* Added rewritten NSUrlSessionHandler that handles memory better
This is a rewrite of the ModernHttpClient version of NSUrlSessionHandler, it has better handling for memory that provides a more consistant memory footprint. It accomplishes this by using NSInputStream for requests, and reading and disposing directly from NSData instead of transitioning the NSData to a byte[] array.
* Try to fix build of PR #31
* [foundation] Restore compatibility with the new NSUrlSessionHandler
Mostly my comments in PR #174
* Add support for redirection [1]
* Add support for credentials [1]
* Add support for caching [2]
* Remove 2nd dictionary lookup in GetHeaderSeparator
* Avoid extraneous cast for credentialsToUse
PR 177 [3] adds tests that ensure no commits can remove, or change
default values, for handlers.
[1] breaking changes (feature, not API)
[2] breaking change (API removal)
[3] https://github.com/xamarin/xamarin-macios/pull/177
* [foundation] Restore compatibility with the new NSUrlSessionHandler
Mostly my comments in PR #174
* Add support for redirection [1]
* Add support for credentials [1]
* Add support for caching [2]
* Remove 2nd dictionary lookup in GetHeaderSeparator
* Avoid extraneous cast for credentialsToUse
PR 177 [3] adds tests that ensure no commits can remove, or change
default values, for handlers.
[1] breaking changes (feature, not API)
[2] breaking change (API removal)
[3] https://github.com/xamarin/xamarin-macios/pull/177
* Try to fix build of PR #31
* prevent DEADLOCK in UI code
* Added ConfigureAwait(false) to Task.Delay to prevent DEADLOCK when the stream is being awaited on the UI thread
* added a few more ConfigureAwait(false) statments that were missed on first pass
* Fix some small style issues.
* Set the default value of AllowAutoRedirect to true.
At startup we send a request to all the IP addresses we have,
so we must make sure to not get confused if we get responses
from more than one of them.
This fix also requires an updated mlaunch.
https://bugzilla.xamarin.com/show_bug.cgi?id=44568
The watchOS device has limited networking support; in particular
it does not allow inbound/output network connections using 'bind'
(kernel-level sandbox restrictions).
This means that we can't use BSD sockets to connect to the debugger
in the IDE on the desktop. Instead we create an http tunnel that
knows how to convert socket send/recv data into http requests on
both sides.
https://bugzilla.xamarin.com/show_bug.cgi?id=41554
The watchOS device has limited networking support; in particular
it does not allow inbound/output network connections using 'bind'
(kernel-level sandbox restrictions).
This means that we can't use BSD sockets to connect to the debugger
in the IDE on the desktop. Instead we create an http tunnel that
knows how to convert socket send/recv data into http requests on
both sides.
https://bugzilla.xamarin.com/show_bug.cgi?id=41554
* Bump [watch-]mono to master to get fix for #43658.
https://bugzilla.xamarin.com/show_bug.cgi?id=43658
* [mtouch/mmp] Fix build after breaking cecil update in mono.
Also use mono's cecil instead of our own cecil submodule for mtouch.
* Bump [watch-]mono to get compilation fixes after cecil bump in mono.
* Remove cecil submodule, we only use the one in mono now.
This fixes the following problem:
* xamarin-macios is cloned
* xamarin mode is enabled
* make world is executed
* The maccore repo will be cloned immediately
* make reset-versions will be executed, which will reset the maccore version.
If this brings a new version of maccore, maccore dependencies might be
out of date.
* make all will be executed, which may fail because some dependencies
are out of date.
Instead check if the maccore directory exists.
This prevents an issue where if the stamp file did not exist and
the maccore directory did, we'd reset the maccore directory even
for make targets that shouldn't change anything (such as 'check-versions').
* [XM] Teach XM's mmp tool to handle read only assemblies/native libs
- https://bugzilla.xamarin.com/show_bug.cgi?id=41037
- mmp should also promote any install_name_tool errors to "real" errors
* Bump maccore
[Xamarin.Hosting] Set a default simulator if nothing is provided to mlaunch. Fix#41083 part 3
The string to represent a specific simulator is not something you can
easily remember but it's often not very important.
This will start the most basic simulator, by default, instead of throwing
an NRE that does not mention a (now un-) required parameter is missing.
> Failed to launch the simulator: Object reference not set to an instance of an object
> error MT1008: Failed to launch the simulator: Object reference not set to an instance of an object
https://bugzilla.xamarin.com/show_bug.cgi?id=41083
Also bump maccore to get new test.
commit xamarin/maccore@eb9c34d875
Author: Rolf Bjarne Kvinge <rolf@xamarin.com>
Date: Fri May 13 12:46:04 2016 +0200
[monotouch-test] Add test to ensure mtouch properly links with CoreAudioKit when not using simlauncher.
It is apparently common to have references to assemblies that can't
be resolved when building apps with the linker disabled, so we need
to support that.
So add a custom reference loading step that only shows a warning
for unresolvable assemblies.
Also bump maccore to add a test.
commit xamarin/maccore@190c67d855
Author: Rolf Bjarne Kvinge <rolf@xamarin.com>
Date: Tue May 10 18:14:04 2016 +0200
[tests] Add test case for bug #40786.
https://bugzilla.xamarin.com/show_bug.cgi?id=40786https://bugzilla.xamarin.com/show_bug.cgi?id=40786
To get a version of mdtool that works without any Xamarin.Mac
licenses.
Also bump maccore to run the mmp regression tests now that
we have a working Xamarin Studio.
commit xamarin/maccore@9a5e6f02f3
Author: Rolf Bjarne Kvinge <rolf@xamarin.com>
Date: Tue May 10 18:20:48 2016 +0200
[tests] There is now a released version of Xamarin Studio whose mdtool doesn't care about the XM license, so we can run the mmp regression tests again.
commit xamarin/maccore@8cc7ade11c
Author: Rolf Bjarne Kvinge <rolf@xamarin.com>
Date: Tue May 10 12:16:47 2016 +0200
[xharness] Find and parse config files in xamarin-macios as well.
Should make the watchos tests run again on wrench.
It's based on our work, with a few extra commits. That should reduce a
bit the checkout time and ensure more consistent results.
Also bump Touch.Unit since that requires a small API change.
Fix ApiSelectorTest for NSImage initByReferencingFile:
Commit 8b400722fb introduced
a new InitByReferencingFile internal method bound to initByReferencingFile:
Therefore the mac don't link tests were complaining because that selector
was used on a method and not a constructor.