Re: [boost] Beta posted for new regression tests

From: Beman Dawes (bdawes_at_[hidden])
Date: 2002-07-24 09:22:30

At 06:43 AM 7/24/2002, John Maddock wrote:

 >> It only covers three compilers, but should be enough to get the flavor.
 >
 >That look excellent, however a couple of comments:
 >
 >* If the program is run and the run succeeds then no output is present in
 >the table, so we can't
 >check for "informative" messages - from config_info or limits_test for
 >example.

That was a deliberate decision to hold down the size of the status file.

I should probably add an option to include them, and then see if it 
actually has much impact.  If a particular test is generating excessive 
output, we can always (1) badger the developer to cut back on messages, and 
(2) truncate messages over a certain size, as is already being done with 
compiler messages.

 >* If building a library fails, and the table contains "Missing", then we
 >don't have a link to the library build failures to see what the problem 
is.

Yes, and fixing that is very high priority on my "do list".

Thanks for the comments,

--Beman