On Tue, 2006-06-20 at 13:43 -0700, Richard L. Hamilton wrote:
> [...]
> > When I was halfway or even 2/3 done with my changes
> > both libxml2
> > and libxslt had new micro releases, with lots of
> > nice bug fixes.
> > Even though I'm all for using the latest and
> > greatest community code,
> > I decided not to waste a week and start again with
> > the new releases.
> [...]
>
> Would it have taken a week even just updating rather than restructuring?
Well, let's see, once the restucturing was done, if I had wanted to
update to the new microversions, I would still have had to update:
DISTFILES x 2
DISTDIRS x 2
EXFILES x 2
Makefile.sfw x 2
mapfile x 2
install-libxml2
install-libxml2-64
install-libxslt
install-libxslt-64
prototype_com x 6
pkginfo.tmpl x 6
Some of these (prototype files, mapfiles) may or may not have changed,
some of the others (pkginfo files) are just version number changes.
Then rebuild, retest, test in a full nightly, if problems found, repeat.
Laca
> Because if so, not only is the overhead likely to be an issue (in
> cost-conscious
> times especially), but the odds of getting in one last update in time for some
> {testing, packaging} deadline are that much slimmer. Apparently people want
> any given sfw to be as up to date as possible (although I suspect they're not
> willing
> to give much if anything up on reliability for that). So the time to drop in
> a micro
> update of any one external element, rebuild everything and recheck that not
> only
> was that one item still working at least as well as before, but nothing
> dependent
> on it broken either, has to be as short as possible, IMO.
