs of toBlob, before crbug.com/569301 feature is
available.

BUG=568603
==========

to

==========
Use BlobCallback in toBlob instead of FileCallback

Recently there's a spec change: https://github.com/whatwg/html/pull/391.
As FileCallback is used extensively in existing DOM File System, I try not to
simply rename it to BlobCallback. Adding a new BlobCallback.idl is a temporary
definition to suit the needs of toBlob, before crbug.com/569301 feature is
available.

BUG=568603
==========

[xlai (Olivia), 2016-01-19 21:31:00]

xlai@chromium.org changed reviewers:
+ junov@chromium.org

[xlai (Olivia), 2016-01-19 21:35:14]

Patchset #1 (id:1) has been deleted

[xlai (Olivia), 2016-01-19 21:42:46]

xlai@chromium.org changed reviewers:
+ noel@chromium.org

[xlai (Olivia), 2016-01-19 21:42:46]

I had replaced FileCallback in toBlob to BlobCallback due to recent spec change.
In http://crbug.com/568603, there're different solutions suggested to implement
this. I chose to add BlobCallback.idl as a temporary and simpler solution,
because I find that FileCallback.h is used extensively in
FileEntry,DOMFileSystem,and so on.

[xlai (Olivia), 2016-01-19 21:51:32]

Patchset #1 (id:20001) has been deleted

[Justin Novosad, 2016-01-19 22:09:37]

junov@chromium.org changed reviewers:
+ philipj@opera.com

[Justin Novosad, 2016-01-19 22:09:37]

Adding philipj as reviewer (API owner + author of the spec change)

[Justin Novosad, 2016-01-19 22:15:49]

lgtm with nits.

https://codereview.chromium.org/1599413005/diff/2/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/fileapi/BlobCallback.h (right):

https://codereview.chromium.org/1599413005/diff/2/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/fileapi/BlobCallback.h:2: * Copyright (C) 2010
Google Inc. All rights reserved.
New file -> (C) 2016

https://codereview.chromium.org/1599413005/diff/2/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/fileapi/BlobCallback.idl (right):

https://codereview.chromium.org/1599413005/diff/2/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/fileapi/BlobCallback.idl:2: * Copyright (C) 2010
Google Inc. All rights reserved.
New file -> (C) 2016

[Justin Novosad, 2016-01-19 22:17:37]

https://codereview.chromium.org/1599413005/diff/2/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/fileapi/BlobCallback.h (right):

https://codereview.chromium.org/1599413005/diff/2/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/fileapi/BlobCallback.h:40: class BlobCallback :
public GarbageCollectedFinalized<BlobCallback> {
Why "Finalized"? destructor is a no-op.

[philipj_slow, 2016-01-19 22:28:14]

Aren't there any tests that need updating? Given that has been confused, I'd say
that a test that checks that the returned blob is *not* instanceof File wouldn't
be overkill.

https://codereview.chromium.org/1599413005/diff/2/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/fileapi/Blob.h (right):

https://codereview.chromium.org/1599413005/diff/2/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/fileapi/Blob.h:68: static Blob* create(const
unsigned char* data, size_t bytes, const String& mimeType);
s/mimeType/contentType/ to match other bits of Blob.h/cpp

https://codereview.chromium.org/1599413005/diff/2/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/fileapi/BlobCallback.idl (right):

https://codereview.chromium.org/1599413005/diff/2/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/fileapi/BlobCallback.idl:31: callback interface
BlobCallback {
Actually, the conclusion of
http://code.google.com/p/chromium/issues/detail?id=569301 was that callback
functions already *are* supported, but missing some of the convenience of
callback interfaces.

Since it should already be possible to use a callback function, can you try
doing that and see if there's any trouble. It would be sad to introduce a new
callback interface if it can be avoided.

[xlai (Olivia), 2016-01-20 17:54:11]

philipj@: I am trying out a different solution in this issue:
https://codereview.chromium.org/1609313002/. The changes about toBlob between
that code review issue and this one are the same. The only difference is that I
replace everywhere in FileSystem that use FileCallback to use BlobCallback
instead. As a result, there's 1 place that I had to change using Blob instead of
File in function argument. If the other issue works, this code review issue will
be abandoned. Looks like it compiles in my local machine and pass existing
fast/filesystem layout tests.

junov@: Do I still need to change the (C) to 2016 if the whole file is just a
renamed file from a older one?

[Justin Novosad, 2016-01-20 17:57:22]

On 2016/01/20 17:54:11, Olivia wrote:

> junov@: Do I still need to change the (C) to 2016 if the whole file is just a
> renamed file from a older one?

It's not technically a rename (old file still exists). It's a new file -> new
date.