Failed to allocate a memoryblock with 8 bytes in size


In several of my apps, I occasionally get an exception telling “Failed to allocate a memoryblock with 8 bytes in size.” and the app shuts down. I’ve not been able to track the issue (seems like it always happens in the built app :man_shrugging:), but just wondered whether someone else already saw this message.

Maybe your app could be suffering some memory leaks while continuously requesting 8 bytes allocations (also not releasing them as necessary or in time?).

Or you made an extremely memory hunger app tested ok in a 16Gb+ machine that blows up in 8Gb ones?

I’m rarely using memory blocks and don’t do memory intensive things in these apps. Also, in Activity Monitor, the memory isn’t increasing much.

Not either. Not dealing much with memory-intensive things. :man_shrugging:

I’ve studied the stack of the crash, which didn’t tell anything obvious (IMO). Next time it happens, I’ll copy&paste it here.

Thanks for your answer.

8bytes is only a 64bit integer, isnt it?
Are you using pointers or AddressOf anywhere in the app, which might trample over other areas of memory?

This could help.

Good guess. I use AddressOf and delegates quite a bit in my apps. I’ll check that.
Thank you.


Always remove any handlers added with AddHandler when you are finished with them using RemoveHandler. Balance AddHandler with RemoveHandler in the appropriate place, e.g., if AddHandler is used when a Window opens, use RemoveHandler when the Window closes.

Thanks. I usually try to do that.
In the current project, I don’t use AddHandler, but I use AddressOf; I’m recompiling using WeakAddressOf to check if it goes better.
As it’s a random crash, I don’t know when I’ll validate whether it’s working better.

Ok, I got the issue again, and it was early (first and quick use); so I don’t think it’s about a memory full condition.

Here’s the stack log (shown in a MessageDialog from App.UnhandledException):

Unhandled exception in method CIPCS1.Error
Number: 0
Message: Failed to allocate a memoryblock with 8 bytes in size.
-[NSApplication(NSEvent) _nextEventMatchingEventMask:untilDate:inMode:dequeue:]
-[NSApplication run]

And, again, this happened in the built app (I can’t reproduce in the debug version).

The location of the exception is “CIPCS1.Error”, as seen in this log. CIPCS1 is an IPCSocket, and here’s the Error event code:

app.CurrentMethodDebug="CIPCS1.Error" //What's shown in the log
if PDlgReply<>nil then PDlgReply.Invoke "TERM" //A delegate
if PartialReceivedData<>"" then MessageBox "Warning: not all data was processed."+EndOfLine+PartialReceivedData //When the error occurs before the socket has handled the whole data.

The delegate is assigned that way:
Socket.PDlgReply=WeakAddressOf ResultReceived,Opt

And ResultReceived looks like this:

select case Data.Left(4)
case "rslt" //Got a result from the helper
  Var Item As FolderItem
  Var t as string=Data.Middle(4) //After “rslt”
  Var SearchID as integer=t.NthField(chr(4),2).ToInteger
  Var What as string=Left(t,4)
  Var s as string=t.Middle(4).NthField(chr(4),1) //Path to item received
  if s<>"" then Item=new FolderItem(s)
  if What="ifnd" then SearchParams.Result.Add Item //Add the found item to an array
case "TERM" 'Termination (clean or not)
  SearchIn SearchStep.ApplyRules
end select

Does this log help?

Wait, is that a MBS call there?
REALBuildMemoryblock is a function in my C++ code.

Could it be an old plugin used with newer Xojo, so the constructor call to make the memory block fails?

Possibly. So I’ve added more debug properties to my code; I’m currently awaiting for the next crash to point its location.

I recently updated both Xojo and your plugins, so no.
Thanks for the pointer.

Ok, I found where the problem occurred. For some reasons, under certain circumstances, the code was trying to access a timer on a window that was closed; this led to a NilObjectException. So I added a check against nil.

I have no idea why that leads to this “Failed to allocate a memoryblock with 8 bytes in size” message nor why the stack refers to REALBuildMemoryblock.
In a test project, I couldn’t reproduce the issue (I correctly get a NilObjectException). Also, with the new check against nil, the original app won’t crash anymore, so that was clearly the cause. Puzzling :man_shrugging:

If there’s anything I could try (as I don’t like to keep a mystery unsolved), feel free to ask.

1 Like