Hi.
I've found the problem with current implementation for the library: it corrupts some binary files (resubmitting of them breaks them (for example zip files). The cause of it, I think, is that you serialiser text data (name, size etc) and binary data to the same cache file. In my implementation where we have two separate caches everything works fine.
Hi. I've found the problem with current implementation for the library: it corrupts some binary files (resubmitting of them breaks them (for example zip files). The cause of it, I think, is that you serialiser text data (name, size etc) and binary data to the same cache file. In my implementation where we have two separate caches everything works fine.