Keith M Wesolowski <[EMAIL PROTECTED]> wrote:
> On Fri, Jul 01, 2005 at 05:39:22PM +0200, Joerg Schilling wrote:
> > Eric Schrock <[EMAIL PROTECTED]> wrote:
>
> > If a change is really small, it should be possible to verify the change
> > by just looking at it (doing code review).
>
> Code review is not a substitute for testing. If the expectation is
> that the binary has not changed, elfcmp or equivalent should be used
> to aid in verifying that assertion. If the binary has changed or if
> you are modifying something like a script which is delivered directly,
> you must at the very least run manual sanity tests on the affected
> functionality. If you don't at least build what you changed, how will
> you know you haven't introduced a compiler or lint warning or broken
> the build? If you don't test at all, how will you really be sure
> you've fixed the bug? The amount of testing required scales with the
> change but - like the change itself - is never zero.
For a lot of the byte-size changes, it is possible to look at the
source changes to know whether it is correct. If course not for all
of them.....
Jörg
--
EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin
[EMAIL PROTECTED] (uni)
[EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/
URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
