* Updated dependencies.
* Reverted Microsoft.Bcl.AsyncInterfaces back to version 1.0.0.
Also added a note so we no longer try to upgrade it.
* Reverted back to Moq.
---------
Co-authored-by: Wojciech Nagórski <wojtpl2@gmail.com>
* Added guard clauses to various timeouts to ensure they don't exceed an Int32 in milliseconds.
* Fixed guard clauses.
* Updated build tags.
* Added guard clauses to various timeouts to ensure they don't exceed an Int32 in milliseconds.
* Fixed tests.
* Added additional tests.
* Replaced NoWarn with .editorconfig setting
* Fixed references to parameter names.
* Restore Write/Flush tests
* Restore write buffer to ShellStream and painfully fix the mocks
---------
Co-authored-by: Wojciech Nagórski <wojtpl2@gmail.com>
* Fix a few issues with ShellStream
The main change is to replace the Queue<byte> with a byte[] and a couple of
variables which index the start and end of the data. The remainder is mainly
slightly more careful locking semantics.
It also implements Expect(string) separately so that it can work on the bytes
and skip a lot of encoding work (this is where I wish ShellStream derived from
StreamReader).
One possibly contentious point: in fixing the Write behaviour I chose to
remove the "outgoing" buffer and immediately send the data across the channel.
Write(string) and WriteLine(string) were already doing this, and I felt it
was better to change Write(byte[]) to match rather than changing the string
methods.
* Integrate expectSize (as "windowSize" parameter)
* Rename "windowSize" to "lookback"
---------
Co-authored-by: Wojciech Nagórski <wojtpl2@gmail.com>
* Support ETM (Encrypt-then-MAC) variants for HMAC
* Support ETM (Encrypt-then-MAC) variants for HMAC
* Remove `ETM` property from `HashInfo`
* Add support for HmacMd5Etm, HmacMd5_96_Etm, HmacSha1Etm and HmacSha1_96_Etm
Explicitly specify etm even if false
* Add Encrypt-then-MAC variants to README.md
* Add back empty span to prevent auto link
* Store ETM in `HashInfo`;
Change `HMAC Create[...]Hash()` to `HashAlgorithm Create[...]Hash(out bool isEncryptThenMAC)`
* Making all unit tests pass locally.
Excluded MD5 tests on net462 because I get System.InvalidOperationException: 'This implementation is not part of the Windows Platform FIPS validated cryptographic algorithms.'
SshdConfig: do not throw for "Include", just do nothing.
Modified failing dos2unix parameters in Dockerfile.TestServer.
Forceing LF line ending for key files used by integration tests, otherwise using them causes error.
SftpClientTest.Test_Sftp_Multiple_Async_Upload_And_Download_10Files_5MB_Each times out for maxFiles=10, decreasing to 2 to make the test pass.
* Added SshCommand.InputStream.
* Added an integration test for SshCommand.InputStream.
* Reverting changes made to unit tests unrelated to this PR.
* Moved ChannelInputStream's EOF sending from Write to Dispose. Replace SshCommand.InputStream with CreateInputStream to emphasise that a (disposable) resource is created here. EndExecute also closes the _inputStream to make sure that EOF is sent (in case the user forgot to dispose the input stream). Added more unit tests: sending the input one byte at a time, not disposing the input stream, calling CreateInputStream before BeginExecute or AfterEndExecute throws exception.
* Fixing review comments.
* Fix build error after #1286
---------
Co-authored-by: ferenc.vizkeleti <ferenc.vizkeleti@innopay.hu>
Co-authored-by: Wojciech Nagórski <wojtpl2@gmail.com>
Co-authored-by: Rob Hague <rob.hague00@gmail.com>
* Allow the integration tests to run on .NET Framework
* Update appveyor.yml
* Update appveyor.yml
---------
Co-authored-by: Wojciech Nagórski <wojtpl2@gmail.com>
- Change _connectAndLazySemaphoreInitLock to a SemaphoreSlim and use it in
ConnectAsync.
