U++ forum: Welcome to the forum

Search on this site

Search in forums

Home » U++ Library support » U++ Core » PROPOSAL: SerializeStore helper

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

PROPOSAL: SerializeStore helper [message #29413]

Tue, 19 October 2010 12:05

kohait00
Messages: 939
Registered: July 2009
Location: Germany

Experienced Contributor

just a short add..

XmlizeStore exists, to be able to handle 'const T& x'
I couldnt find anything similar for Serialize..
here is a solution.

template<class T>
void SerializeStore(Stream& s, const T& x)
{
	ASSERT(s.IsStoring());
	s % const_cast<T&>(x);
}

this makes thing possible like

SerializeStore(s, mymap.GetKey(i));

without the hassle of a const_cast each time..and an ASSERT is there..

any better solution or didnt i think of every pitfall..

Report message to a moderator

Re: PROPOSAL: SerializeStore helper [message #29422 is a reply to message #29413]

Tue, 19 October 2010 15:28

mirek
Messages: 13975
Registered: November 2005

Ultimate Member

kohait00 wrote on Tue, 19 October 2010 06:05

just a short add..

XmlizeStore exists, to be able to handle 'const T& x'
I couldnt find anything similar for Serialize..
here is a solution.

template<class T>
void SerializeStore(Stream& s, const T& x)
{
	ASSERT(s.IsStoring());
	s % const_cast<T&>(x);
}

this makes thing possible like

SerializeStore(s, mymap.GetKey(i));

without the hassle of a const_cast each time..and an ASSERT is there..

any better solution or didnt i think of every pitfall..

Well, the equivalent function is there, called "Store", but it now accepts only non-const parameter - so the change would be there...

The problem I possibly see with it (and in fact, maybe the XmlizeStore is wrong) is that you are giving option to client code to change const object... I mean, nothing prevents Serialize to mutate the object.

But perhaps I am wrong... Another opinion?

Report message to a moderator

Re: PROPOSAL: SerializeStore helper [message #29426 is a reply to message #29422]

Tue, 19 October 2010 15:47

kohait00
Messages: 939
Registered: July 2009
Location: Germany

Experienced Contributor

i think at a cetain point one needs to handle over control to tested and working code. there, trust begins Smile

otherwise, i wouldn't see the point of XmlizeStore as well.

also, to leave the user using const_cast at points he maybe doesn't really understand, 'just to make this damn thing compile', is probably more error prone.. it's to hide all that 'hacking' away from user.

just an opinion. you are to decide..

Report message to a moderator

Previous Topic:	no String::Replace() ?
Next Topic:	PolyDeepCopyNew: MSC / GCC differ in behaviour

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Fri Apr 19 04:56:23 CEST 2024

Total time taken to generate the page: 0.02887 seconds