Fix a few typos in the docs

docs/discussions.rst

@@ -133,7 +133,7 @@ This says we can safely expect the file to contain ``abcd``.
 
 Note that Procrastinate will use PostgreSQL to search the jobs table for suitable jobs.
 Even if the database contains a high proportion of locked jobs, this will barely affect
-Procrastinates's capacity to quickly find the free jobs.
+Procrastinate's capacity to quickly find the free jobs.
 
 A good string identifier for the lock is a string identifier of the shared resource,
 UUIDs are well suited for this. If multiple resources are implicated, a combination of
@@ -331,8 +331,8 @@ have to be maintained in parallel.
 How stable is Procrastinate?
 ----------------------------
 
-More and more stable. We've started mentionning the project around, and it's
-now used in production in real projects.
+More and more stable. We've started mentioning the project around, and it's now
+used in production in real projects.
 
 That being said, we'd like to develop real monitoring tools before we call this
 really ready for production.

docs/howto/blueprints.rst

@@ -29,7 +29,7 @@ Firstly, create a Blueprint instance and then create some tasks::
     def mytask(argument, other_argument):
         ...
 
-In your projcet register the blueprint with the `App` after you have created it::
+In your project register the blueprint with the `App` after you have created it::
 
     from procrastinate import AiopgConnector, App
 

docs/howto/defer.rst

@@ -36,7 +36,7 @@ Create a job pattern, launch multiple jobs
 
 ::
 
-    pattern = my_task.configure(task_kwagrs={"a": 1})
+    pattern = my_task.configure(task_kwargs={"a": 1})
 
     pattern.defer(b=2)
     pattern.defer(b=3)