Inspect Manual

Using Inspect With TNS/R Native Programs

Inspect Manual—429164-006

17-6

Code Breakpoints

When debugging a TNS/R program, you can set code breakpoints at any location in

the program. To set a breakpoint in native code, you can either specify a scope name

or a machine-level code address.

Signals

TNS and accelerated processes use a trap mechanism for exception handling. TNS

and accelerated processes in the OSS environment can also use a signal mechanism

for exception handling. (A signal is a means by which a process can be notified of or

affected by an event occurring in the system.)

TNS/R native processes, regardless of whether they are in the Guardian or OSS

environment, use a signal mechanism for exception handling. (OSS processes can

also use signals for communicating between processes.) The SIG_DEBUG parameter

was added to the MODIFY SIGNAL command to specify that when a signal is

delivered, the debugger will be invoked.

For example, where a TNS process might enter a trap handler on an arithmetic

overflow, a TNS/R native process might enter a signal handler.

Refer to the Guardian Programmer’s Guide and the Open System Services

Programmer’s Guide for details on signal handlers.

Commands Useful When Debugging Native

Programs

The commands listed are beneficial when debugging TNS/R native programs. For

more information, see Section 6, High-Level Inspect Commands.

ADD PROGRAM

The ADD PROGRAM command accepts the names of one or more SRLs. The SRL

clause is a single filename, or a comma separated list of SRL object files within

parentheses. If the process used more SRLs than you specified, Inspect will emit a

warning indicating this. By default, Inspect does not load public SRL object files.

BREAK

The BREAK command supports memory access breakpoints on TNS/R native

addresses. The support of MABs on native addresses applies to all data breakpoint

subtypes (ACCESS, CHANGE, READ, WRITE, READ WRITE, WRITE READ).