|
Hi guys,
I have COM DLL Servers creating a lot of COM objects with CoCreateInstance and it's quite slow.
I wondered if replacing the CoCreateInstance(CLSID_Foo) by a new CComObject < CFoo > would speed up things ?
Or to be clearer, how slower is CoCreateInstance compared to new operator ?
thanks !
|
|
|
|
|
Alexandre GRANVAUD wrote: I wondered if replacing the CoCreateInstance(CLSID_Foo) by a new CComObject < CFoo > would speed up things ?
No.
Since in both ways a COM object is created. Using the "new operator", as you call it, will eventually call ::CoCreateInstance(), so there's no difference.
"It's supposed to be hard, otherwise anybody could do it!" - selfquote
"High speed never compensates for wrong direction!" - unknown
|
|
|
|
|
calling new CComObject < CThing > calls CoCoreateInstance????
i'm sure it doesnt !!
|
|
|
|
|
Alexandre GRANVAUD wrote: calling new CComObject < CThing > calls CoCoreateInstance????
i'm sure it doesnt !!
No, of course; allocating memory for a CComObject doesn't call ::CoCreateInstance().
But in order to create the server you'd have to call CComObject::CreateInstance().
Otherwise the server won't be created and your comparison wouldn't make sense at all. 
"It's supposed to be hard, otherwise anybody could do it!" - selfquote
"High speed never compensates for wrong direction!" - unknown
|
|
|
|
|
why do you need to call it ?
as i can see, the CCOmObject constructor call FinalConstruct, so what extra initialization do we need ?
|
|
|
|
|
You simply cannot use new to create COM objects (COM objects lifetime has its own rules  ). If you need to create multiple objects, then you may obtain a speed enhancement using directly the IClassFactory interface, see the remarks section of this page [^].
If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong.
-- Iain Clarke
[My articles]
|
|
|
|
|
yes you can in the same DLL, i do it everyday : in inprocess it's ok to do so 
new CComObject < CThing > then doing a Queryinterface or a AddRef on it and your object is ok
|
|
|
|
|
And what is the rationale behind that (other than revealing your bad practices )?
If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong.
-- Iain Clarke
[My articles]
|
|
|
|
|
speed is always a good practice
having a bad practice is ok if you know you have (i do ) and know the consequences
|
|
|
|
|
|
lol i agree for the speed (but can't slowdown)
i don't break the com creation mechanism : the CreateInstance in a ClassFactory does exactly the same : a new CComObject<cthing> !
|
|
|
|
|
Alexandre GRANVAUD wrote: i don't break the com creation mechanism : the CreateInstance in a ClassFactory does exactly the same : a new CComObject !
Avoiding such call you're breaking it by definition. Of course you'll see the practical implications of such a wicked act just on 'strange' servers (i.e. servers that don't "do exactly the same", such a behaviour is allowed by COM).
If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong.
-- Iain Clarke
[My articles]
|
|
|
|
|
hi,
Please, I developed an application with C + + and I need help; I am trying to test in my application that usually requires a COM + dll and I wonder how I could test that this dll is registred or no and it should display a message if this dll is not in the register or it does not exist?
thanks
|
|
|
|
|
What about calling CoGetClassObject [^]?
If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong.
-- Iain Clarke
[My articles]
|
|
|
|
|
i just create an instance of a class in this dll and after i do a test if it is registred or no .this are some lines of code which can explain what i need exactly:
CoInitializeEx(NULL, COINIT_APARTMENTTHREADED);
hRes = TContext::CreateInstanceNoAppID(CLSID_TLibraryInformation_SeeEdbUI, &piLibInfoSeeEdbUI);
if(FAILED(hRes))
{
if(hRes == REGDB_E_CLASSNOTREG)
{
strTemp.LoadString(IDS_SEEEDBUIDLL_NOT_FOUND);
TErrorTool::DisplayErrorStack(TRUE, _bstr_t(strTemp));
}
else
{
strTemp.LoadString(IDS_SEEEDBUIDLL_NOT_FOUND);
TErrorTool::DisplayErrorStack(TRUE, _bstr_t(strTemp));
}
exit(0);
}
|
|
|
|
|
You don't you trust CoGetClassObject, do you? Why?
If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong.
-- Iain Clarke
[My articles]
|
|
|
|
|
frankly, i don't try it and i would save the same structure of my code. so if you can help me with this?
|
|
|
|
|
I don't know Borland's (or CodeGear or whatever) classes so I cannot help with those.
I would keep the whole thing simple using directly COM library API, anyway, if your code is able to detect DLL's registration, that's fine, isn't it?
If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong.
