Re: Ada Core Technologies and Ada95 Standards

[john@assen.demon.co.uk (John McCabe)]

bobduff@world.std.com (Robert A Duff) wrote:

>>We do this by using LDRA Testbed with limits on the minimum level of
>>statement and branch coverage of 100%, and 70% on LCSJ's. I'm not sure
>>exactly where those figures are derived from, but they seem
>>reasonable. The only problem here is that we've found a few bugs in
>>that tool as well!

>Surely that's not the *only* problem.  Surely the test cases fail to
>cover every combination of requirements, and therefore some bugs slip
>through.

To some extent this is obviously true. However, one way we have tried
to minimise the effect of this is to keep a close working relationship
with our customer, to ensure that we are as completely aware as
possible of what the possible combinations of requirements are. In
this way our testing can simulate the real operation of the instrument
as possible. One particular case of this is where we have been in
discussions with our customer as to the how they intend to implement
the command sequencing on the 1553 interface. Through these
discussions we have found out that the timing requirements of events
at our end of the bus are also used at the other end of the bus as
minimum inter-command gaps. Therefore if we complete our command
processing (i.e. initiate events and finish any housekeeping etc)
within those times, we do not need to consider the case where the next
command arrives while we are still processing the last one.

>>... Based on
>>the methods you and I use, would it not be better to use the ACVC
>>suite as a basis for the compiler vendors tests, and also require the
>>compiler vendors to submit their own test suites for approval.

>At least *some* of the compiler vendors' tests include proprietary code
>from their customers, and they simply cannot release that code.

I can see that being a minor problem. I would have thought it would be
fairly simple to produce a minimal, modified version of most problem
reports that could be released without jeopardising the
confidentiality of the customer's code. I, for example, tend to cut my
problem reports down to the bare minimum of code which causes a
problem.



Best Regards
John McCabe <john@assen.demon.co.uk>