Revert of Cancel image loads if decoding failed.

Revert of Cancel image loads if decoding failed. (patchset #8 id:140001 of https://codereview.chromium.org/2108033003/ )

Reason for revert:

Likely cause of failures:

webkit_tests webkit_tests:
http/tests/images/restyle-decode-error.html

https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.11%20%28dbg%29/builds/3384

---
10:19:32.937 5109 worker/0 http/tests/images/restyle-decode-error.html started
10:19:32.953 5111 worker/2 accessibility/animation-policy.html passed
10:19:32.955 5111 worker/2 accessibility/anonymous-render-block-in-continuation-causes-crash.html started
10:19:32.999 5109 renderer crash, pid = None, error_line = #CRASHED - renderer
10:19:33.000 5109 killed pid 5357
10:19:33.000 5109 worker/0 http/tests/images/restyle-decode-error.html crashed, (stderr lines):
10:19:33.000 5109   Received signal 11 <unknown> 000000000000
10:19:33.000 5109    [0x0001196ef23e]
...
10:19:33.002 5109   [end of stack trace]
10:19:33.002 5109   [5358:13319:0722/101932:598964195783:ERROR:node_controller.cc(1099)] Could not be introduced to peer C9E4F59A639F0E59.C2F8E762ADA8C89E
10:19:33.006 5084 [4832/40217] http/tests/images/restyle-decode-error.html failed unexpectedly (renderer crashed)
10:19:33.003 5109 worker/0 killing primary driver
10:19:33.004 5109 worker/0 killing secondary driver
10:19:33.004 5109 worker/0 http/tests/images/restyle-decode-error.html failed:
10:19:33.004 5109 worker/0  renderer crashed
---

Also same failure on other bots such as
https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Leak/builds/21244

(ignore other webkit failures on this bot; those all have a known cause with a fix incoming)

Original issue's description:
> Cancel image loads if decoding failed.
>
> 'Cancel' isn't really the right word for what this change does.
> In the event of an image decoding failure, ImageResource will
> synthesize a call to ResourceLoader::didFinishLoading, killing
> the actual network request but reporting it as a successful
> completion to the rest of blink. This matches our traditional
> behavior of decoding errors looking like a successful
> resource load (for the most part).
>
> This requires some plumbing changes to image decoding, because
> the decoder selection logic doesn't report why it wasn't able to
> create an ImageDecoder. It might be because insufficient data has
> been received to sniff the image type, or it may be that we
> definitely don't have a valid image type. This change exposes enough
> information to tell the difference.
>
> BUG=471272
>
> Committed: https://crrev.com/520519605b961a7d979bc0eaa3c057a973dc26f1
> Cr-Commit-Position: refs/heads/master@{#407171}

TBR=pdr@chromium.org,peter@chromium.org,japhet@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=471272
Committed: https://crrev.com/bc36d89c757eedc36f73e1caa9948d384a2b8e60
Cr-Commit-Position: refs/heads/master@{#407228}

[Mark P, 2016-07-22 19:16:53]

The CQ bit was checked by mpearson@chromium.org

[Mark P, 2016-07-22 19:16:54]

Created Revert of Cancel image loads if decoding failed.

[commit-bot: I haz the power, 2016-07-22 19:17:34]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[Mark P, 2016-07-22 19:18:04]

Also note the stack in
https://storage.googleapis.com/chromium-layout-test-archives/WebKit_Linux_ASA...
makes this almost certain.

--mark

[commit-bot: I haz the power, 2016-07-22 19:18:52]

Description was changed from

==========
Revert of Cancel image loads if decoding failed. (patchset #8 id:140001 of
https://codereview.chromium.org/2108033003/ )

Reason for revert:

Likely cause of failures:

webkit_tests webkit_tests:
http/tests/images/restyle-decode-error.html

https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.11%20%28...

---
10:19:32.937 5109 worker/0 http/tests/images/restyle-decode-error.html started
10:19:32.953 5111 worker/2 accessibility/animation-policy.html passed
10:19:32.955 5111 worker/2
accessibility/anonymous-render-block-in-continuation-causes-crash.html started
10:19:32.999 5109 renderer crash, pid = None, error_line = #CRASHED - renderer
10:19:33.000 5109 killed pid 5357
10:19:33.000 5109 worker/0 http/tests/images/restyle-decode-error.html crashed,
(stderr lines):
10:19:33.000 5109   Received signal 11 <unknown> 000000000000
10:19:33.000 5109    [0x0001196ef23e]
...
10:19:33.002 5109   [end of stack trace]
10:19:33.002 5109  
[5358:13319:0722/101932:598964195783:ERROR:node_controller.cc(1099)] Could not
be introduced to peer C9E4F59A639F0E59.C2F8E762ADA8C89E
10:19:33.006 5084 [4832/40217] http/tests/images/restyle-decode-error.html
failed unexpectedly (renderer crashed)
10:19:33.003 5109 worker/0 killing primary driver
10:19:33.004 5109 worker/0 killing secondary driver
10:19:33.004 5109 worker/0 http/tests/images/restyle-decode-error.html failed:
10:19:33.004 5109 worker/0  renderer crashed
---

Also same failure on other bots such as
https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Leak/b...

