Due to incorrect escaping, the wrapper script on Darwin currently bakes
in "installerPhase" rather than "$@". Postman doesn't really seem to do
anything with cli args as far as I can tell, but fixing this aligns it
better with other packages that also provide app wrappers on Darwin.
So we are adding a simplified version that builds a monolithic nix binary to get finished
in time for the release. Afterwards we will switch to the modular build again.
This reverts commit 0ee2449821.
In Nixpkgs, we consider the linker a property of the platform.
Individual packages should not be using toolchains other than the ones
that are part of the platform without a strong justification. In this
case, the only justification appears to have been that upstream uses
this linker. Just like we don't force every package whose upstream
happens to use Clang in their own builds to build with Clang in
Nixpkgs rather than the platform compiler, upstream's choice alone
isn't grounds for overriding the platform linker here.
This mostly reverts commit 63a365d8ff.
I've left the deletion of the postPatch from the original commit stand
— it seems to have been included in the original commit by mistake
instead of being a separate commit, as it has nothing to do with mold.
In Nixpkgs, we consider the linker a property of the platform.
Individual packages should not be using toolchains other than the ones
that are part of the platform without a strong justification. In this
case, the only justification appears to have been that upstream uses
this linker. Just like we don't force every package whose upstream
happens to use Clang in their own builds to build with Clang in
Nixpkgs rather than the platform compiler, upstream's choice alone
isn't grounds for overriding the platform linker here.
This reverts commit 801b31bb36.
In Nixpkgs, we consider the linker a property of the platform.
Individual packages should not be using toolchains other than the ones
that are part of the platform without a strong justification. In this
case, the only justification appears to have been that upstream uses
this linker. Just like we don't force every package whose upstream
happens to use Clang in their own builds to build with Clang in
Nixpkgs rather than the platform compiler, upstream's choice alone
isn't grounds for overriding the platform linker here.
This reverts commit d9344e9f67.
In Nixpkgs, we consider the linker a property of the platform.
Individual packages should not be using toolchains other than the ones
that are part of the platform without a strong justification. In this
case, the only justification appears to have been that upstream uses
this linker. Just like we don't force every package whose upstream
happens to use Clang in their own builds to build with Clang in
Nixpkgs rather than the platform compiler, upstream's choice alone
isn't grounds for overriding the platform linker here.
Josh Heinrichs