Macports needs a little marketing ....
cal at macports.org
Tue Nov 8 22:15:01 CET 2016
On Tue, Nov 08, 2016 at 08:26:21PM +0100, Bachsau wrote:
> Stable, as I see it, means every port in every available variant
> builds out of the box
That alone is not testable. For example, the nginx port has 32 variants.
For each variant, you can either enable it, or not, so for each variant
you have two choices. The first variant gives you two choices. Adding a
second variant brings you up to four. The third variant doubles the
number of choices again, so you're at 8.
For 32 variants, that 2^32 = 4294967296 possibilities. On my MacBook
Pro, a build of nginx takes 13.25 wallclock seconds. That's 1850 years
to do a testbuild of every possible combination, and that's just for one
port on one OS version, and you're suggesting that we do this for every
port on every OS version.
> and does not have major incompatibilities with other ports.
This requires time to test (i.e. manpower) which is already a scarce
resource for MacPorts.
> > Who decides when to update it?
> Apple. A new stable version should be released for every iteration of
And when do we backport fixes? Do we fix all bugs, critical bugs,
security bugs only, or not even those? Who does the work of backporting
bugfixes to older versions? (And you're guessing it, that work requires
More information about the macports-users