Re: [ros-dev] [ros-diffs] [ekohl] 53201: [ADVAPI32] Fix EnumServicesStatusEx[A/W]: - If lpServices is NULL or cbBufSize is less than sizeof(ENUM_SERVICE_STATUS_PROCESS) pass a pointer to an internal status buffer to REnumSe...

[Jérôme Gardou]

Passing [in] and [unique] to RQueryServiceonfigW is not a solution 
either. It was a quick and dirty hack to get things working.

I'm working on a proper solution right now, unless you already have 
something in your hat :-)

Le 13/08/2011 12:53, [e-mail] a écrit :
> Author: ekohl
> Date: Sat Aug 13 10:53:15 2011
> New Revision: 53201
>
> URL: http://svn.reactos.org/svn/reactos?rev=53201&view=rev
> Log:
> [ADVAPI32]
> Fix EnumServicesStatusEx[A/W]:
> If lpServices is NULL or cbBufSize is less than sizeof(ENUM_SERVICE_STATUS_PROCESS) pass a pointer to an internal status buffer to REnumServicesStatusExA/W. Adding 'in' and 'unique' attributes in the idl file is NOT an option because this is not compatible with Windows.
> Check the InfoLevel.
>
> Modified:
>      trunk/reactos/dll/win32/advapi32/service/scm.c
>
> Modified: trunk/reactos/dll/win32/advapi32/service/scm.c
> URL: http://svn.reactos.org/svn/reactos/trunk/reactos/dll/win32/advapi32/service/...
> ==============================================================================
> --- trunk/reactos/dll/win32/advapi32/service/scm.c [iso-8859-1] (original)
> +++ trunk/reactos/dll/win32/advapi32/service/scm.c [iso-8859-1] Sat Aug 13 10:53:15 2011
> @@ -1150,7 +1150,9 @@
>                         LPDWORD lpResumeHandle,
>                         LPCSTR pszGroupName)
>   {
> ENUM_SERVICE_STATUS_PROCESSA ServiceStatus;
>  LPENUM_SERVICE_STATUS_PROCESSA lpStatusPtr;
> DWORD dwBufferSize;
>  DWORD dwError;
>  DWORD dwCount;
>
> @@ -1166,6 +1168,18 @@
>       {
>           SetLastError(ERROR_INVALID_HANDLE);
>           return FALSE;
>
> }
>
> if (lpServices == NULL ||
>    cbBufSize<  sizeof(ENUM_SERVICE_STATUS_PROCESSA))
>
> {
>    lpStatusPtr =&ServiceStatus;
>    dwBufferSize = sizeof(ENUM_SERVICE_STATUS_PROCESSA);
>
> }
>
> else
>
> {
>    lpStatusPtr = (LPENUM_SERVICE_STATUS_PROCESSA)lpServices;
>    dwBufferSize = cbBufSize;
> }
>
> RpcTryExcept
>
> @@ -1174,8 +1188,8 @@
>                                            InfoLevel,
>                                            dwServiceType,
>                                            dwServiceState,
>                                     (LPBYTE)lpServices,
>                                     cbBufSize,
>                                     (LPBYTE)lpStatusPtr,
>                                     dwBufferSize,
>                                      pcbBytesNeeded,
>                                      lpServicesReturned,
>                                      lpResumeHandle,
>
> @@ -1189,18 +1203,20 @@
>   if (dwError == ERROR_SUCCESS || dwError == ERROR_MORE_DATA)
>   {
>    lpStatusPtr = (LPENUM_SERVICE_STATUS_PROCESSA)lpServices;
>    for (dwCount = 0; dwCount<  *lpServicesReturned; dwCount++)
>    if (InfoLevel == SC_ENUM_PROCESS_INFO)
>     {
>        if (lpStatusPtr->lpServiceName)
>            lpStatusPtr->lpServiceName =
>                (LPSTR)((ULONG_PTR)lpServices + (ULONG_PTR)lpStatusPtr->lpServiceName);
>        if (lpStatusPtr->lpDisplayName)
>            lpStatusPtr->lpDisplayName =
>                (LPSTR)((ULONG_PTR)lpServices + (ULONG_PTR)lpStatusPtr->lpDisplayName);
>        lpStatusPtr++;
>        for (dwCount = 0; dwCount<  *lpServicesReturned; dwCount++)
>        {
>            if (lpStatusPtr->lpServiceName)
>                lpStatusPtr->lpServiceName =
>                    (LPSTR)((ULONG_PTR)lpServices + (ULONG_PTR)lpStatusPtr->lpServiceName);
>            if (lpStatusPtr->lpDisplayName)
>                lpStatusPtr->lpDisplayName =
>                    (LPSTR)((ULONG_PTR)lpServices + (ULONG_PTR)lpStatusPtr->lpDisplayName);
>            lpStatusPtr++;
>        }
>     }
> }
>
> @@ -1234,11 +1250,31 @@
>                         LPDWORD lpResumeHandle,
>                         LPCWSTR pszGroupName)
>   {
>
> ENUM_SERVICE_STATUS_PROCESSW ServiceStatus;
>  LPENUM_SERVICE_STATUS_PROCESSW lpStatusPtr;
>
> DWORD dwBufferSize;
>  DWORD dwError;
>  DWORD dwCount;
>
> TRACE("EnumServicesStatusExW() called\n");
>
> if (InfoLevel != SC_ENUM_PROCESS_INFO)
>
> {
>    SetLastError(ERROR_INVALID_LEVEL);
>    return FALSE;
>
> }
>
> if (lpServices == NULL ||
>    cbBufSize<  sizeof(ENUM_SERVICE_STATUS_PROCESSW))
>
> {
>    lpStatusPtr =&ServiceStatus;
>    dwBufferSize = sizeof(ENUM_SERVICE_STATUS_PROCESSW);
>
> }
>
> else
>
> {
>    lpStatusPtr = (LPENUM_SERVICE_STATUS_PROCESSW)lpServices;
>    dwBufferSize = cbBufSize;
>
> }
>
> RpcTryExcept
>  {
>
> @@ -1246,8 +1282,8 @@
>                                            InfoLevel,
>                                            dwServiceType,
>                                            dwServiceState,
>                                     (LPBYTE)lpServices,
>                                     cbBufSize,
>                                     (LPBYTE)lpStatusPtr,
>                                     dwBufferSize,
>                                      pcbBytesNeeded,
>                                      lpServicesReturned,
>                                      lpResumeHandle,
>
> @@ -1261,18 +1297,20 @@
>   if (dwError == ERROR_SUCCESS || dwError == ERROR_MORE_DATA)
>   {
>    lpStatusPtr = (LPENUM_SERVICE_STATUS_PROCESSW)lpServices;
>    for (dwCount = 0; dwCount<  *lpServicesReturned; dwCount++)
>    if (InfoLevel == SC_ENUM_PROCESS_INFO)
>     {
>        if (lpStatusPtr->lpServiceName)
>            lpStatusPtr->lpServiceName =
>                (LPWSTR)((ULONG_PTR)lpServices + (ULONG_PTR)lpStatusPtr->lpServiceName);
>        if (lpStatusPtr->lpDisplayName)
>            lpStatusPtr->lpDisplayName =
>                (LPWSTR)((ULONG_PTR)lpServices + (ULONG_PTR)lpStatusPtr->lpDisplayName);
>        lpStatusPtr++;
>        for (dwCount = 0; dwCount<  *lpServicesReturned; dwCount++)
>        {
>            if (lpStatusPtr->lpServiceName)
>                lpStatusPtr->lpServiceName =
>                    (LPWSTR)((ULONG_PTR)lpServices + (ULONG_PTR)lpStatusPtr->lpServiceName);
>            if (lpStatusPtr->lpDisplayName)
>                lpStatusPtr->lpDisplayName =
>                    (LPWSTR)((ULONG_PTR)lpServices + (ULONG_PTR)lpStatusPtr->lpDisplayName);
>            lpStatusPtr++;
>        }
>     }
> }

