I'm curious if this project would have any interest in moving away from using openssl from homebrew by default? Similar to what ruby-build has already done.
While the dynamic linking to brew installed openssl versions at brew --prefix openssl@1.1 is nice, it will eventually break python installs whenever openssl bumps their formula version and a user brew upgrades it (or another package that declares a dependency on openssl updates it).
It appears like most of the code in python-build would support this change pretty easily. I replicated the changes from the ruby-build PR linked above and was able to install python versions successfully. Admittedly, I haven't tested this extensively.
Thoughts?
I'm curious if this project would have any interest in moving away from using openssl from homebrew by default? Similar to what ruby-build has already done.
While the dynamic linking to brew installed openssl versions at
brew --prefix openssl@1.1is nice, it will eventually break python installs whenever openssl bumps their formula version and a user brew upgrades it (or another package that declares a dependency on openssl updates it).It appears like most of the code in python-build would support this change pretty easily. I replicated the changes from the ruby-build PR linked above and was able to install python versions successfully. Admittedly, I haven't tested this extensively.
Thoughts?