-- Iain Clarke
[My articles]
|
|
|
|
|
|
Using Matt Curland's Type Library Editor, I wrote a Typelib for Nick Yakowlew's 'Charset Detector Library' . This is a standard DLL (written in Delphi as it happens) and not a COM DLL.
Up until this point I had been building ODL and thus TLB files "by hand". The benefit of this was that I could get Intellisense working in Visual Studio 6 (esp. VB6) and also be able to pass widestrings to DLLs without said strings being mangled into ANSI in the process. Once I had generated the TLB from the ODL using MIDL, I had but to register the TLB using `regtlib`, add a reference to that typelib in VB6 and cut code. And once the binary had been built, I didn't need to have the typelib around.
The typelib made by Curland's tool seems to work a little differently. It forces me to have the typelib registered for the binary to access the DLL. I exported the TLB to IDL and, well it looks a bit different to the ODL I've been writing. For one thing, there are typedefs and structs and enums. Okay, I put those in there via Curland's TLE and they need to remain. But I don't want to have to register the TLB -- I want to do things the old way.
So what do I do to the IDL below, to make it the kind of ODL I'm familiar with, yet maintain the functionality, all without needing the TLB to be registered once the binary has been produced?
<br />
[<br />
uuid(316A83D7-8BF4-490E-BDDE-75EBC332C355),<br />
version(1.0),<br />
helpstring("Charset Detector - as the name says - is a stand alone executable module for automatic charset detection of a given text.\r\n\t\r\nIt can be useful for internationalisation support in multilingual applications such as web-script editors or Unicode editors.\r\n\t\r\nGiven input buffer will be analysed to guess used encoding. The result can be used as control parameter for charset conversation procedure.\r\n\t\r\nCharset Detector can be compiled (and hopefully used) for MS Windows (as dll - dynamic link library) or Linux.\r\n\t\r\nBased on Mozilla's i18n component - http://www.mozilla.org/projects/intl/. \r\n\r\nCharset Detector is open source project and distributed under Lesser GPL.\r\nSee the GNU Lesser General Public License for more details - http://www.opensource.org/licenses/lgpl-license.php\r\n\r\nNikolaj Yakowlew \xFFFFFFA9 2006-2008 \r\nTypeLib by Bruce M. Axtens, 2008.")<br />
]<br />
library CHSDET<br />
{<br />
<br />
[<br />
dllname("CHSDET.dll"),<br />
version(1.0),<br />
helpstring("Functions in CHSDET.DLL")<br />
]<br />
module CHSDETFunctions {<br />
[entry(0x60000000), helpstring("Returns rAbout record (qv)")]<br />
void _stdcall GetAbout([in, out] rAbout* AboutRec);<br />
[entry(0x60000001), helpstring("Reset detector. Prepares for new analysis.")]<br />
void _stdcall Reset();<br />
[entry(0x60000002), helpstring("Analyse given buffer of specified length.<br />
<br />
Return value is of eHandleDataErrors, either <br />
NS_ERROR_OUT_OF_MEMORY (Unable to create internal objects) or NS_OK.<br />
<br />
Function can be called more that one time to continue guessing. Charset Detector remembers last state until Reset called.")]<br />
void _stdcall HandleData(<br />
[in] BSTR aBuf, <br />
[in] short aLen, <br />
[out, retval] short* retVal);<br />
[entry(0x60000003), helpstring("Returns either TRUE (Charset Detector is sure about text encoding.) or FALSE.<br />
<br />
NB: If input buffer is smaller then 1K, Charset Detector returns FALSE.")]<br />
void _stdcall IsDone([out, retval] short* retVal);<br />
[entry(0x60000004), helpstring("Signal data end. If Charset Detector hasn't sure result (IsDone = FALSE) the best guessed encoding will be set as result.")]<br />
void _stdcall DataEnd();<br />
[entry(0x60000005), helpstring("Returns guessed charset as rCharsetInfo record")]<br />
void _stdcall GetDetectedCharset([out, retval] rCharsetInfo* retVal);<br />
[entry(0x60000006), helpstring("Returns all supported charsets in form "0x0A Name - CodePage"")]<br />
void _stdcall GetKnownCharsets(<br />
[in, out] long* sList, <br />
[out, retval] long* retVal);<br />