- Rename it to _connectLock and only use it for connecting. Replace its
other usages (on SessionSemaphore and NextChannelNumber) with Interlocked
operations.
- Remove AuthenticationConnection semaphore. This static member placed a
process-wide limit on the number of connections an application can make.
I agree with the argument in
https://github.com/sshnet/SSH.NET/issues/409#issuecomment-457415542
(and in several other issues/PRs) that this should not be something
that the library attempts to control.
The last change broke a few tests which do things like making 100 connections.
I was tempted to delete these tests as I don't think they have much value, but
instead I just limited their concurrency.
Co-authored-by: Wojciech Nagórski <wojtpl2@gmail.com>
This makes it easier to reason about Key instances in e.g.
DigitalSignature implementations, because we know that the
Key is initialised with its data and will not change.
Co-authored-by: Wojciech Nagórski <wojtpl2@gmail.com>
* Exludes the test projects from code coverage reports
* Removes MessageAttribute in favor of properties on Message class
* Benchmark for removal of MessageAttribute
This paves the way for asynchronous synchronisation via WaitAsync
(and eliminates a timing test which sometimes fails in CI)
Co-authored-by: Wojciech Nagórski <wojtpl2@gmail.com>
* Updates .gitignore to exclude test results
* Adds .editorconfig for xml, csproj, and props files
Formats xml, csproj, and props files
* Hides stylecop.json from Solution Explorer projects
* 🐛 AsyncResult contains invalid value
- AsyncResult should contain invalid value immediately after async operation is marked as completed
- there was race condition problems with callback method which is invoked on different thread so updating of value is done without any synchronization. So in some cases async operation is marked as completed but async result value is not yet updated and contains invalid value
* Revert test
---------
Co-authored-by: Miroslav Pokorný <m.pokorny@quadient.com>
* Add support for .NET 8.0
* Fix CA1512: Use 'ArgumentOutOfRangeException.ThrowIfNegative' instead of explicitly throwing a new exception instance (https://learn.microsoft.com/dotnet/fundamentals/code-analysis/quality-rules/ca1512)
* Use `ThrowIfGreaterThan` instead of `ThrowIfNegative`+calculation, to supply the correct parameter name.
* .NET 8 is faster :-)
* Use the explicit version, in case newer version SDK breaks our tests
* Add FEATURE_AES_CSP to use hardware-accelerated AesCryptoServiceProvider
Reduces CPU usage dramatically, allowing more performance on slower machines
* Restructure, move most of the feature code to AesCipher.cs
Fix padding for non-AES blockciphers
Fix IV exception for non-AES blockciphers
* Fix the AES Padding
It looks like the legacy code doesn't correctly remove padding, so this code needs to do the same.
* fix rebase issues
restructure AES CSP code into its own class
* Minor fixes
* Rework based on suggestions
* Move all changes to AesCypher.cs, as per Rob-Hague suggestion
Remove FEATURE_AES_CSP conditional
Fix OFB CipherMode
* update AesCipherTest.cs generator
* Fix continuous session encrypt/decrypt
(preserve IV between calls when Padding is None)
* Reduce CTR memory usage in Net 6+
Small performance increase in CTR buffer mode
Cosmetic changes
* Factor out the implementations and re-add the existing constructor
* remove ctor; revert tests; remove unused _iv member
* Reorder Encryption cipher preference list
* Remove redundant AES tests
Add tests for stream cipher state preservation
* Refactor ArrayXOR()
* Add test for IV overflow
* Performance bump for AES CTR (thanks @robhague)
* fix merge conflict
* Move AesCipherMode enum to its own file
---------
Co-authored-by: Pedro Fonseca <pfonseca@qti.qualcomm.com>
Co-authored-by: Rob Hague <rob.hague00@gmail.com>
* Data lengths longer than stream position when data lengths are greater than int.maxvalue are ignored and do not throw an exception
* Removed unreachable test
* Do not try to load the data (just ignore it)
---------
Co-authored-by: Steve Evans <sevans1@oeconnection.com>
Co-authored-by: Rob Hague <rob.hague00@gmail.com>
* Fix https://github.com/sshnet/SSH.NET/issues/1191
* Expose `SshIdentificationReceived` event so that lib consumer can adjust based on server identification
* revert unrelated code style change
* revert OpenSSH 6.6 related tests
* revert ConnectionBase
* Add unit tests
* Rename to `ServerIdentificationReceived`
* rename
* Use System.Security.Cryptography in AesCipher
* don't set the IV - it is unused by ECB
* Dispose ciphers in Session if applicable
---------
Co-authored-by: Wojciech Nagórski <wojtpl2@gmail.com>
Change the data to be in 64-wide columns instead of 70, and add a newline
at the end of the file. ssh-keygen does not like it otherwise.
