Closed frol closed 6 years ago
Could you please explain what your new strategy is, and why it's an improvement? Do you know whether /tmp
is guaranteed to be large enough inside a Docker container for building a large C++ library, for example? Thank you!
/tmp
and any other location except the permissions (i.e. /tmp
is not mounted separately, so if /home/rust/libs
can fit all the libs then /tmp
will also do). Thus, there is no point in a separate location for temporary files.cd /tmp
) just simplifies reading (if compare it to cd ..
you will notice that you will need to scan for all the other cd
s to reason about the location you end up)./tmp
after you are done and that is it, so you don't need to list all the separate files and folders (you could have done the same with /home/rust/libs
as well).OPENSSL_VERSION
variables is just more explicit way which is just future-proof (imagine you dropped the second VERS
and that would just silently install zlib 1.0.1 instead of 1.2.11)Great! Thank you for the explanation of what you've changed. Have you also run the test-image
script to make sure that everything still works? If so, I'll go ahead and merge this.
My apologies for not having a CONTRIBUTING.md
suggesting what things to include with PRs. I should get around to that. :-) And thank you very much for the PR.
...
Successfully tagged rust-musl-builder-using-diesel:latest
Hello, world!
No DATABASE_URL set, so doing nothing
Created binary (application) `testme` project
Compiling testme v0.1.0 (file:///home/rust/src/testme)
Finished dev [unoptimized + debuginfo] target(s) in 0.46 secs
Checking to make sure it is not a dynamic executable
not a dynamic executable
OK. Tests passed.
Thank you for the PR! I've merged it.
Docker Hub should rebuild the official images sometime in the next few hours.
The main reason I touched this part was that postgres sources were left in the image and while I could fix it with a single
/..
change, I decided to make the script more reliable overall.