Uncommented the sources and update some mistakes after following the
sample provided by Apple. Initially tests were going to be added but
they resulted to be to flacky and would make the CI red too often to be
adding value.
Porting the sample will ensure that it works are the bindings are not
broken.
The DSymUtil tool not only generates the debug symbol files but also modifies the executable file. Marking that property as Output (and changing it to ITaskItem type) makes Visual Studio on Windows aware of that change. Under certain scenarios this was making the build on VS produce an app bundle that was not fully signed on incremental builds. For instance, the DSymUtil task was run for a framework on an incremental build, but as the executable file of that framework was not modified on Windows the inputs/outputs check for CodesignFrameworks did not fail so that target was skipped. This led to a failure on the CodesignVerify target.
Partial fix for https://developercommunity.visualstudio.com/content/problem/729766/codedesign-exited-with-code-1.html
state.
Some of the tests were creating connections when not needed, this
'apparetly' was leaving the internal state of the Networking API in a
bad state in which it would throw a SIGSEGV and will make the connection
of other tests fail. In this case, the Tls tests that use the Network
API.
Cleaning the tests and removing those badly managed connections ensured
that the SIGSEGV would not be thrown and got all the other tests to
work.
This is a blackbox, test now passes without problems.
Fixes: https://github.com/xamarin/maccore/issues/2048
* [Tests] Fix the Network tests.
The pattern used to Cancel then Dispose the connection is wrong and
ended up crashing the tests sometimes. It is a race condition and is
very rare, but makes the tests flaky. The changes ensures that we
Dispose the connection, which internally closes it.
Tests have been added for the NWConnection Close calls.
Also cleaned some on the CWL that were not needed.
* [runtime] Simplify Vision dlsym'ed functions
More code sharing - DRYer :)
It also remove complaints (from static analysis tools) that a `dlclose`
should be present. That's not really an issue here since it's reference
counted and won't be unloaded if you use the code (and such a call is
not added in the PR). It silence the warning since it's not a local
variable anymore.
We have only encountered a case in which we had to add a nested
class/enum as the return type of a property. This fix ensures that we
can work with 1 level nested classes. A more general situation with
nested/nested/nested/.. classes is not taken into account because:
1. Would complicate too much the code.
2. Is fixing problems we do not have AFAIK.
So I'm keeping the fix simple, as I said, we have never faced anything
deeper than one level.
and by friends I mean `mmp` and `btouch`
What does this do ?
1. Assume that output of `mtouch` (and other similar tools) is **always** of high importance. Why ?
- If not then it's not saved in the binary log (even if visible on the console/text logs).
- The logging of `mtouch` (and friends) is dynamic, based on a supplied verbosity level.
- If a verbosity level _anywhere_ then it's a clear sign that the developer wants that extra output (and that includes binary logs).
2. Assume the _global_ verbosity of `msbuild` from the console is just as valid/useful than the one from VSfM.
- CI/bots produce logs and they should be useful to diagnose build issues.
- Setting verbosity in several places is error-prone, which delay investigations and results.
- Running the same project, with the same `msbuild` verbosity, should be identical between IDE and console.
What does that mean ?
Using `msbuild /v:diag /bl:out.binlog` you get a small(er) binary log that has everything[1] you need to diagnose a Xamarin.iOS (or Mac) build. It's also identical to the output what VSfM produce (for the same `msbuild` verbosity level).
[1] we might need to review what we log if we're missing interesting stuff
References:
https://github.com/xamarin/xamarin-macios/issues/7035
* Bump mono to a hash with archives and use them.
New commits in mono/mono:
* mono/mono@6af4ae7635 [2019-06][ci] Add Xcode 11.2beta2 for XI/XM Mono SDK builds
Diff: 476d72b9e3..6af4ae7635
* Bump mono to get min iOS version fix.
New commits in mono/mono:
* mono/mono@3775d5ac0a [sdks] Bump min iOS version to 7.0.
Diff: 6af4ae7635..3775d5ac0a
* [xharness] Bump mtouch tests timeout to 3h, we have a couple of new PR bots which are old and slow.
The new PR bots are late 2012 mac minis, so quite slow.
`calloc` can return `null` and we're writing to the memory location
which would crash the process.
An `OutOfMemoryException` is the correct way to handle this (even if
will likely crash the process anyway).
If `old_gchandle` is not set (0) then the function returns early,
so can't be 0 after the `MONO_THREAD_ATTACH` meaning:
* the `else` branch is never executed;
* the `exception_gchandle` is never set to something other than 0
* ` if (exception_gchandle == 0) {` is always `true` and can be removed (diff looks bigger than it is because of indentation change)
= the call to `xamarin_process_managed_exception_gchandle` is unneeded
* Add xamarin_os_log function
See the comment in the function for an explanation
of why this wrapper function is required.
* Add Darwin/OSLog.cs
* Add xamarin_os_log to header
This ensures that the symbol will not be subject
to C++ name mangling, therefore breaking mmp.
With this change applied, OSLog works as expected.
* Resolve stylistic PR feedback
* Move OSLog into CoreFoundation namespace
This is where the NativeObject class lives, and it
also feels like a better fit for a low-level API
that is available on non-Mac platforms than the
macOS-only Darwin namespace.
It is the more common and optimized idiom to just substract two integer values rather than do conditional checks. This may yield better performance for the qsort function, improving a bit of the startup time in case of many items
* Use a better, non-overflowing version with bit twiddling hacks.
Credits to @mandel-macaque for the magic
* Only notarize CI builds; timeout after 90 min
* Update jenkins/Jenkinsfile
Co-Authored-By: Rolf Bjarne Kvinge <rolf@xamarin.com>
* Don't upload notarized pkgs if it fails
* Update jenkins/Jenkinsfile
Co-Authored-By: Manuel de la Pena <mandel@microsoft.com>
The block/delegate passed to NWTxtRecord.Apply is supposed to return a bool.
Not returning anything ends up with random behavior, which causes a test
failure on Catalina:
3) TestApply (MonoTouchFixtures.Network.NWTxtRecordTest.TestApply)
keycount
Expected: 4
But was: 1
at MonoTouchFixtures.Network.NWTxtRecordTest.TestApply () [0x000a3] in /Users/builder/jenkins/workspace/xamarin-macios-pr-builder/tests/monotouch-test/Network/NWTxtRecordTest.cs:134
at (wrapper managed-to-native) System.Reflection.RuntimeMethodInfo.InternalInvoke(System.Reflection.RuntimeMethodInfo,object,object[],System.Exception&)
at System.Reflection.RuntimeMethodInfo.Invoke (System.Object obj, System.Reflection.BindingFlags invokeAttr, System.Reflection.Binder binder, System.Object[] parameters, System.Globalization.CultureInfo culture) [0x0006a] in /Users/builder/jenkins/workspace/xamarin-macios-pr-builder/external/mono/mcs/class/corlib/System.Reflection/RuntimeMethodInfo.cs:395
I've fixed the existing API to return 'true' always in the block callback, to
get consistent behavior, and in addition I've added a new overload that takes
a delegate that returns a bool, which allows the consumer of the API to decide
what to return.
Fixes https://github.com/xamarin/maccore/issues/2036.
There are a number of namespaces that do not support the API and can be
completly skipped. Also some of the classes in those namspaces are
expensive to create and the test is not providing any useful
information.
Fixes: https://github.com/xamarin/maccore/issues/2038