[Webkit-unassigned] [Bug 36903] Implement BlobBuilder internal class for BlobBuilder support as defined in FileWriter

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=36903





--- Comment #5 from Kinuko Yasuda <kinuko at chromium.org>  2010-04-02 11:42:42 PST ---
(In reply to comment #4)
> (From update of attachment 52227 [details])
> General comment: it seems you are taking the path of converting the parts
> appended by append() method to ByteStore immediately and synchronously on
> append(). This will have 2 undesired effects:

For append(Blob) I was/am going to put them into a list and wasn't going to
read them immediately, but I was missing the point that BlobBuilder.endings may
be changed after Strings are added.

In the spec there's a notion that BlobBuilder.append(string) may throw encoding
exceptions, so I had assumed append(string) does some conversion right away. 
(Actually I think we can drop this part as we now do not have encoding
parameters in BlobBuilder.)

It makes me wonder when we should really convert the strings;
1. BlobBuilder.append, 2. BlobBuilder.getBlob or 3. when we read Blob.
1. is not efficient as you suggested, but I think we will want some raw data at
2., because Blob.size and Blob.slice seem to be byte-based methods/operations. 
Does this sound correct?

I'm attaching a new patch that do not convert Strings into ByteStore at
append() but at getBlob() timing.  This time I'm including changes in Blob too
to give more overview.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.