FidoNet Echomail Archive

<<< Previous Index Next >>>

From: Frank Adam
To: Roy McNeill
Date: 1997-08-07 19:45:36
Subject: Knock-knock

On Aug 03 21:18, 1997, Roy McNeill of 3:712/610.16 wrote:
G'day Roy,

FA>> Anyone home here ? :-)
RM> It has gone a bit quiet here lately, true.
They've probably all moved to some java echo.:-(

FA>> member functions. It has a bug somewhere, but when i put it through
FA>> Turbo debugger, the bugger(no doubt the real origin of the name)
FA>> refuses to trace through two of the 3 functions, won't even let me
FA>> set a breakpoint there. It will happily trace through the third,
RM> You can get odd debugging effects if you didn't have the "include
RM> debugging info in objs" compiler option turned on when you compiled 
I thought i've had everything turned on here, i even looked at dirty 
pictures, just to be sure.:-) 
RM> file containing the actual source code (not the declarations) of
RM> the offending functions. The fix is to turn it on, then Rebuild All
RM> to recompile and link all source files.
Thanks, i always have that switched on. But your reply did make me go and 
check again and i found an option, which i've never touched or looked at before.
To force inline functions out of line. 
That was it, i can trace it now. Shame about all those printfs i've wasted
fixing the bug .:-) Worse, i should've thought of it, one of the first
lines i put into a C++ program is #pragma warn -inl to turn off that stupid
Oh well, learn something every day.

 Regards, Frank

--- Msged 4.00
 * Origin: The ticking point, Melbourne, Australia. (3:635/728.21)
SEEN-BY: 622/419 632/360 633/267 270 635/728 670/218
@PATH: 635/728 633/267

<<< Previous Index Next >>>