Re: Re: SourceComment: Let the library do the work..

Hi Atul!

I pushed the second patch in revision 84 of your public calitko-Gnutella branch.

Regarding you Packet::Data command, we don’t necessary need to promote it to a full-pledged class if we want to automate some behavior. Take a look a QSharedData and QSharedDataPointers and their “polymorphic” variants in Utils/Memory/SharedData.*. They take care of reference counting of the private data behind the scenes and we still use Data as a POD ;-). As Anders correctly suggested before, we should prefer to use QSharedData and QSharedDataPointer instead of implementing ref counting ourselves each time.

Regards,

Peter

atul wrote:Dear all,

I was able to do it…
Here is the latest patch… I think this idea is applicable elsewhere

( other places ref counting is done… )

– cheerio atul

atul wrote:

Dear all

While browsing thru the code I found these two small refactorings…. Thanks to the tests as they told me where I went wrong ;-)
P.S. I, very much wanted to use qAtomicDetach in
Packet::invalidateHeader(). Elevating Packet::Data to a class status would let us use it… but, opinions vary - CHUCKLE ;-)

— cheerio atul

Would you like to post a relpy?


This post is a reply to:
Re: SourceComment: Let the library do the work..
Dear all, I was able to do it... Here is the latest patch... I think this idea is applicable elsewhere ( other places ref counting is done... ) -- cheerio atul atul wrote: Dear all While browsing (more...)

No follow-ups yet.