Merge pull request #26 from thedodd/20-update-should-take-filter-doc

Closes #20.

The Model.update method has been updated a decent bit. It now takes an
optional filter doc & returns Result<Option<Self>>. See the changelog
and #20 for more details.

Author: thedodd

CHANGELOG.md

@@ -1,6 +1,11 @@
 changelog
 =========
 
+## 0.7
+Minimal changes per this release. The main issue being addressed here is [#20](https://github.com/thedodd/wither/issues/20). It is arguable that this is just a bug fix, but the interface for `Model.update` has changed enough with these updates that a minor version increment is merited.
+
+The main thing to observe here is that `Model.update` now takes an optional filter document. If none is given, then this method will create a filter doc and populate it with the model's ID. If a doc is given, then it will have the `_id` key set (potentially overwritten) to the model's ID. Lastly, this method now returns `Result<Option<Self>>` to reflect the fallible nature of the update with a filter.
+
 ## 0.6
 Wow! So much stuff here. `0.6` is a big step forward for the ergonomics & usability of the wither system. A custom derive has been introduced (`#[derive(Model)]`), and it is now the recommended way to use this system. This should greatly simplify the process of getting started. Overall this has been an awesome experience putting this together and delving into the custom derive system in Rust. Here are a few of the changes to highlight.
 - Use `#[derive(Model)]` to turn your struct into a wither `Model`.

CONTRIBUTING.md

@@ -37,5 +37,12 @@ For the compile tests, you will need to use nightly.
 
 ```bash
 # Run the compile tests.
-cargo +nightly test -p wither_derive --tests --lib
+cargo +nightly test -p wither_tests -p wither_derive --tests --lib
+```
+
+For doc tests, you will also need to use nightly.
+
+```bash
+# From the wither dir.
+cargo +nightly test --features docinclude --doc
 ```

Cargo.lock

@@ -62,7 +62,7 @@ fn main() {
     me.save(db.clone(), None);
 
     // Update user's email address.
-    me = me.update(db.clone(), doc!{"$set": doc!{"email": "[email protected]"}}, None).unwrap();
+    me.update(db.clone(), None, doc!{"$set": doc!{"email": "[email protected]"}}, None).unwrap();
 
     // Fetch all users.
     let all_users = User::find(db.clone(), None, None).unwrap();

README.md

@@ -9,7 +9,7 @@ license = "Apache-2.0"
 name = "wither"
 readme = "../README.md"
 repository = "https://github.com/thedodd/wither"
-version = "0.6.3"
+version = "0.7.0"
 
 [dependencies]
 chrono = "0.4"
@@ -20,7 +20,7 @@ serde_derive = "1"
 
 [dev-dependencies]
 lazy_static = "1"
-wither_derive = { version = "0.6", path = "../wither_derive" }
+wither_derive = { version = "0.7", path = "../wither_derive" }
 
 [features]
 docinclude = [] # Used only for activating `doc(include="...")` on nightly.

wither/Cargo.toml

@@ -6,7 +6,7 @@ All `Model` struct & field attributes are declared inside of `model(...)` attrib
 Deriving `Model` for your struct is straightforward.
 
 - Ensure that your struct has at least the following derivations: `#[derive(Model, Serialize, Deserialize)]`.
-- Ensure that you have a field named `id`, of type `Option<bson::oid::ObjectId>`, with the following serde attributes: `#[serde(rename="_id", skip_serializing_if="Option::is_none")]`.
+- Ensure that you have a field named `id`, of type `Option<ObjectId>`, with the following serde attributes: `#[serde(rename="_id", skip_serializing_if="Option::is_none")]`.
 
 For now, it seems logical to disallow customization of the PK. An argument could be made for allowing full customization of the PK for a MongoDB collection, but there really is no end-all reasoning for this argument which I am aware of. If you need to treat a different field as PK, then just add the needed index to the field, and you are good to go. More on indexing soon.
 
@@ -67,7 +67,7 @@ As you can see, everything is declared within `#[model(index(...))]` attributes.
 ##### index
 Everything related to an index declaration must be declared within these parens. If the field is using a serde `rename` attribute, this system will account for that and use the value of `rename` as the initial field name for the new index.
 
-##### type
+##### index_type
 This declares the type of index for the field which this attribute appears on, which will also be the first field of the generated index. The value must be one of the valid MongoDB index types:  `"asc"`, `"dsc"`, `"2d"`, `"2dsphere"`, `"geoHaystack"`, `"text"` & `"hashed"`.
 
 ##### with