Update to secp 0.26 and secp-sys 0.8 #62
Conversation
Tibo-lg
force-pushed
the
update-upstream-dependencies
branch
from
6a154b7 to
737bfd1
Compare
|
I had to implement the traits to make it work with 1.41.1 but not sure what is causing the CI failures. I'll have a look at it again but if someone has a suggestion let me know. |
|
I'm quite confused, unless I'm missing something it looks like actions on the same commit passed in my forked repo: |
Tibo-lg
force-pushed
the
update-upstream-dependencies
branch
from
737bfd1 to
bd24a81
Compare
|
Investigating the CI failure I found that for some reason it only happens on recent version of OSX (I've updated recently and before the update I could not reproduce the issue, now I can). The issue also seem to happen on master and also on |
Tibo-lg
force-pushed
the
update-upstream-dependencies
branch
from
bd24a81 to
fa5d316
Compare
Tibo-lg
changed the title
|
CI failure seems to be same as rust-pcap/pcap#280 (to take an example from a random project). Apparently "rust 1.41 does not work on macos >= 12". So we will just need to adjust CI. @Tibo-lg could you add a commit which changes CI to not test 1.41 on Mac? |
Tibo-lg
force-pushed
the
update-upstream-dependencies
branch
from
0ce330b to
a1cff98
Compare
@apoelstra Done! I've modified a bit the git action file removing the target field which seemed unused, making it easier to exclude 1.41 on macos, but I hope that doesn't break anything. |
|
Looks good. I'll ACK and merge this since it fixes CI. But cc @jonasnick is there any specific reason we had the target field here? Was it to test different endianness or anything? |
|
I don't remember. It seems redundant. |
This PR updates the rust-secp256k1(-sys) dependencies to latest released ones. I tried to mimic what was done upstream. Regarding the trait implementations (
Eq,PartialEq,...), since the sys types in this crate don't seem to haveserializefunctions I felt like it was ok to just derive then but please correct me if that's wrong.