- #23 part 4
- depends on aspnet/HttpAbstractions@8c120a0
nits:
- correct name of a field in `AntiforgerySerializationContext`
- avoid allocations when returning an `AntiforgerySerializationContext` in (unlikely) case `Stream` is unused
- name literal `int` parameters
- rewrite existing methods in terms of the new ones
- don't allocate multiple 0-length arrays
nits:
- clarify a couple of doc comments e.g. using `<paramref/>`
- move an error message into a resource
- pass parameter names into new resource
- rename parameters for consistency e.g. `inputLength` -> `count`
- name literal `int` parameters
- more `var`
- This means connections become untracked sooner than before and not all blocks will
necessarily be returned.
- The assertion in the MemoryPoolBlock2 finalizer was weakened because FilteredStreamAdapter
will continue to use blocks after libuv stops tracking the associated connection.
- Make 100% sure we don't accept new connections after we dispose the listen socket by using a flag.
- Add a (currently unused) AllowStop method to KestrelThread. This is meant to be called from
listeners when we stop accepting new connections, but needs investigation to prevent flakiness.
- Even when safe handles are disposed explicitly, ReleaseHandle is sometimes
called on another thread which breaks uv_close.
- Ensure we close the UvAsyncHandle the uv loop so that the second call
to uv_run always completes without a timeout/Thread.Abort.
- Re-enable some tests. Add skip conditions for those that aren't passing.
- Make the time given for requests to complete gracefully configurable.
- Complete all async reads so calling code can re-check whether to stop
request processing and exit if in between requests.
- Don't wait for a FIN from the client since some browsers (e.g. IE & Chrome)
will take a long time to send one.
- Ensure all ConnectionFilters complete before the memory pool is disposed.
- Ensure blocks get returned even when a ConnectionFilter produces a failed read
Also correct tests to avoid warnings and to ensure scenarios are actually tested
- tests previously went `async` without waiting for completion
nit: add parameter `null` checks
This code was popping up everywhere this method is called. Seems bad to
duplicate it. Really what the caller wants to know is 'is the request
valid or a potential CSRF exploit?'. This gets the API closer to that.
Before moving to dotnet dnx on OSx on Mono would load libuv.dll into the process and Kestrel would use DllImport with __Internal to import functions from the native libraries loaded into the Mono process. With the move to dotnet nothing is preloading libuv to the process so Kestrel must no longer use DllImports with __Internal but the regular ones. Mono seems to load native libraries from the app folder so, the user won't have to install libuv on their own since libuv.dylib is published with the app.
Chris R
Cesar Blum Silveira
Pranav K
Doug Bunting
Ajay Bhargav Baaskaran
Stephen Halter
Nate McMaster
ryanbrandenburg
Ben Adams
John Luo
Victor Hurdugaci
Kiran Challa
David Fowler
N. Taylor Mullen
SteveSandersonMS
Ryan Nowak
Brennan
moozzyk
Kristian Hellang