Re: Just installed Delphi 2007 April 08 patch...

{$DEFINE EnableMemoryLeakReporting}
the stack exception does not happen... Quite odd...

I would like to correct myself...


I am not sure April patch is alone / at-all to blame.
I tracked the problem down to Indy snapshot (few days old)

System.InitUnits;
{$IFDEF MSWINDOWS}
TProc(P)();
{$ENDIF}

IdStack.Initialization
{$IFDEF REGISTER_EXPECTED_MEMORY_LEAK}
IndyRegisterExpectedMemoryLeak(GStackCriticalSection);
{$ENDIF}

IdGlobal
{$IFDEF VCL10ORABOVE}
Result := System.RegisterExpectedMemoryLeak(AAddress);
{$ELSE}

*System.RegisterExpectedMemoryLeak* then calls itself recursively...
The reason is that since my FastMM options have not overriden
*MemoryManager.RegisterExpectedMemoryLeak* the call
calls itself again and again and again...




I am still wondering if perhaps April patch has
messed with some default define or...?

This is simply first time I have tried making a release build
without *EnableMemoryLeakReporting* since updating Indy.


Anyways, this is probably the wrong group! Sorry!


best regards
Thomas Schulz


.

Relevant Pages

  • Re: April08 Hofix released
    ... I am not sure April patch is alone / at-all to blame. ... I tracked the problem down to Indy snapshot ... {$IFDEF MSWINDOWS} ...
    (borland.public.delphi.non-technical)
  • RE: [Full-Disclosure] Whos to blame for malicious code?
    ... I don't blame you. ... > months after the patch is released? ... Shouldn't a vendor always assume the worst in users? ... This debate has never been about Open Source. ...
    (Full-Disclosure)
  • RE: [Full-Disclosure] Whos to blame for malicious code?
    ... The patch for Blaster came out 26 days prior to the ... Is *that* Microsoft's fault ... Do those users not share at least *some* of the blame? ... The point why I brought up OpenBSD is ...
    (Full-Disclosure)
  • Re: AIM Password theft VU#865940
    ... >then trying avoid the blame by passing it to someone else? ... The patch does not work. ... keep in mind that an admin may have to consider many patches ... since in fact they did not fix it. ...
    (Cert)
  • RE: [Full-Disclosure] Whos to blame for malicious code?
    ... deploying patches to an enterprise in a timely manner. ... My point is not that Microsoft is blameless. ... the problem *still* won't be solved if users don't patch. ... Tobias wants to lay *all* the blame at ...
    (Full-Disclosure)