- Use conditional code inclusion for the dart_no_snapshot and dart executables using the MACRO DART…

- Use conditional code inclusion for the dart_no_snapshot and dart executables using the MACRO DART_NO_SNAPSHOT
- Do some minor modifications to the intrinsifier code to not do redundant fingerprint checks and avoid code duplication.

These changes result in a text segment size savings of about 110k.

Old Sizes:
   text    data     bss     dec     hex filename
3801418    7696   19261 3828375  3a6a97 (TOTALS)

New sizes:
   text    data     bss     dec     hex filename
3690342    7696   19261 3717299  38b8b3 (TOTALS)

Committed: https://code.google.com/p/dart/source/detail?r=45704

[siva, 2015-05-08 23:09:22]

asiva@google.com changed reviewers:
+ iposva@google.com

[Ivan Posva, 2015-05-08 23:52:00]

LGTMwC

-Ivan

https://codereview.chromium.org/1134623004/diff/1/runtime/bin/bin.gypi
File runtime/bin/bin.gypi (right):

https://codereview.chromium.org/1134623004/diff/1/runtime/bin/bin.gypi#newcod...
runtime/bin/bin.gypi:352: 'target_name': 'libdart_withcore',
Should we name this also _nosnapshot?

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/dart.cc
File runtime/vm/dart.cc (right):

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/dart.cc#newcode234
runtime/vm/dart.cc:234: // of Object::Init(..) in a regular isolate creation
path.
Please update the comment.

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/intrinsifier.cc
File runtime/vm/intrinsifier.cc (right):

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/intrinsifier.cc#...
runtime/vm/intrinsifier.cc:194: compiler->assembler()->Comment("Intrinsic");
Not exactly: If we should throw on JS overflows and this is a integer library
intrinsic then we could have generated the comment without any actual
instructions.

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/method_recognize...
File runtime/vm/method_recognizer.cc (right):

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/method_recognize...
runtime/vm/method_recognizer.cc:32: #undef KIND_TO_STRING
Can you move the #define and the #undef outside of the structure being defined?
Probably makes it easier to read.

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/method_recognizer.h
File runtime/vm/method_recognizer.h (right):

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/method_recognize...
runtime/vm/method_recognizer.h:467: #define CHECK_FINGERPRINT3(f, p0, p1, p2,
fp) \
Aren't these also only needed if DART_NO_SNAPSHOT is set?

[siva, 2015-05-11 18:22:14]

https://codereview.chromium.org/1134623004/diff/1/runtime/bin/bin.gypi
File runtime/bin/bin.gypi (right):

https://codereview.chromium.org/1134623004/diff/1/runtime/bin/bin.gypi#newcod...
runtime/bin/bin.gypi:352: 'target_name': 'libdart_withcore',
On 2015/05/08 23:52:00, Ivan Posva wrote:
> Should we name this also _nosnapshot?

Done.

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/dart.cc
File runtime/vm/dart.cc (right):

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/dart.cc#newcode234
runtime/vm/dart.cc:234: // of Object::Init(..) in a regular isolate creation
path.
On 2015/05/08 23:52:00, Ivan Posva wrote:
> Please update the comment.

Refactor the code a bit to do the Object::Init(..) stuff unconditionally and put
the snapshot reading code alone under the condition.

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/intrinsifier.cc
File runtime/vm/intrinsifier.cc (right):

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/intrinsifier.cc#...
runtime/vm/intrinsifier.cc:194: compiler->assembler()->Comment("Intrinsic");
On 2015/05/08 23:52:00, Ivan Posva wrote:
> Not exactly: If we should throw on JS overflows and this is a integer library
> intrinsic then we could have generated the comment without any actual
> instructions.

Good point, pulled it back inside the macro.

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/method_recognize...
File runtime/vm/method_recognizer.cc (right):

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/method_recognize...
runtime/vm/method_recognizer.cc:32: #undef KIND_TO_STRING
On 2015/05/08 23:52:00, Ivan Posva wrote:
> Can you move the #define and the #undef outside of the structure being
defined?
> Probably makes it easier to read.

Done.

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/method_recognizer.h
File runtime/vm/method_recognizer.h (right):

https://codereview.chromium.org/1134623004/diff/1/runtime/vm/method_recognize...
runtime/vm/method_recognizer.h:467: #define CHECK_FINGERPRINT3(f, p0, p1, p2,
fp) \
On 2015/05/08 23:52:00, Ivan Posva wrote:
> Aren't these also only needed if DART_NO_SNAPSHOT is set?

they are used in flow_graph_builder.h and Library::CheckFunctionFingerprints()
needs some more cleanup, next CL.

[siva, 2015-05-11 18:32:09]

Committed patchset #2 (id:20001) manually as r45704 (presubmit successful).