This change optimizes allocations by RouteValueDictionary based on usage.
First, implement a struct Enumerator, and expose the concrete RVD type
from all extensibility points. We wanted to try and decouple this code
from RVD originally and use IDictionary everywhere. After doing that we've
found that it allocates an unacceptable number of enumerators.
Secondly, optimize copies of RVD for the case where you're copying an RVC
to another (common case). When doing this we can copy the count to get the
right capacity, and copy the entries without allocating an enumerator.
Lastly, optimize RVD for the case where it's a wrapper around a poco
object. We 'upgrade' to a writable full dictionary if you try to write to
it, or call one of a number of APIs that are uncommonly used. We could
produce optimized versions of things like `Keys` and `CopyTo` if necessary
in the future.
This change removes the call to string.Split and a few substrings, and
replaces it with a tokenizer API. The tokenizer isn't really optimized
right now for compute - it should probably be an iterator- but it's a
significant improvement on what we're doing.
This change makes the allocation of DataTokens and Values on RouteData
lazy, and elides copies when copying an 'empty' RouteData.
In our current architecture this change will eliminiate 2 * (N + 1)
dictionary allocations/copies per request, where N is the number of routes
processed. In a large system with lots of attribute routes, this number
could be very significant.
For a small MVC site (ModelBinding, Validation, Views) with one route, it
still shows a modest reduction of dictionary allocations without adding
much complexity.
- Removed existing logger scopes as we want to minimize the number of scopes being created.
- Cleaned up tests related to removal of scopes.
- Added new log statements.
- Removed old logger structure base implementation and related tests. Added new tests also.
This change adds a feature needed for aspnet/Mvc#302
There's a new option in routing that allows link-generation to proceed
when the route values cannot be validated. The key scenario for this is
during development of an MVC site. Routing will refuse to generate a link
to actions which don't exists, this is a breaking change from the MVC5
behavior. Setting UseBestEffortLinkGeneration will allow routing to return
a value even when we can't match the action.
This option will remain off by default - setting this to on will impact
link-generation in a bunch of scenarios involving areas where we've
improved the logic for MVC6. If you're considering leaving this on outside
of development scenarios, then make sure to be as explicit with route values
as possible (don't rely on ambient values).
Functional tests to follow up in the MVC repo.
Added OptionalRouteConstraint class to take care of optional inline parameter. It will create the OptionalRouteConstraint for a inline parameter that is optional with real constraint on the parameter as inner constraint of OptionalRouteConstraint.
The properties on TemplateRoute for DataTokens and Defaults are now
readonly. This prevents modifying these collections in a way that
invalidates cached data, or violates thread-safety.
To do the same for constraints, this change includes a substantial refactor
of how we realize inline constraints, and moves the constraint resolver
out of the parsing phase.
This allow creates a builder for the constraint map, that will make it
easier to implement features like optional constraints, and is reusable
for anyone building their own type of routing system.
This is the routing part of the fix. MVC will be updated as well
(attribute routing).
As the graph of routers is traversed, routers add themselves to the
current 'path', which unwinds on a failed path.
This mechanism is opt-in. Whoever adds something needs to remove it as
part of cleanup. If a router in the tree doesn't interact with the
.Routers property, then there are no consequences for those that do.
Additionally, fixing #116 as part of the same change. This means that we
create a nested 'RouteData' and then restore it on the way out. This is
simpler than just dealing with the .Routers property in isolation.
Changed the implementation of route template to merge the existing route data
with the values obtained from parsing the request path with the given template.
Restored original route data values in case the route template data does not match.
This change makes RouteValueDictionary a full IDictionary implementation
instead of a subclass of Dictionary.
Followed the patterns used in the old implementation, namely preserving
the struct-returning behavior of Keys/Values/GetEnumerator.
This change adds tests and makes the behavior consistent with legacy MVC
as far as what values are visible in constraints.
This is important because it allows constraints to make decisions based on
whether or not a value is present even if it's not in the template. This
is similar to the behavior of WebAPI link generation or Area link
generation in MVC 5 - but without hardcoding.
This is the routing part of these changes, and just the breaking changes
parts.
Follow-ups will add:
- DataTokens
- Tracking the logical stack of routers
Moving out Default Handler and IInlineConstraintResolver from RouteCollection.
These are moved to a new interface IRouteBuilder, (instead of an IInlineConstraintResolver, it takes in an IServiceProvider).
This means all RouteCollectionExtensions are moved to RouteBuilderExtensions.
For link generation to areas, we need to provide the set of values that
the route could potentially provide.
Basically if we know what action we want to reach, we want to know whether
or not a given route could hit that action before giving it the OK to
generate a link.
For instance a route like '{controller}' couldn't hit an action like
'HomeController:DoACoolThing', since it can never provide a value for
'action'. This makes it possible for WebFX to make the right decision
without changing the behavior of any of the routing constructs. This also
has the side-effect of removing a class of order dependencies in routing
that cause bad links to be generated.
We'll need to access the accepted values to do proper link generation, so
separating this process out into 2 parts.
Also moving defaults into the TemplateBinder because they are conceptually
part of the route, not part of the request. I'll do the same for
TemplateMatcher soon, but it's a big change and worth separating.
Ryan Nowak
Pranav K
Anthony van der Hoorn
Ajay Bhargav Baaskaran
N. Taylor Mullen
Kiran Challa
ianhong
Praburaj
Brennan
Mugdha Kulkarni
Doug Bunting
SonjaKhan
jacalvar
Stephen Halter
Max Toro
harshgMSFT
Ben Brown
Andrew Peters
Yishai Galatzer