[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Test for DejaGnu support that runs DejaGnu [PATCH]

From: Richard Dawe
Subject: Re: Test for DejaGnu support that runs DejaGnu [PATCH]
Date: Tue, 15 Apr 2003 09:35:18 +0100


Thanks for reviewing the patch! I agree with all your points and I'll fix
them. I'll respond to some points directly.

Alexandre Duret-Lutz wrote:
> >>> "Rich" == Richard Dawe <address@hidden> writes:
>  Rich> Perhaps the check for 'runtest' in tests/defs should be a little
>  Rich> more strict - perhaps it should grep for some string that
>  Rich> 'runtest --version' outputs.
> IMHO it's fine as is.  (Except for the locally inconsistent
> spacing you used inside the parentheses -- I say "locally",
> because this file is globally inconsistent anyway.)

I used Emacs to write the patch and use "tab" to ensure the lines were lined
up like the other cases in the case statement. Any tips on how to correct the

> (tests/README contains a sort of small checklist about such
> things -- it's probably incomplete, but it's a start)

Thanks for the hint. I've read that now!

>  Rich> +
>  Rich> +chmod ug+x hammer
> Is `ug' motivated by some dejagnu requirement?

No, that's just habit. I don't usually let "others" execute the scripts I

> I'd prefer `chmod +x hammer' for consistency with other tests.


> It would be nice to ensure that dejagnu was actually run by
> `$MAKE check'.  Can you see a way to check for this?  Perhaps by
> building a file during the dejagnu test case and checking for
> this file afterwards?

In this case DejaGnu creates a log file called hammer.log and a summary file
called hammer.sum. It always creates these files, according to the manual. I
will make the test check for these files.

> Also, this seems a good occasion to make sure that dejagnu files
> are cleaned appropriately.  There have been some longstanding
> bug fixed in this area last year (grep ChangeLog.02 for
>; they would have been obvious if someone had written
> a test case.  Calling `make distcheck' (in addition to `$MAKE
> check') is probably a good way to check for cleaning, plus it
> will also make sure that files are correctly distributed and
> that VPATH builds work.

I will add a test for distcheck.

Thanks! Bye, Rich =]

Richard Dawe [ ]

reply via email to

[Prev in Thread] Current Thread [Next in Thread]