Introduce ZLibInflater and ZLibDeflater to dart:io.

Introduce ZLibInflater and ZLibDeflater to dart:io.

BUG=

Committed: https://code.google.com/p/dart/source/detail?r=19551

[Søren Gjesse, 2013-03-05 15:59:05]

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.cc
File runtime/bin/filter.cc (right):

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.cc#newcode23
runtime/bin/filter.cc:23: || filter == NULL) {
Shouldn't we call Dart_PropagateError when we Dart_IsError is true? The goes for
a number of places below as well. We loose the actual error and just get an
InternalError.

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.cc#newcode38
runtime/bin/filter.cc:38: if (!filter->Init() ||
NULL check as well.

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.cc#newcode63
runtime/bin/filter.cc:63: if (!filter->Init() ||
NULL check here as well.

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.cc#newcode83
runtime/bin/filter.cc:83: if (Dart_IsError(Dart_ListGetAsBytes(data_obj, 0,
data, length))) {
Before using ListGetAsBytes you should try Dart_TypedDataAcquireData (see
Socket_WriteList in socket.cc).

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.h
File runtime/bin/filter.h (right):

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.h#newcode23
runtime/bin/filter.h:23: * On a succesfull call to Process, Process will take
ownership of data.
Please expand comment on how the data will be freed.

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.h#newcode44
runtime/bin/filter.h:44: virtual bool Process(uint8_t* data, intptr_t length);
Does the same ownership handover happen here?

[Anders Johnsen, 2013-03-06 10:32:46]

Thanks, improved quite a few areas. PTAL.

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.cc
File runtime/bin/filter.cc (right):

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.cc#newcode23
runtime/bin/filter.cc:23: || filter == NULL) {
On 2013/03/05 15:59:05, Søren Gjesse wrote:
> Shouldn't we call Dart_PropagateError when we Dart_IsError is true? The goes
for
> a number of places below as well. We loose the actual error and just get an
> InternalError.

Done.

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.cc#newcode38
runtime/bin/filter.cc:38: if (!filter->Init() ||
On 2013/03/05 15:59:05, Søren Gjesse wrote:
> NULL check as well.

Done.

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.cc#newcode63
runtime/bin/filter.cc:63: if (!filter->Init() ||
On 2013/03/05 15:59:05, Søren Gjesse wrote:
> NULL check here as well.

Done.

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.cc#newcode83
runtime/bin/filter.cc:83: if (Dart_IsError(Dart_ListGetAsBytes(data_obj, 0,
data, length))) {
On 2013/03/05 15:59:05, Søren Gjesse wrote:
> Before using ListGetAsBytes you should try Dart_TypedDataAcquireData (see
> Socket_WriteList in socket.cc).

Done.

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.h
File runtime/bin/filter.h (right):

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.h#newcode23
runtime/bin/filter.h:23: * On a succesfull call to Process, Process will take
ownership of data.
On 2013/03/05 15:59:05, Søren Gjesse wrote:
> Please expand comment on how the data will be freed.

Done.

https://codereview.chromium.org/12451004/diff/1/runtime/bin/filter.h#newcode44
runtime/bin/filter.h:44: virtual bool Process(uint8_t* data, intptr_t length);
On 2013/03/05 15:59:05, Søren Gjesse wrote:
> Does the same ownership handover happen here?

It does, but since it overrides the method, I think it's fine to just state it
in the super class?

[Søren Gjesse, 2013-03-06 11:09:52]

lgtm

https://codereview.chromium.org/12451004/diff/6001/runtime/bin/filter.cc
File runtime/bin/filter.cc (right):

https://codereview.chromium.org/12451004/diff/6001/runtime/bin/filter.cc#newc...
runtime/bin/filter.cc:28: "Failed to get destroyed filter"));
You actually got a destroyed filter :-). How about just "Filter destroyed"?

[Anders Johnsen, 2013-03-06 11:37:24]

Landing.

https://codereview.chromium.org/12451004/diff/6001/runtime/bin/filter.cc
File runtime/bin/filter.cc (right):

https://codereview.chromium.org/12451004/diff/6001/runtime/bin/filter.cc#newc...
runtime/bin/filter.cc:28: "Failed to get destroyed filter"));
On 2013/03/06 11:09:53, Søren Gjesse wrote:
> You actually got a destroyed filter :-). How about just "Filter destroyed"?

You are right, changed :)

[Anders Johnsen, 2013-03-06 11:38:55]

Committed patchset #3 manually as r19551 (presubmit successful).