Tink is a multi-language, cross-platform, open source library that provides cryptographic APIs that are secure, easy to use correctly, and hard(er) to misuse.
I want to use tink inside a linux docker container running on apple silicon. This is problematic because there are no prebuilt binaries of tink for that plattform available on pypi. Ordinarily this wouldn't be a problem, because pip will just download and build the wheel on demand. However, if bazel is not available in that linux plattform (in my case centos7) that fails. This makes it fairly hard to use tink.
What sort of feature would you like to see?
I would like to see prebuilt binaries for tink for arm / apple silicon on pypi to ease and speed up usage of tink in docker, especially on apple silicon.
Have you considered any alternative solutions?
Building bazel myself to then build tink later: Possible, but involved. I may need to go that route though. :/
Adding a builder sidecar docker container with a more modern linux distribution that already ships an arm version of bazel. Possible, also more involved than I'd like, but maybe necessary?
Search for someone to provide a repo that provides bazel builds for arm already… Haven't found anything yet, but pointers would be welcome.
If there are other possible workarounds, especially if they involve less work, that would be most welcome.
Is your feature request related to a problem?
I want to use
tink
inside a linux docker container running on apple silicon. This is problematic because there are no prebuilt binaries of tink for that plattform available on pypi. Ordinarily this wouldn't be a problem, becausepip
will just download and build the wheel on demand. However, ifbazel
is not available in that linux plattform (in my casecentos7
) that fails. This makes it fairly hard to usetink
.What sort of feature would you like to see?
I would like to see prebuilt binaries for tink for arm / apple silicon on pypi to ease and speed up usage of tink in docker, especially on apple silicon.
Have you considered any alternative solutions?
If there are other possible workarounds, especially if they involve less work, that would be most welcome.