Zephyr Lykos
+1 on @WillPower3309, I think a fallback to opacity is required.
I guess after the gpgme release, so we could finish removing the replace directive in one PR
A few questions: - Could anyone post a strace log of GPG daemon? - Is the SQLite connection persisted? - What's the journal mode & locking mode on the SQLite...
seems great, will try to implement this after I finish my work on some Win32 drivers :sweat_smile:
done that. will look at it later this weekend
seems this is quite trivial: we just need to poll `fcntl(2)` to detect the lock. or we could link to sqlite3, but that might be overkill just for this task.
to my knowledge, no. the underlying syscall is the same after all.
But it would be easier to find the offset to lock instead of reading the database ourselves
I already poked around using inotify, unfortunately, with no avail. 😿
https://fars.ee/UpO4.log `clang version 18.1.3 (Fedora 18.1.3-2.fc41) Target: x86_64-w64-windows-gnu` https://fars.ee/Ycjz.log `gcc version 14.0.1 20240127 (Fedora MinGW 14.0.1-1.fc40) (GCC) Target: x86_64-w64-mingw32`