[fa.info-vax] Debugger problem

info-vax@ucbvax.ARPA (01/19/85)

From: Allan A. Miller <AAA@SU-AI.ARPA>

I am using the debugger on a Pascal program - it allows me to set a breakpoint
in a procedure, and when I hit the breakpoint I can examine a boolean variable
in the procedure, but when I try to examine a subfield of a record that is
a dereferenced pointer (eg. s^.length), I get the error
  %DEBUG-E-NOFREE, no free storage available
This error happens even while examining global pointer variables.  The manual
suggests using the CANCEL MODULE command, but since the whole program is one
module, that won't work (the program is only about 5400 lines).

I tried increasing various memory quotas to no avail.  Also, when using the
SHOW MODULE command, the debugger indicates that there is still space
available!  Has anyone ever seen this problem?
					Allan

info-vax@ucbvax.ARPA (01/19/85)

From: engvax!KVC@cit-vax

There were a lot of problems in the VMS V3 debugger, especially with
Pascal.  According to one of the Pascal developers, there was some
disagreement between the specification for the debugger (that the
compiler writers follow) and what the debugger actually did with
regard to debugger symbol tables.  Most of the problems have been
fixed in the VMS V4 debugger.  I have been using Pascal with the V4
debugger quite a bit, and it works much better.

Besides, screen mode is sooooo much fun that you almost wanna write a
few bugs into the code just so you can use it!  :-)

	/Kevin Carosso            engvax!kvc @ CIT-VAX.ARPA
	 Hughes Aircraft Co.

rcb@rti-sel.UUCP (Randy Buckland) (01/21/85)

> From: Allan A. Miller <AAA@SU-AI.ARPA>
> 
> I am using the debugger on a Pascal program - it allows me to set a breakpoint
> in a procedure, and when I hit the breakpoint I can examine a boolean variable
> in the procedure, but when I try to examine a subfield of a record that is
> a dereferenced pointer (eg. s^.length), I get the error
>   %DEBUG-E-NOFREE, no free storage available
> This error happens even while examining global pointer variables.  The manual
> suggests using the CANCEL MODULE command, but since the whole program is one
> module, that won't work (the program is only about 5400 lines).
> 
> I tried increasing various memory quotas to no avail.  Also, when using the
> SHOW MODULE command, the debugger indicates that there is still space
> available!  Has anyone ever seen this problem?
> 					Allan

Check if the current module has it's symbols loaded. The module probably has
too many symbols for the free memory.

					Randy Buckland
					Research Triangle Institute
					...!mcnc!rti-sel!rcb