[entry(0x60000007), helpstring("Return eBOMKind value matching byte order mark (if any) of input data.")]<br />
void _stdcall GetDetectedBOM([out, retval] eBOMKind* retVal);<br />
[entry(0x60000008), helpstring("Remove CodePage from consideration as a possible match")]<br />
void _stdcall DisableCharsetCP([in] long CodePage);<br />
};<br />
<br />
typedef [uuid(91694067-30AB-44A9-A210-F5602935475F)]<br />
struct tagrAbout {<br />
long lMajor;<br />
long lMinor;<br />
long lRelease;<br />
long sAbout;<br />
} rAbout;<br />
<br />
typedef [uuid(3C8B7420-D40B-458B-8DE8-9B3D28607396)]<br />
enum {<br />
BOM_Not_Found = 0,<br />
BOM_UCS4_BE = 1,<br />
BOM_UCS4_LE = 2,<br />
BOM_UCS4_2143 = 3,<br />
BOM_UCS4_3412 = 4,<br />
BOM_UTF16_BE = 5,<br />
BOM_UTF16_LE = 6,<br />
BOM_UTF8 = 7<br />
} eBOMKind;<br />
<br />
typedef [uuid(9B231DEF-93FB-440D-B06B-D760AECE09D0)]<br />
struct tagrCharsetInfo {<br />
long Name;<br />
short CodePage; <br />
long Language;<br />
} rCharsetInfo;<br />
<br />
typedef enum {<br />
NS_OK = 0,<br />
NS_ERROR_OUT_OF_MEMORY = -2147024882<br />
} eHandleDataErrors;<br />
};<br />
|
|
|
|
|
I am having a problem with safearrays. On the client side, I create a CComSafeArray and populate it with variants. Then I assign the safearray to a variant, which I pass as an in/out parameter across COM to the server. The server extracts the safearray from the variant, and attempts to extract the elements from the safearray, but fails. Any ideas why?
Here is the client function:
void CMyClass::GetData(VARIANT* pvData)
{
...
CComSafeArrayBound bound[2];
bound[0].SetCount(parms.GetSize()/2);
bound[0].SetLowerBound(0);
bound[1].SetCount(2);
bound[1].SetLowerBound(0);
CComSafeArray<variant> sa;
sa.Create(bound, 2);
LONG index[2];
for (int i = 0; i < parms.GetSize(); i+=2)
{
index[0] = i / 2;
index[1] = 0;
ATLASSERT(S_OK == sa.MultiDimSetAt(index, CComVariant(parms.GetAt(i))));
index[1] = 1;
ATLASSERT(S_OK == sa.MultiDimSetAt(index, CComVariant(parms.GetAt(i+1))));
}
pvData->vt = VT_ARRAY|VT_VARIANT;
pvData->parray = sa.Detach();
}
</variant>
The COM call looks like this:
[id(1),helpstring("...")] HRESULT SetData([in,out] VARIANT * pvntData,[out,retval] EnumResult *pVal);
And the server function(s) looks like this:
...
if (pvntData->vt == VT_ARRAY | VT_VARIANT)
{
SafeArrayCopy(pvntData->parray, &m_psaData);
SafeArrayDestroy(pvntData->parray);
}
...
m_psaData is a member variable of type SAFEARRAY *
...
long lNumRows,lNumCols;
long idxColRow[2];
SafeArrayGetUBound(m_psaData, 1, &lNumRows);
SafeArrayGetUBound(m_psaData, 2, &lNumCols);
ATLASSERT(lNumCols == 1);
VARIANT vParm1;
VARIANT vParm2;
for (int row = 0; row <= lNumRows; row++)
{
idxColRow[0] = row;
idxColRow[1] = 0;
SafeArrayGetElement(m_psaData, idxColRow, (void *)&vParm1);
idxColRow[1] = 1;
SafeArrayGetElement(m_psaData, idxColRow, (void *)&vParm2);
...
At this point, vParm1 and vParm2 are empty. Why am I not seeing the correct values?
To make matters more interesting...I DO see the expected values when I step through the debugger. But when I build a release executable, the values are empty.
Can anyone here help??
Thank you,
Tim
|
|
|
|
|
We product two COM components that our customers use. We are currently using a tool based loosly on the old NDoc tool to find differences between versions (yes its odd to use NDoc on type libraries, but it used to do a lot more). I am looking to replace that tool with one designed to report differences between two releases of a TypeLib.
Does anyone have any suggestions?
Tanks for your support
Pat O
<a href="http://currentchaos.blogspot.com/">Blog</a>
_ _ _
|
|
|
|
|
I use OleView.exe that comes along with Visual Studio to view type libraries and check for differences.
«_Superman_»
|
|
|
|
|
Thanks, I was hoping for more of an automated solution.
Tanks for your support
Pat O
<a href="http://currentchaos.blogspot.com/">Blog</a>
_ _ _
|
|
|
|
|
You could write a simple program to do that.
Use the ITypeLib and ITypeInfo interfaces.
«_Superman_»
I love work. It gives me something to do between weekends.
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
