ATL / WTL / STL

CodeGuy3-Sep-02 4:23

CodeGuy

3-Sep-02 4:23

Chris Hambleton wrote:
Is there a built-in function that would allow me to do this without having to create a new map and do this through a loop?

Nope. A map is normally a red-black tree implementation, so by swapping a key-value pair, you've essentially invalidated the location of that node in the tree.

You can always remove the key-value pair first and then add the swapped values back in the tree. (More overhead, less memory usage of course.)

Brandon

Joao Vaz3-Sep-02 4:56

Joao Vaz

3-Sep-02 4:56

Humm, perhaps you could use a multimap, since insertion and remotion from the multimap doesn't invalidate the iterator(if I'm not mistaken here).

Cheers,
Joao Vaz
A Programming Language is a tool that has profound influence on our thinking habits -The late giant Edsger Dijkstra 1930 - 2002

And if your dream is to care for your family, to put food on the table, to provide them with an education and a good home, then maybe suffering through an endless, pointless, boring job will seem to have purpose. And you will realize how even a rock can change the world, simply by remaining obstinately stationary. -

Alexandru Savescu3-Sep-02 22:23

Alexandru Savescu

3-Sep-02 22:23

Well, until Joaquin bidirectional map comes out I suggest you use two maps and use one of them depending on what the keys are. This is what I have been doing. Smile | :)

Best regards,
Alexandru Savescu

Todd Smith5-Sep-02 13:18

Todd Smith

5-Sep-02 13:18

Why not store as the value both they value and the key

struct
{
int key;
int value;
} KeyValuePair;

std::map<int, KeyValuePair> mymap;

Then you can search through the map (although not as fast) by eithe the key or value.

Todd Smith

Jörgen Sigvardsson3-Oct-02 14:43

Jörgen Sigvardsson

3-Oct-02 14:43

You can already do a linear search for values. Why add the key to the data set?

--
Please state the nature of your medical emergency.

Jörgen Sigvardsson3-Oct-02 14:37

Jörgen Sigvardsson

3-Oct-02 14:37

You can always to this:

template <typename key_t, typename value_t>
class bidir_map {
	typedef map<key_t, value_t>	key_to_value_map;
	typedef map<value_t, key_t>	value_to_key_map;

	key_to_value_map	key_map;
	value_to_key_map	value_map;
 
	/* Implement map-interface here which manipulate
  	   key_map and value_map accordingly during inserts
	   and etc. 
	 */
};

Although, this may not be the most efficient implementation in terms of size. But if your collections is small, then this may be a cheap solution.

--
Please state the nature of your medical emergency.

IDL question

Jeremy Pullicino2-Sep-02 23:00

Jeremy Pullicino

2-Sep-02 23:00

Hello guys,

I have an ATL project with a COM object called IJeremy it has a method called SetMsg which takes an IMessage* as an [in] param.

This object fails to compile - it tells me:
error MIDL2025 : syntax error : expecting a type specification near "IMessage"

What is the best way around to get it to work?

I have tried importing CDOSys.Idl from my IDL, however it does not compile because of many errors.

Has anyone done this before?

Thanks for the help,
Jeremy

Jeremy Pullicino
Professional C++ Developer

Done any hacking lately?

Joaquín M López Muñoz3-Sep-02 2:49

Joaquín M López Muñoz

3-Sep-02 2:49

Have you tried inserting a forward declaration for IMessage?

Joaquín M López Muñoz
Telefónica, Investigación y Desarrollo

Jeremy Pullicino3-Sep-02 3:44

Jeremy Pullicino

3-Sep-02 3:44

Yes, I did that, however, I get another error when I do that.

error MIDL2011 : unresolved type declaration : IMessage [ Parameter 'pIMessage' of Procedure 'Submit' ( Interface 'ISubmitMsg2' ) ]

'Submit' is my method
'ISubmitMsg2' is my interface

Thanks,
Jeremy.

Jeremy Pullicino
Professional C++ Developer

Done any hacking lately?

Joaquín M López Muñoz3-Sep-02 4:04

Joaquín M López Muñoz

3-Sep-02 4:04

Try moving the forward declaration of IMessage into the library block where ISubmitMsg2 is defined.

Joaquín M López Muñoz
Telefónica, Investigación y Desarrollo

Bill S3-Sep-02 4:24

Bill S

3-Sep-02 4:24

Jeremy Pullicino wrote:
I have an ATL project with a COM object called IJeremy it has a method called SetMsg which takes an IMessage* as an [in] param.