(ignore other webkit failures on this bot; those all have a known cause with a
fix incoming)


Original issue's description:
> Cancel image loads if decoding failed.
> 
> 'Cancel' isn't really the right word for what this change does.
> In the event of an image decoding failure, ImageResource will
> synthesize a call to ResourceLoader::didFinishLoading, killing
> the actual network request but reporting it as a successful
> completion to the rest of blink. This matches our traditional
> behavior of decoding errors looking like a successful
> resource load (for the most part).
> 
> This requires some plumbing changes to image decoding, because
> the decoder selection logic doesn't report why it wasn't able to
> create an ImageDecoder. It might be because insufficient data has
> been received to sniff the image type, or it may be that we
> definitely don't have a valid image type. This change exposes enough
> information to tell the difference.
> 
> BUG=471272
> 
> Committed: https://crrev.com/520519605b961a7d979bc0eaa3c057a973dc26f1
> Cr-Commit-Position: refs/heads/master@{#407171}

TBR=pdr@chromium.org,peter@chromium.org,japhet@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=471272
==========

to

==========
Revert of Cancel image loads if decoding failed. (patchset #8 id:140001 of
https://codereview.chromium.org/2108033003/ )

Reason for revert:

Likely cause of failures:

webkit_tests webkit_tests:
http/tests/images/restyle-decode-error.html

https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.11%20%28...

---
10:19:32.937 5109 worker/0 http/tests/images/restyle-decode-error.html started
10:19:32.953 5111 worker/2 accessibility/animation-policy.html passed
10:19:32.955 5111 worker/2
accessibility/anonymous-render-block-in-continuation-causes-crash.html started
10:19:32.999 5109 renderer crash, pid = None, error_line = #CRASHED - renderer
10:19:33.000 5109 killed pid 5357
10:19:33.000 5109 worker/0 http/tests/images/restyle-decode-error.html crashed,
(stderr lines):
10:19:33.000 5109   Received signal 11 <unknown> 000000000000
10:19:33.000 5109    [0x0001196ef23e]
...
10:19:33.002 5109   [end of stack trace]
10:19:33.002 5109  
[5358:13319:0722/101932:598964195783:ERROR:node_controller.cc(1099)] Could not
be introduced to peer C9E4F59A639F0E59.C2F8E762ADA8C89E
10:19:33.006 5084 [4832/40217] http/tests/images/restyle-decode-error.html
failed unexpectedly (renderer crashed)
10:19:33.003 5109 worker/0 killing primary driver
10:19:33.004 5109 worker/0 killing secondary driver
10:19:33.004 5109 worker/0 http/tests/images/restyle-decode-error.html failed:
10:19:33.004 5109 worker/0  renderer crashed
---

Also same failure on other bots such as
https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Leak/b...

(ignore other webkit failures on this bot; those all have a known cause with a
fix incoming)


Original issue's description:
> Cancel image loads if decoding failed.
> 
> 'Cancel' isn't really the right word for what this change does.
> In the event of an image decoding failure, ImageResource will
> synthesize a call to ResourceLoader::didFinishLoading, killing
> the actual network request but reporting it as a successful
> completion to the rest of blink. This matches our traditional
> behavior of decoding errors looking like a successful
> resource load (for the most part).
> 
> This requires some plumbing changes to image decoding, because
> the decoder selection logic doesn't report why it wasn't able to
> create an ImageDecoder. It might be because insufficient data has
> been received to sniff the image type, or it may be that we
> definitely don't have a valid image type. This change exposes enough
> information to tell the difference.
> 
> BUG=471272
> 
> Committed: https://crrev.com/520519605b961a7d979bc0eaa3c057a973dc26f1
> Cr-Commit-Position: refs/heads/master@{#407171}

TBR=pdr@chromium.org,peter@chromium.org,japhet@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=471272
==========

[commit-bot: I haz the power, 2016-07-22 19:18:54]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-07-22 19:20:30]

Description was changed from

==========
Revert of Cancel image loads if decoding failed. (patchset #8 id:140001 of
https://codereview.chromium.org/2108033003/ )

Reason for revert:

Likely cause of failures:

webkit_tests webkit_tests:
http/tests/images/restyle-decode-error.html

https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.11%20%28...

---
10:19:32.937 5109 worker/0 http/tests/images/restyle-decode-error.html started
10:19:32.953 5111 worker/2 accessibility/animation-policy.html passed
10:19:32.955 5111 worker/2
accessibility/anonymous-render-block-in-continuation-causes-crash.html started
10:19:32.999 5109 renderer crash, pid = None, error_line = #CRASHED - renderer
10:19:33.000 5109 killed pid 5357
10:19:33.000 5109 worker/0 http/tests/images/restyle-decode-error.html crashed,
(stderr lines):
10:19:33.000 5109   Received signal 11 <unknown> 000000000000
10:19:33.000 5109    [0x0001196ef23e]
...
10:19:33.002 5109   [end of stack trace]
10:19:33.002 5109  
[5358:13319:0722/101932:598964195783:ERROR:node_controller.cc(1099)] Could not
be introduced to peer C9E4F59A639F0E59.C2F8E762ADA8C89E
10:19:33.006 5084 [4832/40217] http/tests/images/restyle-decode-error.html
failed unexpectedly (renderer crashed)
10:19:33.003 5109 worker/0 killing primary driver
10:19:33.004 5109 worker/0 killing secondary driver
10:19:33.004 5109 worker/0 http/tests/images/restyle-decode-error.html failed:
10:19:33.004 5109 worker/0  renderer crashed
---

Also same failure on other bots such as
https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Leak/b...

