Guillaume E
Guillaume E
Glad to see it's already reported upstream :)
Glad to see it's already reported upstream :)
Looks like the `cargo build` check is not happy about that.
Thank you for the contribution. - As far as I understand it, the main difference is about making the cache size a `u64`. The `normalize()` function doesn't need to be...
See #9 and the "Encoder" section in the README: > For now, there is also a dumb encoder that only uses byte literals, with many hard-coded constants for code simplicity....
- Breaking API changes are not a problem per se, as they can be guarded by features and a SemVer bump at the next release. - As far as I...
This filter is not yet implemented. For the moment, only the LZMA2 filter is implemented within the XZ container format: https://github.com/gendx/lzma-rs/blob/v0.3.0/src/decode/xz.rs#L173-L183.
Looking at the [tukaani project](https://github.com/tukaani-project/xz), this appears to be a filter specific to ARM: https://github.com/tukaani-project/xz/blob/v5.4/src/liblzma/api/lzma/bcj.h#L41.
FYI, given that the title mentions "unsoundness", I've opened an issue to file a security advisory for this: https://github.com/rustsec/advisory-db/issues/1984.
Thank you a lot for the feedback, that's really valuable! There are several things here. 1. Byte arrays (representing keys or signatures) of an invalid length. These aren't strictly speaking...