Marcono1234
Marcono1234
For reference, it seems the [way R8 does it](https://r8.googlesource.com/r8/+/490bc53d9310d4cc2a5084c05df4aadaec8c885d/src/main/java/com/android/tools/r8/JdkClassFileProvider.java#113) is to use the `jrt:/` file system to load the JDK class files[^1]. Which seems to work even if the jmod...
@piazzesiNiccolo-GS, I have created a proof-of-concept for that: #504 Hopefully that is useful for you.
Thanks! Feel free to close this issue if you think this is resolved.
If I understand it correctly, that was only a bug without any security impact. Zip4j treats an entry with the name `/` as being the directory to which the files...
It looks like this has been fixed by #68468.
`Value` has similar `is_...` methods as well. Maybe there could be a common `ValueKind` (or similar) used for both, instead of adding this specific `RawKind` (and risking that it needs...
Ah sorry, I did not mean merging _`Value`_ and `RawKind`, but rather have a `ValueKind` (or similar) which is used by `Value` and `RawValue`, e.g.: - `Value::kind(&self) -> ValueKind` -...
Maybe this is not that important, or needed at all, because Dependabot keeps the versions up to date and makes sure `Cargo.toml` and `Cargo.lock` are in sync. Might mainly be...
An implementation for this is in the https://github.com/Marcono1234/struson/tree/rust-2024-edition branch
Side note: It is probably not possible to create an integration test here for ModiTect because it looks like this issue cannot be reproduced when using the Eclipse compiler in...