will try and see #1
Open
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
1st shims for the LibSsl library.
…onse stream Discovered during testing with DEBUG version of CoreCLR running internal ToF tests. The design of the WinHttpRequestState is such that it stays alive due to a GCHandle.Alloc(this) in its constructor. This keeps the object alive during any of the native WinHTTP callbacks. However, the object was being prematurely disposed during the Dispose() of the WinHttpResponseStream. Instead, it should have been explicitly disposed during the final HANDLE_CLOSING callback from WinHTTP when the request handle was disposed. The early disposing of the WinHttpRequestState object was doing a GCHandle.Free() of the state object. This later caused an Assert() in the DEBUG version of the CoreCLR when doing GCHandle.FromIntPtr() indicating that the handle was already deallocated.
When using standard input redirection on Unix with System.Diagnostics.Process, the code would emit a BOM. This screws up many unix tools which don't understand the BOM.
Address PR feedback for System.Net.Sockets.
There are a number of clean up tasks that came out of the first 2 rounds of libssl shim work. 1. Add some Debug.Asserts when reading/writing to buffers. 2. Split the SSL extern methods into a separate class from the Interop.Crypto class. This allows the minimum amount of intialization to occur as needed. It also better keeps libssl separate from libcrypto shims. 3. Use a configure.cmake and #cmakedefine01 feature in our CMake file to create defines. This is consistent with CoreCLR and the other of CoreFX shims.
Clean up current libssl shim work.
Enable System.Console.ReadKey
Fix assert and leak of WinHttpRequestState object
Update license attribution
Stop emitting a BOM into stdin when using System.Diagnostics.Process
Enabling the interpreter in all builds.
Made Encode methods virtual to make it easier to write mocks
Fixing interpreter for NullReferenceException behavior in Unbox.
If it does not exist, `typedef in_addr in_pktinfo`.
I also moved the SSL SafeHandle classes out of the Interop class to be better aligned with the rest of the SafeHandles.
Instead of having managed code duplicate the SSL_CIPHER struct, moving all the code that uses the cipher struct into a native shim method. That also pulled a lot of algorithm constants to be defined in native instead of managed.
When calculating generations I wasn't considering transient dependencies. I was assuming that if a dependency wasn't tracked by generations it wouldn't impact the generation of the depender. That was incorrect since every assembly effectively has a generation. This updates the algorithm to walk the entire dependency graph when calculating the generation I also wasn't validating support for targeting generations: PCLs consuming the packages. I was only testing for specific platforms. I added validation for generation monikers and derived the support matrix for these from the specific platforms. This uncovered a handful of bugs which I have fixed. [tfs-changeset: 1540300]
Add mvid and metadata token API to S.R.TypeExtensions
Fixes #2542.
native: Check if in_pktinfo exists
OpenSSL doesn't have a direct encoder for KeyUsage. Utilizing its encoder would require one of
* calling ASN1_BIT_STRING_set_bit for each bit
* Presuming its data encoding and building it directly
* The complicated part is setting the "number of unused bits" value
* Making a native function which does the bit-testing and calls ASN1_BIT_STRING_set_bit itself
* But what order should the bits be assigned in?
So, instead just add support for encoding a "NamedBitList" bit string to the fledgling DerEncoder.
Add a few Marshal.*HGlobal tests
Trim response headers in CurlHandler
Fixes #4200
Adding a native shim for curl_version_info and removing the managed constants that are no longer needed.
This removes all direct dependencies from managed code to native libc functions. Fix #4106
…to getdomainname and looking for a warning. Fixing unused parameters build error on OSX. Update for PR feedback. - Collapse the 3 HAVE_INOTIFY defines into a single define.
…s and fail on errors, instead of checking for -Wshorten in the build output.
Add shim for libcurl version info.
Adding Compile(bool) overload to ref file for System.Linq.Expressions.
Make RuntimeInformation work on Net 4.5
Add shims for the rest of the libc native calls.
Make Unix HttpClient use X509Chain instead of the default chain validator
System.Net.WebHeaderCollection already had a SubstringTrim helper that's used instead of Substring(...).Trim() to avoid first creating the substring and then trimming it, instead just offsetting the startIndex and length appropriately to do the Substring once. I've factored this helper out into Common, added unit tests, cleaned it up a bit, and used it in several more places in System.Net.
Support Correct exception for windows authentication failure on Xplat
Implement EncodeX509KeyUsageExtension for Unix
Replace Substring(...).Trim() with SubstringTrim(...)
Today on Windows if no debugger is attached, Debug.Assert fail fasts, which makes diagnosing difficult in a continuous integration environment; it also causes any other tests in the same process to be abandoned. This commit changes Assert to instead write to stderr and throw an exception in such situations. The exception helps to ensure that code containing the assert is alerted to the failure, and the stderr output helps to ensure that it's noticed even if the exception is eaten.
…dding a leg to the PR builds that will fail the leg if the native code violates our code style; also, running the formatting script
…ements Cross plat dev workflow improvements.
Change Debug.Assert on Windows when no debugger attached
The lack of these statements was breaking the TFS build. Fixes #4143.
If you pipe from a managed app to another app that ends (or doesn't exist to begin with), attempting to write to it via Console will throw an exception due to the pipe getting broken. This commit just changes Console's Write to ignore EPIPE (broken pipe) errors. For now I still have it throw an other failures, but we could choose to add additional ignored errno values to the list, or simply not throw for anything.
…derr Revert "Change Debug.Assert on Windows when no debugger attached"
Add some missing using statements.
Ignore EPIPE errors on Unix when writing to Console
Adding a PR leg to run our style check and bringing native code into style compliance
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.