Age | Commit message (Collapse) | Author |
|
|
|
|
|
Now that we can automatically generate the reference typeforwards from
contracts I can eliminate the ifdef'ing I added with 45b4c71bff924245d8962035bd751d65c45bf31d.
|
|
|
|
|
|
Delete a few unnecessary files from Common
|
|
|
|
* XmlTextAttribute test
* Use generic Assert.Throws method.
|
|
We weren't passing the right setting to CMAKE.
Also fixes issue with current working directory getting changed during the build. Run.exe currently must be run
with the current working directory set to the repo root. One part of this was the config.json file- which I've explicitly specified now.
The other (as of yet unfixable) problem is that we specify the location of the binclash logger as a relative path. Afaik there currently is no way to make this happen with run.cmd without making this logger an explicit (as opposed to default) option. Could possibly extend run.exe's parsing to allow specifying environment variables or reference to other parameters...
I haven't nailed down who is changing the working directory- while it appears to be a 2017 specific problem I can't guarantee that it isn't some other variable.
|
|
|
|
Started copying netstandars shims for netfx under ref\netfx
|
|
Root cause WebClient test failures on NETFX
|
|
Update defaults of SignedXml to use SHA256
|
|
* Disable all tests for M.W32.Registry on uap
These tests all need to be disabled because the
implementation does nothing but throw PlatformNotSupported
on uap
* Don't even try to run registry tests on uap
|
|
Every test using PlatformDetection for anything was
failing due to a NullReferenceException in PlatformDetection's
type initializer (due to Environment.GetEnvironmentVariable()
returning null for everything on UWP.)
Turn PlatformDetection's properties into non-latched
properties to avoid this unnecessary failure multiplication.
(I counted over 500 uses of PlatformDetection across the
corefx test bed.)
|
|
|
|
|
|
Finished investigation failures on .NET Framework for WebClient.
Added some new tests to clarify the open issues.
Added new bugs #18674 and #18680.
Fixes #17882.
|
|
|
|
|
|
|
|
Validate the DER components of user-provided data
|
|
Move XmlSerializer perf tests into own project with correct naming
|
|
Disable Xml quirk test on NetFx
|
|
* Issue #18572 Add ReadonlySpan CopyTo test
* Issue #18572 Fixed missed use of ReadOnlySpan
|
|
Activity: Precise for StartTime and duration on .NET Desktop
|
|
|
|
|
|
|
|
Start Activity and inject headers in HttpHandlerDiagnosticListener on net46
|
|
Fix System.Net.NameResolution tests on desktop
|
|
Fix/disable several System.Net.Security tests on desktop
|
|
A bunch of the unit tests try to validate that sockets were correctly initialized. They do so with fakes that intercept the calls. But on desktop, the real methods were being used, bypassing that interception. I've fixed it by making the unit tests on desktop compile like core so as to use the fakes.
|
|
- SslStream_StreamToStream_WriteAsync_ReadAsync_Pending_Success was hanging / timing out on desktop because it relied on Read/WriteAsync not going through the base Stream's Read/WriteAsync, which serializes the delegated calls to Begin/EndRead/Write. Desktop's implementation calls directly to Begin/EndRead/Write whereas in core (by design) we call Read/WriteAsync (which will in turn use Begin/EndRead/Write if the Async method wasn't overridden), but the VirtualNetworkStream wasn't overriding Begin/EndReadWrite to avoid the serialization, and since the test requires non-serialization, it hung.
- SslStream_StreamToStream_FlushAsync_Propagated and NegotiateStream_StreamToStream_FlushAsync_Propagated both rely on FlushAsync being overridden, which it's not in desktop.
- NegotiateStream_StreamToStream_Authentication_EmptyCredentials_Fails fails due to a known behavioral difference from desktop for which there's already an open issue.
|
|
The WinHTTP implementation sends Connection: "keep-alive", the desktop implementation does not. Just change the test to validate whether Connection: "close" is sent or not.
|
|
No more "Multidimensional arrays of rank 1 are not supported..."
|
|
Modify System.Numerics.Vectors unittest
|
|
|
|
Add support for UseDefaultCredentials to CurlHandler
|
|
(#18622)
* Add test for failfast in event of corrupted state exception. Add test for Assembly.Load by AssemblyName when AssemblyName.CodeBase property is set
* update
* update
|
|
|
|
* Fixed basic XmlMembersMapping tests.
* Add test for HasWrapperElement=true.
* Added tests for types having SpecifiedField.
Fixed wrting types having SpecifiedField.
* Enabled MembersMapping with SpecifiedField.
Refactor ReflectionXmlSerializationReader.
* Enabled using XmlMembersMapping with arrays.
* Removed Unneeded Param: parentMapping.
* Removed unused param: arrayName
* Renamed variable: parent to container.
* Modified the test helper: RoundTripWithXmlMembersMapping
* Support Xmlns member.
* Support XmlMembersMapping with XmlAnyAttribute.
* Test XmlMembersMapping_Member_With_XmlAnyAttribute_Specified*
* Support XmlMembersMapping with ChoiceIdentifier.
* Make XmlSerializer.FromMappings Use Reflection for Soap.
* Support Soap XmlMembersMapping.
* Fix rebase conflict.
* XmlMembersMapping_Soap_HasWrapperElement
* XmlMembersMapping_Soap_MemberSpecified_True
* XmlMembersMapping Soap HasWrapperElement Validate
* Added a few tests.
XmlMembersMapping_MultipleMembers
XmlMembersMapping_MultipleMembers_IsReturnValue
XmlMembersMapping_Soap_MultipleMembers_IsReturnValue
* Replaced two Debug.Assert with exceptions.
* XmlMembersMapping_MultipleMembers_XmlAnyElement
* Some cleanup.
* Fixed some PNSE, and added tests.
* Xml_Soap_MyCollection
* Fixed a bug with ReadOnly collection using Soap.
* Add tests for nullables.
* Clean up. Fix styles.
* Fixed merge conflicts
* Changes based on mconnew's feedback.
|
|
The parameters value for the PublicKey and the generator's SignatureAlgorithm
both are supposed to be DER-encoded; so validate that claim.
For the most part this isn't necessary, since rehydrating the signed cert is going
to fail to parse the certificate anyways; but being defensive is good.
|
|
|
|
|
|
* Added failing test for fAbortOnError not cleared
Issue #17396
* Ensure fAbortOnError is clear when opening port
|
|
|
|
Fix UnixFileSystem.MoveFile on macOS when just changing file casing
|
|
This property is causing many test failures. The wrong assembly was
being included in the ILC input folder
|
|
* Change Unix creation time to min(ctime, mtime)
|