[webkit-reviews] review requested: [Bug 213254] Rename TextEncoding::encode() to TextEncoding::encodeWithNormalization() : [Attachment 402015] Patch

[bugzilla-daemon at webkit.org]

Tetsuharu Ohzeki <tetsuharu.ohzeki at gmail.com> has asked  for review:
Bug 213254: Rename TextEncoding::encode() to
TextEncoding::encodeWithNormalization()
https://bugs.webkit.org/show_bug.cgi?id=213254

Attachment 402015: Patch

https://bugs.webkit.org/attachment.cgi?id=402015&action=review




--- Comment #3 from Tetsuharu Ohzeki <tetsuharu.ohzeki at gmail.com> ---
Comment on attachment 402015
  --> https://bugs.webkit.org/attachment.cgi?id=402015
Patch

(In reply to Alexey Proskuryakov from comment #2)
> Comment on attachment 402015 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=402015&action=review
> 
> > Source/WebCore/ChangeLog:9
> > +	     I'm planning to add a variant of `TextEncoding::encode()` which
does not
> > +	     normalize to fix bug 213253.
> 
> I don't think that this is necessary. Once we perform normalization on
> ingress (macOS file names, keyboard input etc.), TextEncoding::encode()
> won't need to normalize ever. There isn't a need for a separate code path.

The exist `TextEncoding::encode()` do normalization in it and it is widely used
in evenrywhere in WebKit.
https://trac.webkit.org/browser/webkit/trunk/Source/WebCore/platform/text/TextE
ncoding.cpp#L77

To fix bug 213253, we need construcing `Blob` object on JavaScript without
unicode normalization. I also had a plan to add the function which does not do
unicode nomalization as `TextEncoding::encodeWithoutNormalization()` but I
thought it's better to rename the exist implementation and I create this
separated patch.