Mercurial > lbo > hg > leveldb-rs
changeset 410:d3abb5cd983b
meta: Prepare for publishing.
| author | Lewin Bormann <lbo@spheniscida.de> |
|---|---|
| date | Thu, 12 Oct 2017 22:03:59 +0200 |
| parents | f53a4deb9c10 |
| children | 1102c90b157d |
| files | Cargo.toml README.md |
| diffstat | 2 files changed, 14 insertions(+), 3 deletions(-) [+] |
line wrap: on
line diff
--- a/Cargo.toml Thu Oct 12 21:05:36 2017 +0200 +++ b/Cargo.toml Thu Oct 12 22:03:59 2017 +0200 @@ -2,7 +2,12 @@ name = "leveldb_rs" version = "0.1.0" authors = ["Lewin Bormann <lbo@spheniscida.de>"] -publish = false +description = "A compatible re-implementation of LevelDB in Rust" +repository = "https://bitbucket.org/dermesser/leveldb-rs" +readme = "README.md" +keywords = ["LevelDB", "key/value", "database", "SSTable", "Google"] +license = "MIT" +publish = true [dependencies] crc = "1.2"
--- a/README.md Thu Oct 12 21:05:36 2017 +0200 +++ b/README.md Thu Oct 12 22:03:59 2017 +0200 @@ -6,6 +6,9 @@ algorithm translated 1:1, and class (struct) and method names are similar or the same. +**NOTE: Do not use this for production purposes unless you thoroughly evaluated +it first, and are reasonably certain that no data loss occurs.** + ## Status * User-facing methods exist: Read/Write/Delete; snapshots; iteration @@ -15,8 +18,11 @@ takes 0.2-0.5 seconds. * Compatibility with the original: Compression is not implemented so far; this works as long as compression is disabled in the original. -* Performance is usually on par with the original, except in cases where the - original implementation makes use of multithreading (e.g., background compactions). +* Performance is decent; while usually not par with the original, due to multi-threading + in the original and language-inherent overhead (we are doing things the right way), + it will be enough for most use cases. +* Safe: While using many shared pointers, the implementation is generally safe. Many + places use asserts though, so you may see a crash -- in which case you should file a bug. ## Goals