From what I understand failureMode=true is what other clients call failover=false in the sense that if failure mode is on, the client does not failover to other servers but marks the failed server as dead until it come back up. This is confusing.

In the short term this should be documented. In the long term a configuration like builder.setFailover(true) could be added that sets `failureMo

https://pypi.org/project/expiringdict/

Instead of ExpirableDictStorage implementation, replacing dict to expringdict looks simpler and reasonable

Docs can be updated through the normal SVN workflow (for users with a php.net account) or through a PR to https://github.com/php/doc-en

At the time of writing,

• Some functions, such as Memcache::append, aren't documented.
• Many functions have incorrect return types - Memcache::get can now return objects, booleans, etc for string keys

Related to #64

Why do you not state in the documentation that strings in don;t equals strings out? It appears if set a string then get it I get a buffer instead?

We call num.to_string().as_bytes() to serialize and deserialize numbers when setting them in memcache server. Instead, we could use byteorder crate's read_u16, read_u64 etc. and write out the bytes to the server with a specific endianness(big/little) and use the same when deserializing.
deserialization:
https://github.com/aisk/rust-memcache/blob/faed4e117b8108dbcae1901c891844009229190e/s

A few things, in no particular order:

• 1. we should investigate Azure Pipelines: Windows, macOS, Linux, free for OSS
• 2. we should kill all 32-bit builds because it's 2019 and they just waste time
• 3. we should check formatting and not just regular tests
• 4. might be worth switching to cargo-make to ⭐ synergize ⭐with the rest of the NF crates