"concurrent" qt4-mac: request feedback & testing for current port phase in Portfile?

[Brandon Allbery]

On Thu, Dec 11, 2014 at 4:07 AM, René J.V. <rjvbertin at gmail.com> wrote:

> > How about a main port with the new paths, and a stub port or subport that
> > depends on the main port, conflicts with qt4-mac, and installs the
> > symlinks? Then we can replace qt4-mac with the stub port at some point.
>
> I like the idea. It was my first idea (after deciding not to make a new
> port) to make a subport ... but qt4-mac already has sub-ports. My idea had
> been to change the PortGroup through a subport, but that didn't work with
> (= propagate to) the other subports.
> In the end I found that I understand variants better than subports and
> stub ports and the ways to have the dependents accept them transparently...
>

Only up to a point, though, as Ryan has been pointing out. But in this case
I meant have your Qt4 as a main port, and a subport of that that installs
the symlinks to make it compatible with the existing qt4-mac.


> > (Maybe make the stub port qt4-mac-devel, so it's maybe already handled
> as a
> > potential dependency *and* it's announcing that it will eventually
> replace
> > qt4-mac.)
>
> But then I should develop the concurrent version as the qt4-mac-devel port
> . To be honest I didn't think of that ...
> But is everything in place for Qt4 client ports to accept
> port:qt4-mac-devel as an alternative to port:qt4-mac?
>

If it can be done at all, it can only be done with actual ports, not
variants. Mostly I suggested qt4-mac-devel because of the implication that
this is the future qt4-mac, and lower confusion for users; the mechanism
for allowing multiple ports to satisfy a dependency only lists the default
port to install if no existing installed port does so; presumably that
would be qt4-mac for now. My hope is that a significant number of ports
depending on qt4-mac already use a path or bin dependency so that ports
like qt4-mac-devel or your backward compatibility shim port (assuming
previous response) could satisfy it if installed.

-- 
brandon s allbery kf8nh                               sine nomine associates
allbery.b at gmail.com                                  ballbery at sinenomine.net
unix, openafs, kerberos, infrastructure, xmonad        http://sinenomine.net
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.macosforge.org/pipermail/macports-users/attachments/20141211/88f4ad9d/attachment.html>