Why not pass the IUnknown* I/F and requery back for IMessage* I/F?

Jeremy Pullicino3-Sep-02 5:13

Jeremy Pullicino

3-Sep-02 5:13

Bill S wrote:
Why not pass the IUnknown* I/F and requery back for IMessage* I/F?

I am currently doing that Smile | :)

I would like to manage to get an IMessage* through however...

Jeremy.

Jeremy Pullicino
Professional C++ Developer

Done any hacking lately?

Bill S3-Sep-02 5:21

Bill S

3-Sep-02 5:21

Jeremy Pullicino wrote:
I would like to manage to get an IMessage* through however...

I assume that IMessage I/F is your own custom interface. You will have to import the interface definition file that provides the marshalling definition for this interface. This is created when compiling the IDL file that contains the IMessage I/F definition. Give it a try.

Best Regards,

ATL based COM fire-event client ?

Tom Wright5-Sep-02 7:36

Tom Wright

5-Sep-02 7:36

Jeremy,
Try making your "in" look like this
[in]BSTR param

make sure you change the prototype in header file to reflect this change and update the function on the CPP file also.

let me know if this helps

in your code after you get the message that the container is returning to you, you will need to do something like this:

STDMETHODIMP CJeremy::SetMsg(BSTR param)
{
USES_CONVERSION;
LPSTR lpParam = W2A(param);
::MessageBox(NULL, lpParam, "Title", MB_OK);
return S_OK;
}

or very close to that in order to read the .

Tom Wright
tawright915@yahoo.com

alma2-Sep-02 3:25

alma

2-Sep-02 3:25

How can I catch a fired COM event in an ATL based NT/win2k service?
I have found only examples of clients written in VB or MFC.

Alma.

Re: ATL based COM fire-event client ?

Pavel Klocek2-Sep-02 6:20

Pavel Klocek

2-Sep-02 6:20

Try this: http://www.widgetware.com/FAQATL_Misc.htm#Events

Pavel
Sonork 100.15206

Re: ATL based COM fire-event client ?

Todd Smith2-Sep-02 15:43

Todd Smith

2-Sep-02 15:43

http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vcsample/html/vcsmpatlfire.asp[^]

Todd Smith

A question about #define WIN32 and "sbrk"

George Ma1-Sep-02 1:13

1-Sep-02 1:13

Hi, everyone!

Look at the source codes,

--------
/* #define WIN32 */
#ifdef WIN32
#define MORECORE wsbrk
#endif
--------

Such is the comments,

--------
WIN32 (default: undefined)
Define this on MS win (95, nt) platforms to compile in sbrk emulation.
--------

I have two questions,

1. What means WIN32? When the variable "WIN32" becomes
defined? When it is not defined?

2. What means "sbrk" in the comment? What means "wsbrk"?

Btw: the software is written for both Windows and Linux
platforms.

Cheers,
George

Stephane Rodriguez.1-Sep-02 2:15

Stephane Rodriguez.

1-Sep-02 2:15

You need not crosspost, unless you intend to be flamed in no time Unsure | :~

And I swallow a small raisin.

George Ma1-Sep-02 4:08

1-Sep-02 4:08

Thanks, StephaneRodriguez pal!

What means "flamed"? Smile | :)

My English is not
very nice. Frown | :-(

But I will accept your advice.

However my boss is asked me to hand in the
project immediately, but I am in trouble as you
see.

Cheers,
George

Stephane Rodriguez.1-Sep-02 4:21

Stephane Rodriguez.

1-Sep-02 4:21

A couple of people have already argued against you about the crosspost. See for yourself : the lounge.

And I swallow a small raisin.

George Ma2-Sep-02 3:34

2-Sep-02 3:34

Oh, I am sorry about that, StephaneRodriguez.

Yesterday, it is a very urgent case, my boss, ...

I promise that case will not happen again. Smile | :)

Cheers,
George

Anders Molin1-Sep-02 5:01

Anders Molin

1-Sep-02 5:01

George Ma wrote:
What means "flamed"?

Basically it means that people starts writing not-so-nice things to you, because they get angry with you...

- Anders

Money talks, but all mine ever says is "Goodbye!"

George Ma2-Sep-02 3:34