Don't NFC-normalize strings during Blob/File construction

Don't NFC-normalize strings during Blob/File construction

Back in the mists of time, BlobBuilder called UTF8Encoding().encode()
when appending strings, which looks correct. But then mean old
https://crbug.com/117128 came along pointing out that encode() was
actually NFC normalizing strings even when not wanted, so
https://crrev.com/19845004 cleverly renamed the function to
normalizeAndEncode() so future callers would not fall into this trap.
A few existing places such as XHR were eventually changed to just call
encode() (https://crrev.com/22875053) but no-one noticed BlobBuilder
was getting it wrong prior to this bug report.

Change the BlobData::appendText() method to just call encode(), and
add a test case for the Blob constructor.

The new behavior matches FF and IE.

BUG=427183
R=kinuko@chromium.org

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=201497

[jsbell, 2015-08-28 23:45:42]

kinuko@ - please take a look?

[kinuko, 2015-08-29 01:39:08]

On 2015/08/28 23:45:42, jsbell wrote:
> kinuko@ - please take a look?

lgtm, thx for catching this!

[jsbell, 2015-08-29 20:21:32]

The CQ bit was checked by jsbell@chromium.org

[commit-bot: I haz the power, 2015-08-29 20:21:35]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1305393006/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1305393006/1

[commit-bot: I haz the power, 2015-08-29 20:28:56]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-08-29 20:28:57]

Try jobs failed on following builders:
  chromium_presubmit on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)

[kinuko, 2015-08-31 07:16:52]

Sorry, I'm not an owner of Source/platform...

[jsbell, 2015-08-31 17:17:36]

jsbell@chromium.org changed reviewers:
+ adamk@chromium.org

[jsbell, 2015-08-31 17:17:37]

adamk@ - can you OWNERS stamp?

(And thanks for reviewing kinuko@!)

[adamk, 2015-08-31 17:20:05]

lgtm stamp for Source/platform

[jsbell, 2015-08-31 18:10:52]

The CQ bit was checked by jsbell@chromium.org

[commit-bot: I haz the power, 2015-08-31 18:11:05]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1305393006/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1305393006/1

[commit-bot: I haz the power, 2015-08-31 19:45:29]

Committed patchset #1 (id:1) as
https://src.chromium.org/viewvc/blink?view=rev&revision=201497

[tkent, 2015-09-01 00:03:18]

Does this fix crbug.com/341019 too?

[jsbell, 2015-09-01 00:12:55]

On 2015/09/01 00:03:18, tkent wrote:
> Does this fix crbug.com/341019 too?

From a quick glance at the issue, I doubt it. This one is about content being
added to the blob data stream. That bug is about the "metadata" (filename).

I think the issue is here:

https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...


... which is a similar case where a encoding is being done with
normalizeAndEncode() where encode() might be more appropriate.

I'll take that bug, and verify.

[tkent, 2015-09-01 00:26:03]

On 2015/09/01 00:12:55, jsbell wrote:
> On 2015/09/01 00:03:18, tkent wrote:
> > Does this fix crbug.com/341019 too?
> 
> From a quick glance at the issue, I doubt it. This one is about content being
> added to the blob data stream. That bug is about the "metadata" (filename).
> 
> I think the issue is here:
> 
>
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...
> 
> 
> ... which is a similar case where a encoding is being done with
> normalizeAndEncode() where encode() might be more appropriate.
> 
> I'll take that bug, and verify.

Great.  Thank you for the investigation and taking it.