(ignore other webkit failures on this bot; those all have a known cause with a
fix incoming)


Original issue's description:
> Cancel image loads if decoding failed.
> 
> 'Cancel' isn't really the right word for what this change does.
> In the event of an image decoding failure, ImageResource will
> synthesize a call to ResourceLoader::didFinishLoading, killing
> the actual network request but reporting it as a successful
> completion to the rest of blink. This matches our traditional
> behavior of decoding errors looking like a successful
> resource load (for the most part).
> 
> This requires some plumbing changes to image decoding, because
> the decoder selection logic doesn't report why it wasn't able to
> create an ImageDecoder. It might be because insufficient data has
> been received to sniff the image type, or it may be that we
> definitely don't have a valid image type. This change exposes enough
> information to tell the difference.
> 
> BUG=471272
> 
> Committed: https://crrev.com/520519605b961a7d979bc0eaa3c057a973dc26f1
> Cr-Commit-Position: refs/heads/master@{#407171}

TBR=pdr@chromium.org,peter@chromium.org,japhet@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=471272
==========

to

==========
Revert of Cancel image loads if decoding failed. (patchset #8 id:140001 of
https://codereview.chromium.org/2108033003/ )

Reason for revert:

Likely cause of failures:

webkit_tests webkit_tests:
http/tests/images/restyle-decode-error.html

https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.11%20%28...

---
10:19:32.937 5109 worker/0 http/tests/images/restyle-decode-error.html started
10:19:32.953 5111 worker/2 accessibility/animation-policy.html passed
10:19:32.955 5111 worker/2
accessibility/anonymous-render-block-in-continuation-causes-crash.html started
10:19:32.999 5109 renderer crash, pid = None, error_line = #CRASHED - renderer
10:19:33.000 5109 killed pid 5357
10:19:33.000 5109 worker/0 http/tests/images/restyle-decode-error.html crashed,
(stderr lines):
10:19:33.000 5109   Received signal 11 <unknown> 000000000000
10:19:33.000 5109    [0x0001196ef23e]
...
10:19:33.002 5109   [end of stack trace]
10:19:33.002 5109  
[5358:13319:0722/101932:598964195783:ERROR:node_controller.cc(1099)] Could not
be introduced to peer C9E4F59A639F0E59.C2F8E762ADA8C89E
10:19:33.006 5084 [4832/40217] http/tests/images/restyle-decode-error.html
failed unexpectedly (renderer crashed)
10:19:33.003 5109 worker/0 killing primary driver
10:19:33.004 5109 worker/0 killing secondary driver
10:19:33.004 5109 worker/0 http/tests/images/restyle-decode-error.html failed:
10:19:33.004 5109 worker/0  renderer crashed
---

Also same failure on other bots such as
https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Leak/b...

(ignore other webkit failures on this bot; those all have a known cause with a
fix incoming)

Original issue's description:
> Cancel image loads if decoding failed.
>
> 'Cancel' isn't really the right word for what this change does.
> In the event of an image decoding failure, ImageResource will
> synthesize a call to ResourceLoader::didFinishLoading, killing
> the actual network request but reporting it as a successful
> completion to the rest of blink. This matches our traditional
> behavior of decoding errors looking like a successful
> resource load (for the most part).
>
> This requires some plumbing changes to image decoding, because
> the decoder selection logic doesn't report why it wasn't able to
> create an ImageDecoder. It might be because insufficient data has
> been received to sniff the image type, or it may be that we
> definitely don't have a valid image type. This change exposes enough
> information to tell the difference.
>
> BUG=471272
>
> Committed: https://crrev.com/520519605b961a7d979bc0eaa3c057a973dc26f1
> Cr-Commit-Position: refs/heads/master@{#407171}

TBR=pdr@chromium.org,peter@chromium.org,japhet@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=471272

Committed: https://crrev.com/bc36d89c757eedc36f73e1caa9948d384a2b8e60
Cr-Commit-Position: refs/heads/master@{#407228}
==========

[commit-bot: I haz the power, 2016-07-22 19:20:31]

Patchset 1 (id:??) landed as
https://crrev.com/bc36d89c757eedc36f73e1caa9948d384a2b8e60
Cr-Commit-Position: refs/heads/master@{#407228}

[pdr., 2016-07-22 19:21:51]

On 2016/07/22 at 19:18:54, commit-bot wrote:
> Committed patchset #1 (id:1)

I wonder how the original patch passed the CQ which should have run this test
with dchecks on?

In any case, this rollout is correct. Thank you!