Explicitly initialize Bigint members after uninitialized allocations.

Explicitly initialize Bigint members after uninitialized allocations.

Committed: https://code.google.com/p/dart/source/detail?r=6988

[siva, 2012-04-26 04:45:04]

LGTM

http://codereview.chromium.org/10230001/diff/2001/runtime/vm/snapshot.cc
File runtime/vm/snapshot.cc (right):

http://codereview.chromium.org/10230001/diff/2001/runtime/vm/snapshot.cc#newc...
runtime/vm/snapshot.cc:255: obj->ptr()->signed_length_ = bigint_length;
why not add a set_allocated_length() and set_signed_length() in object.h and use
them here

or

Bigint& result = Bigint::Handle();
{
  NoGCScope nogc;
  result = AllocateUninitialized(...);
  result.raw()->ptr()->allocated_length_ = ...
  result.raw()->ptr()->signed_length_ = ...
 }
 BigintOperations::FromHexCString(...);
 return

[cshapiro, 2012-04-26 06:34:53]

http://codereview.chromium.org/10230001/diff/2001/runtime/vm/snapshot.cc
File runtime/vm/snapshot.cc (right):

http://codereview.chromium.org/10230001/diff/2001/runtime/vm/snapshot.cc#newc...
runtime/vm/snapshot.cc:255: obj->ptr()->signed_length_ = bigint_length;
Hmm... I tried to make the code look like the rest of the code in the snapshot
reader.

The NoGCScope seemed unnecessary because of the assert on line 249 and because
the entire read is wrapped in a NoGCScope on line 130.

We have a SetLength but it does something different than one would expect.  It
is confusing to me why we have these two length values in the first place.  
Let's talk about this off-line.

[siva, 2012-04-26 17:17:44]

http://codereview.chromium.org/10230001/diff/2001/runtime/vm/snapshot.cc
File runtime/vm/snapshot.cc (right):

http://codereview.chromium.org/10230001/diff/2001/runtime/vm/snapshot.cc#newc...
runtime/vm/snapshot.cc:255: obj->ptr()->signed_length_ = bigint_length;
Agree, I didn't realize we already had a NoGCScope around the caller.

Maybe in a different CL, I will add an assertion in each of this functions so
that there is no confusion.

On 2012/04/26 06:34:53, cshapiro wrote:
> Hmm... I tried to make the code look like the rest of the code in the snapshot
> reader.
> 
> The NoGCScope seemed unnecessary because of the assert on line 249 and because
> the entire read is wrapped in a NoGCScope on line 130.
> 
> We have a SetLength but it does something different than one would expect.  It
> is confusing to me why we have these two length values in the first place.  
> Let's talk about this off-line.