Bug with AllocateMemory

[lionel_om]

But with using "*mem\s = *str\s", PB made the copy and create a new buffer . Or I'm wrong...?

No, you're quite right.
PB's way of handling the '@' operator for strings can be a bit confusing at times. As you can se from the different answers you're getting

So my procedure is quite good, excepted the null pointer case ?!

Code: Select all

Procedure.l V__Constructor_String(*param = #Null)
  Protected *str.String, *mem.String
  *mem = Vector__Allocate(SizeOf(String))
  If *param
    *str = @*param
    *mem\s = *str\s
  Else
    *mem\s = ""
  EndIf
  ProcedureReturn *mem
EndProcedure

So the problem came from another place ?!

[Pupil]

So my procedure is quite good, excepted the null pointer case ?!

So the problem came from another place ?!

Yes i think the procedure is just fine, so there might be another problem somewhere in the code. One thing i wonder though, how is the result from the string constructor procedure handled?

[lionel_om]

So my procedure is quite good, excepted the null pointer case ?!

So the problem came from another place ?!

Yes i think the procedure is just fine, so there might be another problem somewhere in the code. One thing i wonder though, how is the result from the string constructor procedure handled?

The return is used has a pointer of a STRING structure :

Code: Select all

*a.STRING = CallMyFunction()

[Pupil]

That should be OK.

Sorry, i'm afraid you have to do some more bug hunting

Are you using any threads or are the code compiled into a DLL?

[lionel_om]

That should be OK.

Sorry, i'm afraid you have to do some more bug hunting

Are you using any threads or are the code compiled into a DLL?

Yeah, I'm using Threads, and the only way the program bug is in the Threaded functions. I hope that PB have no more bug with threads and strings.

PS : I'have checked "Thread safe" in compiler options.

Lio