[Eric Kohl]

Hello Jérôme,

I am already cleaning up the remainig service functions. On my machine 
the number of failing tests is down to 14 and shrinking by the hour. ;-)

Regards,
Eric

Am 13.08.2011 13:30, schrieb Jérôme Gardou:
> Passing [in] and [unique] to RQueryServiceonfigW is not a solution
> either. It was a quick and dirty hack to get things working.
>
> I'm working on a proper solution right now, unless you already have
> something in your hat :-)

Hiya

What about advapi32:service test crash?

On Sat, 13 Aug 2011 21:51 +0200, "Eric Kohl" [e-mail] wrote:
> Hello Jérôme,
>
> I am already cleaning up the remainig service functions. On my machine 
> the number of failing tests is down to 14 and shrinking by the hour. ;-)
>
> Regards,
> Eric
>
> Am 13.08.2011 13:30, schrieb Jérôme Gardou:
> > Passing [in] and [unique] to RQueryServiceonfigW is not a solution
> > either. It was a quick and dirty hack to get things working.
> >
> > I'm working on a proper solution right now, unless you already have
> > something in your hat :-)
>
> Ros-dev mailing list
> [e-mail]
> http://www.reactos.org/mailman/listinfo/ros-dev

With best regards
Caemyr

[Eric Kohl]

Hi

Just run the advapi32:service tests from a gcc-rbuild build and you will 
not see any crash.

IMO the cmake builds are still broken and the test systems seem to be 
pretty instable as well.

BTW, I am building/testing gcc-rbuild only.

Regards,
Eric

Am 14.08.2011 00:24, schrieb [e-mail]:
> Hiya
>
> What about advapi32:service test crash?
>
> On Sat, 13 Aug 2011 21:51 +0200, "Eric Kohl"[e-mail]  wrote:
> > Hello Jérôme,
> >
> > I am already cleaning up the remainig service functions. On my machine
> > the number of failing tests is down to 14 and shrinking by the hour. ;-)
> >
> > Regards,
> > Eric
> >
> > Am 13.08.2011 13:30, schrieb Jérôme Gardou:
> > > Passing [in] and [unique] to RQueryServiceonfigW is not a solution
> > > either. It was a quick and dirty hack to get things working.
> > >
> > > I'm working on a proper solution right now, unless you already have
> > > something in your hat :-)
> >
> > Ros-dev mailing list
> > [e-mail]
> > http://www.reactos.org/mailman/listinfo/ros-dev
>
> With best regards
> Caemyr
>
> Ros-dev mailing list
> [e-mail]
> http://www.reactos.org/mailman/listinfo/ros-dev