On 21/11/2024 2:49 pm, Ken Cunningham wrote:
On Nov 21, 2024, at 6:44 AM, Chris Jones <jon...@hep.phy.cam.ac.uk> wrote:
On 21/11/2024 2:44 pm, Ken Cunningham wrote:
On Nov 21, 2024, at 1:29 AM, Chris Jones via macports-dev
<macports-dev@lists.macports.org> wrote:
OOTH: If gcc10 is available and installed, why would you want to call in a full
build of gcc13 unnecessarily to build the port?
If you are suggesting the builds should check to see what the user has
installed and pick a compiler based on that, then no, absolutely not.
Nobody ever suggested that.
Then what precisely are you posing to do ?
Exactly what was stated before:
current status then is we have a proposal to restrict available compilers on
systems < 10.6 to
gcc48, gcc5, gcc6, gcc7, gcc10, and gcc14
OK, The proposals where hard to follow as you keep switching between
making changes for all OSes and only for < 10.6..
It also does not explain how you would achieve your statement
"OOTH: If gcc10 is available and installed, why would you want to call
in a full build of gcc13 unnecessarily to build the port?"
If you allow me to now rewrite this as
"OOTH: If gcc10 is available and installed, why would you want to call
in a full build of gcc14 unnecessarily to build the port?"
then explain how you propose the compiler selection would work ? If a
user has gcc10 installed, but does not have gcc14, then if the gcc
selection remains as it is (use most recent available) a port build will
still pick gcc14 and install that before using it ? I cannot see how you
achieve the above without having the port first peak to see what the
user has installed and base its decisions on that.