Thunderbird 128.3.3esr (64-bit) fails to connect to etesync-dav using Apple Silicon build of etebase #56
Comments
|
I tried to setup etesync-dav under Debian on an AMD64 VM, just to make sure that there's nothing wrong with my account and that ARM64, macOS, or the new version of etebase respectively are actually what causes the issue. But I can't get that to work either. When I try to install etesync-dav with |
|
Okay, the offending instruction is apparently part of the cryptography module: $ python -mpdb etesync-dav
> /home/odin/etesync-test/bin/etesync-dav(3)<module>()
-> import os
(Pdb) b /home/odin/etesync-test/lib/python3.11/site-packages/etesync/crypto.py:15
Breakpoint 1 at /home/odin/etesync-test/lib/python3.11/site-packages/etesync/crypto.py:15
(Pdb) c
> /home/odin/etesync-test/lib/python3.11/site-packages/etesync/crypto.py(15)<module>()
-> from cryptography.hazmat.backends import default_backend
(Pdb) n
> /home/odin/etesync-test/lib/python3.11/site-packages/etesync/crypto.py(16)<module>()
-> from cryptography.hazmat.primitives import serialization, hashes, padding
(Pdb)
Illegal instructionI'd open an issue in the respective repository, but I can't find the repository that hosts the etesync package. At any rate, this should probably be filed with pyca. Maybe I can find some old AMD64 laptop on which I can test things instead. |
|
Filed at pyca/cryptography#12069 |
|
The radicale downgrade fixed that issue! That leaves the build error and the missing |
The Thunderbird calendar fails to connect to etesync-dav using an Apple Silicon build of etebase from this repository as of commit 89f9b0d. The macOS Calendar.app fails, too. But given how tricky it is to setup etesync in Calendar.app, I won't be trying that as long as I'm not sure it works in principle. This blocks providing Apple Silicon builds.
The good news is that compiling etebase on Apple Silicon works in principle. If this issue can be addressed and the etebase version on PyPI is updated to 89f9b0d, building Apple Silicon binaries should be as simple as adding a
macos-14runner tomanual.yml; that runner runs on Apple Silicon. The current version of etebase on PyPI does not build on that runner (for the reasons given in issue #328 over at etesync-dav).System
Steps to reproduce
# Still using the virtual environment in target etesync-dav --logging-level debugThis leads to:
etesync-dav's debug log reads:
The lines
seem particularly interesting.
The text was updated successfully, but these errors were encountered: