Re: [boost] Review Wizard report for December 2009

Subject: Re: [boost] Review Wizard report for December 2009
From: Phil Bouchard (philippe_at_[hidden])
Date: 2009-12-10 16:59:36

Thorsten Ottosen wrote:
> Phil Bouchard skrev:
>> Thank you for this update.  A great introduction to the library can be 
>> found in the following PowerPoint presentation:
>> https://svn.boost.org/svn/boost/sandbox/shifted_ptr/libs/smart_ptr/doc/ShiftedPointer.ppt 
>>
> 
> Thanks.

BTW I haven't corrected the presentation yet but there the disadvantages 
of a smart pointer list slowness.  It was something basically copied 
from a garbage collection list of disadvantages; I forgot removing the 
bullet.

[...]

>> 3) Instantaneous
> 
> What does 3 mean?

Constant complexity.

> Another question: is it possible to say
> 
> shifted_ptr<T> ptr = make_shifted<T>(...)
> 
> instead (to enhance exception-safety)?

Yes just by adding some wrapper but I think the reason I did it this way 
was because 'operator new' can take advantage of a user defined allocator.

>> Furthermore I am looking forward testing shifted_ptr on Mozilla 
>> Firefox...
> 
> Please do tells us what you observe. Is Firefox using shared_ptr 
> already, or?

No but I see memory leaks and a huge consumption of memory ;)

Regards,
-Phil

Note:  This is a good timing because I just completed a paper in 
astrophysics including a simulator to approximate all tests.  It is 
called "Finite Relativism"; for those interested in the subject please ask.