puttygen Key.OPENSSH.ECDSA.Encrypted.txt -O private-openssh-new -C Key.OPENSSH.ECDSA.Encrypted
puttygen Key.OPENSSH.ECDSA.txt -O private-openssh-new -C Key.OPENSSH.ECDSA
puttygen Key.OPENSSH.ECDSA384.Encrypted.txt -O private-openssh-new -C Key.OPENSSH.ECDSA384.Encrypted
puttygen Key.OPENSSH.ECDSA384.txt -O private-openssh-new -C Key.OPENSSH.ECDSA384
puttygen Key.OPENSSH.ECDSA521.Encrypted.txt -O private-openssh-new -C Key.OPENSSH.ECDSA521.Encrypted
puttygen Key.OPENSSH.ECDSA521.txt -O private-openssh-new -C Key.OPENSSH.ECDSA521
puttygen Key.OPENSSH.ED25519.Encrypted.txt -O private-openssh-new -C Key.OPENSSH.ED25519.Encrypted
puttygen Key.OPENSSH.ED25519.txt -O private-openssh-new -C Key.OPENSSH.ED25519
puttygen Key.OPENSSH.RSA.Encrypted.txt -O private-openssh-new -C Key.OPENSSH.RSA.Encrypted
puttygen Key.OPENSSH.RSA.txt -O private-openssh-new -C Key.OPENSSH.RSA
(the -C flag changes the comment, otherwise puttygen refuses to do anything)
* Increase test coverage for AesCipher
The tests were generated by a script which is also added for posterity.
The script works by running "openssl enc [...]" (via WSL) to generate the
expected encrypted values, and also verifies those values against the .NET
BCL implementation as an extra validation (it uncovered a difference in
CFB mode between the two relating to the feedback size).
* Fix OfbCipherMode
It was an exact copy of CfbCipherMode
* Fix analyzer errors in Renci.SshNet and Renci.SshNet.TestTools.OpenSSH.
Suppress all errors in unit tests and integration tests.
* Update unit tests now that we pass 'mode' as argument name when we throw ArgumentException.
* Remove stale comment and add unit tests for SshData.ReadBytes(int length).
* Remove unnessary suppression.
* Remove Visual Studio magic.
* Removed duplicate source file.
* Clarified that suppression hides a false positive.
* Remove suppressions for S2372.
* Update ReadExtensionPair() to return concrete dictionary.
* Move test projects to test folder.
Move global.json to root of repo.
Update solution items in solution.
* Move test projects to test folder.
Move global.json to root of repo.
Update solution items in solution.
Update appveyor configuration/
* Attempt to have appveyor use the correct .NET SDK.
* Update .NET SDK to version 7.0.402.
* Move Data folder below Renci.SshNet.Tests.
* Make csinst less chatty.
* Move Data folder directly below test folder as it's used by multiple test projects.
* Remove CS1591 nowarn from concrete test projects as this is already defined in the Directory.Build.props that is in the test folder.
* Fix integration test after moving test projects
---------
Co-authored-by: drieseng <gert.driesen@telenet.be>
Remove support for legacy / deprecated target frameworks while adding support for .NET 6.0 (and higher).
The supported target frameworks are now:
* .NETFramework 4.6.2 (and higher)
* .NET Standard 2.0
* .NET 6.0 (and higher)
Jean-Sebastien Carle