Re: [RFA] Testcase for bug report 11531

[Joel Brobecker <brobecker at adacore dot com>]

> +# Do not use run_to main
> +# as this sets a breakpoint at main.
> +# If the breakpoint is at the same instruction as the 
> +# watchpoint value assignment
> +# you can fall into the problem of the stepping over the breakpoint
> +# location that can also trigger a watchpoint miss
> +# This is not the problem reported here.
> +
> +gdb_test "start" ".*Temporary breakpoint.*"

Unfortunately, this does not work when testing with the remote protocol.
Do use "runto", but use "*main" instead of "main", or something like
that. That way, you should always stop before the instruction that
causes the watchpoint to trigger.

> +gdb_test "next" \
> +    "Old value = 0.*New value = 5.*" \
> +    "watchpoint variable triggers at next"
> +
> +gdb_test "continue" \
> +    "Old value = .*New value = 78.*" \
> +    "watchpoint variable triggers at continue"

This is a bit on the paranoid side, but can we use end-of-line
expressions rather than ".*" to match the line break? For instance,
I often use:

    set eol "\[\r\n\]+"
    gdb_test "next" \
        "Old value = 0${eol}New value = 5${eol}" \
        "watchpoint variable triggers at next"

It's unlikely that you'll have any other number than zero if the
leading digit is zero - but conceivably, we could print the new
value as 55 and still pass the test...

I did suggest at one time to provide a routine that takes a list of
regexp strings, and builds a regexp string joined with the typical
new-line regexp.  That way, I could write the test as follow:

    test_gdb_complete "pck" \
                      [multi_line "(p pck\\.ad\[sb\])?" \
                                  "(p pck\\.ad\[sb\])?" \
                                  "p pck.external_identical_one" \
                                  "p pck.inner.inside_variable" \
                                  "p pck.local_identical_one" \
                                  "p pck.local_identical_two" \
                                  "p pck.my_global_variable" \
                                  "p pck.proc" ]


Daniel made another suggestion, which I needed time to investigate
and yet never did :-(.

-- 
Joel