Remove hack to treat ERR_INCOMPLETE_CHUNKED_ENCODING error as success.

Remove hack to treat ERR_INCOMPLETE_CHUNKED_ENCODING error as success.

Only applies to downloads.  We were treating it as an error elsewhere
(Unless there's also a content-length header which has the same value
as the number of bytes the response decompresses to...Which is probably
a bug).

0.00% of main frame loads end with this error, so seems like this hack
isn't needed.

BUG=501384, 610126
Committed: https://crrev.com/3d551ff2ccf5599ebc08037d34b2c1f3b3038874
Cr-Commit-Position: refs/heads/master@{#401499}

[mmenke, 2016-06-22 20:44:42]

Description was changed from

==========
Remove hack to treat ERR_INCOMPLETE_CHUNKED_ENCODING error as success.

Only applies to downloads, we were treating it as a code elsewhere
(Unless there's also a content-length header which has the same value
as the number of bytes the response decompresses to...Which is probably
a bug).

0.00% of main frame loads end with this error, so seems like this hack
isn't needed.

BUG=610126
==========

to

==========
Remove hack to treat ERR_INCOMPLETE_CHUNKED_ENCODING error as success.

Only applies to downloads, we were treating it as a code elsewhere
(Unless there's also a content-length header which has the same value
as the number of bytes the response decompresses to...Which is probably
a bug).

0.00% of main frame loads end with this error, so seems like this hack
isn't needed.

BUG=501384,610126
==========

[mmenke, 2016-06-22 22:10:17]

mmenke@chromium.org changed reviewers:
+ asanka@chromium.org

[mmenke, 2016-06-22 22:10:17]

WDYT?  The incidence of this happening is extremely low.  Also, we treat it as
an error for non-downloads, so it seems weird to treat it differently here. 
Alternatively, we could always treat it as success, though I've seen only one or
two complains about this over the years.

Of course, since we only fail a request after we've read the entire response,
it's possible that the error isn't noticeable - I'm not sure if we're guaranteed
to bubble up the entire response body before the error or not.

Davidben has also pointed out that SSL shutdown is sufficiently fraught with
peril that servers can't really depend on this working correctly in that case.

*And*, if we're talking to a "Connection: keep-alive" server that fails to send
the terminal chunk, we think the response is hung, so think the response is
taking forever, and never time out the socket.  So relying on the old code's
behavior just doesn't really work.

[mmenke, 2016-06-22 22:11:04]

Description was changed from

==========
Remove hack to treat ERR_INCOMPLETE_CHUNKED_ENCODING error as success.

Only applies to downloads, we were treating it as a code elsewhere
(Unless there's also a content-length header which has the same value
as the number of bytes the response decompresses to...Which is probably
a bug).

0.00% of main frame loads end with this error, so seems like this hack
isn't needed.

BUG=501384,610126
==========

to

==========
Remove hack to treat ERR_INCOMPLETE_CHUNKED_ENCODING error as success.

Only applies to downloads.  We were treating it as an error elsewhere
(Unless there's also a content-length header which has the same value
as the number of bytes the response decompresses to...Which is probably
a bug).

0.00% of main frame loads end with this error, so seems like this hack
isn't needed.

BUG=501384,610126
==========

[asanka, 2016-06-23 00:54:29]

On 2016/06/22 at 22:10:17, mmenke wrote:
> WDYT?  The incidence of this happening is extremely low.  Also, we treat it as
an error for non-downloads, so it seems weird to treat it differently here. 
Alternatively, we could always treat it as success, though I've seen only one or
two complains about this over the years.
> 
> Of course, since we only fail a request after we've read the entire response,
it's possible that the error isn't noticeable - I'm not sure if we're guaranteed
to bubble up the entire response body before the error or not.
> 
> Davidben has also pointed out that SSL shutdown is sufficiently fraught with
peril that servers can't really depend on this working correctly in that case.
> 
> *And*, if we're talking to a "Connection: keep-alive" server that fails to
send the terminal chunk, we think the response is hung, so think the response is
taking forever, and never time out the socket.  So relying on the old code's
behavior just doesn't really work.

LGTM.

We are also working on dealing better with content length mismatches. See
https://bugs.chromium.org/p/chromium/issues/detail?id=501384

[mmenke, 2016-06-23 01:13:36]

The CQ bit was checked by mmenke@chromium.org

[commit-bot: I haz the power, 2016-06-23 01:13:54]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/2092523002/1

[commit-bot: I haz the power, 2016-06-23 01:21:21]

Description was changed from

==========
Remove hack to treat ERR_INCOMPLETE_CHUNKED_ENCODING error as success.

Only applies to downloads.  We were treating it as an error elsewhere
(Unless there's also a content-length header which has the same value
as the number of bytes the response decompresses to...Which is probably
a bug).

0.00% of main frame loads end with this error, so seems like this hack
isn't needed.

BUG=501384,610126
==========

to

==========
Remove hack to treat ERR_INCOMPLETE_CHUNKED_ENCODING error as success.

Only applies to downloads.  We were treating it as an error elsewhere
(Unless there's also a content-length header which has the same value
as the number of bytes the response decompresses to...Which is probably
a bug).

0.00% of main frame loads end with this error, so seems like this hack
isn't needed.

BUG=501384,610126
==========

[commit-bot: I haz the power, 2016-06-23 01:21:22]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-06-23 01:23:33]

Description was changed from

==========
Remove hack to treat ERR_INCOMPLETE_CHUNKED_ENCODING error as success.

Only applies to downloads.  We were treating it as an error elsewhere
(Unless there's also a content-length header which has the same value
as the number of bytes the response decompresses to...Which is probably
a bug).

0.00% of main frame loads end with this error, so seems like this hack
isn't needed.

BUG=501384,610126
==========

to

==========
Remove hack to treat ERR_INCOMPLETE_CHUNKED_ENCODING error as success.

Only applies to downloads.  We were treating it as an error elsewhere
(Unless there's also a content-length header which has the same value
as the number of bytes the response decompresses to...Which is probably
a bug).

0.00% of main frame loads end with this error, so seems like this hack
isn't needed.

BUG=501384,610126

Committed: https://crrev.com/3d551ff2ccf5599ebc08037d34b2c1f3b3038874
Cr-Commit-Position: refs/heads/master@{#401499}
==========

[commit-bot: I haz the power, 2016-06-23 01:23:34]

Patchset 1 (id:??) landed as
https://crrev.com/3d551ff2ccf5599ebc08037d34b2c1f3b3038874
Cr-Commit-Position: refs/heads/master@{#401499}