develooper Front page | perl.perl5.porters | Postings from September 2001

Re: [PATCH] make coretest

Thread Previous | Thread Next
From:
Paul Johnson
Date:
September 6, 2001 04:43
Subject:
Re: [PATCH] make coretest
Message ID:
20010906134316.H5776@pjcj.net
On Thu, Sep 06, 2001 at 12:09:26PM +0100, Simon Cozens wrote:
> On Thu, Sep 06, 2001 at 12:02:13PM +0100, Hugo van der Sanden wrote:
> > Paul Johnson <paul@pjcj.net> wrote:
> > :Here's something designed to reduce the frustration of waiting for those
> > :loooong tests.  The coretest target runs all the basic tests and the
> > :pragma tests, but doesn't run the module tests.  This takes about a
> > :third of the time of a full test.
> > 
> > Will be able to distinguish the two in a 'make ok' report? If not, it
> > would be helpful if some means to do so could be added.
> 
> The other problem with this is that it encourages people to make patches
> and send them in without going through the whole regression suite, which
> could catch unexpected interactions.

I'd hope that people wouldn't send in a "make ok" without having first
done a "make test".  I'd also hope that people wouldn't send in patches
without having done the same.  :-)  That's something I mentioned in the
doc patch when I sent it in, but that part of the patch didn't get
applied.

My preference was to document the target and provide the caveats.  The
other view, with which I have some sympathy, is to leave the target
undocumented on the assumption that the people who know about it will
use it prudently, and the people who need it will find it.

My motivation behind this patch was to enhance testing rather than to
detract from it.  Its target audience was pumpkings and core hackers.
My reasoning was that since the test target now takes much longer than
it did, primarily due the the assimilation of new modules, it can be
tempting to test less frequently than might be prudent.  A test target
which tested core functionality, used in the knowledge that it was not a
complete test, seemed a useful addition.

If it turns out that this target is weakening the test suite, then
something should be done about it.  Ultimately this might mean removing
the target.  However, I would imagine that in a short time, if things
are left as they are, only those people who need the target will
remember it, and they will use it responsibly.  I hope I am right.

In the meantime, any suggestions to make use of this target a little
safer will be gratefully received.

-- 
Paul Johnson - paul@pjcj.net
http://www.pjcj.net

Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About