This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [RFA/doco/testsuite] Document new gdb_test_timeout global variable.


> > +If not specifically defined, this variable gets automatically defined
> > +to the same value as @code{timeout} during the testsuite initialization
> > +(the actual value depends on the exact running parameters).
> 
> It would be worth saying more about how to find out the default
> value.  If I'm a reader of this section, I might ask myself whether I
> need to set the variable to a non-default value, but it is impossible
> to answer that question without knowing what would be the value if I
> don't do anything.

The problem is that this is tricky and really depends a lot on whether
the user uses a board file or not, since dejagnu is so configurable
(you should see the list of files that get loaded at startup).  I'm not
even certain that *I* know the entire story: As far as I can tell,
gdb/testsuite/config/unix.exp sets a timeout of 60 seconds.  I don't
think we set the timeout anywhere else.  So that leaves the question
of board files. If the user used board files, then they may or may not
have also set the timeout there.

On the other hand, one easy way to determine the default timeout is
just to add a trace in one testcase that prints it. Crude, but
effective, since this is how I noticed that the default timeout was
not the value that I thought it would be (and hence ended up noticing
the change in unix.exp, only after some heavy grepping).

So what should do we do in this case?  I've fixed all the other nits
that you mentioned, except this... Perhaps we should just refer the user
to the dejagnu documentation somewhere earlier.

-- 
Joel


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]