- 04 12月, 2015 13 次提交
-
-
由 John Hamby 提交于
-
由 John Hamby 提交于
-
由 Dustin Campbell 提交于
Ensure that an EnvDTE.CodeParameter.Name returns the correct value for parameters defined in metadata
-
由 John Hamby 提交于
-
由 Kevin Halverson 提交于
Fix assert failure for FileThatCannotBeDecoded under default codepage 936
-
由 Dustin Campbell 提交于
-
由 Paul Chen 提交于
-
由 Dustin Campbell 提交于
Ensure that an EnvDTE.CodeParameter.Name returns the correct value for parameters defined in metadata The issue is that the symbol display system will not return a name for a parameter unless SymbolDisplayParameterOptions.IncludeName is set. This behavior seems like it might ultimately be a bug in symbol display as that option is intended to control the display the parameters owned by the targeted symbol, but probably shouldn't affect the display of the targeted symbol. However, rather than updating symbol display (which would potentially have larger downstream impact), we'll just include the option in this case.
-
由 David Poeschl 提交于
Theme the rename dashboard's "New name:" TextBlock
-
由 Jared Parsons 提交于
Cache test results locally
-
由 Jared Parsons 提交于
Remove LTTNG warnings on Linux
-
由 Ty Overby 提交于
change all [Fact] private tests to be public
-
由 Jared Parsons 提交于
-
- 03 12月, 2015 21 次提交
-
-
由 Gen Lu 提交于
Fix TypeChar handler in Interactive Window
-
由 Dustin Campbell 提交于
Code Model: GetTypeSymbolFromFullName should not return unpredictable results when the type name was in an unsupported format
-
由 CyrusNajmabadi 提交于
Do not run Add-Using on syntax errors.
-
由 Jason Malinowski 提交于
Enforce that WpfFact tests must run serially
-
由 David Poeschl 提交于
Fixes #6721
-
由 Gen Lu 提交于
Move handling of TypeChar in InteractiveWindow from PreLanguageCommandFilter to PreEditorCommandFilter
-
由 Jared Parsons 提交于
Fix test issues
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Jason Malinowski 提交于
For now, WpfTestCase, when running tests, sets various bits of global static state. Because of this, it's not safe to let two tests interleave on the same test thread, and so we should explicitly prevent this. This change we should revert in the hypothetical future where we can run multiple WpfFact threads in parallel.
-
由 Ty Overby 提交于
-
由 Jared Parsons 提交于
This fixes as couple of test issues: - CultureAttributeMismatch: this was a flaky suite whose flakiness was fixed via the PEVerify changes. The hardened rules revealed the suite was setup to consistently fail verification because it didn't fully load it's dependencies into the Compilation object. This behavior is actually expected, we want to get the compiler errors that result from such a compilation. Hence we disable verification in those cases. - Return07_2: This was directly fixed by the PEVerify hardenning. closes #4300 closes #5866
-
由 Jason Malinowski 提交于
-
由 Jared Parsons 提交于
The LTTNG library version used in CoreCLR requires LTTNG_HOME be set else it issues a warning on process startup. This was cluttering up our linux build output. Fixing this by exporting the value.
-
由 Gen Lu 提交于
Fix `InteractivePasteCommandHandler` to handle line-copy and box-copy
-
由 Gen Lu 提交于
-
由 Dustin Campbell 提交于
-
由 Jared Parsons 提交于
Fix handling of abstract/virtual in structs
-
由 Dustin Campbell 提交于
Throw an exception in Code Model when an attempt is made to get a metadata name for a type parameter This should never happen, but there is a case where it was (https://github.com/dotnet/roslyn/issues/7179). That particular scenario has been corrected but we'll throw here in case it every happens again. Otherwise, we could stack overflow.
-
由 Dustin Campbell 提交于
Fix Code Model's GetTypeFromFullName to better handle situations where unsupported name formats are passed `GetTypeFromFullName` first attempts to treat the specified type name as a metadata name. Failing that, it parses the type name and tries to speculatively bind it. However, it was entirely possible that a type name could be provided that we don't support, such as an assembly-qualified name. So, we now check to see if the parsed type name includes an skipped token trivia (e.g. a trailing "Foo, 0.0.0.0") and does not attempt to bind in that case. Otherwise, speculative binding could return something non-sensical.
-
由 Paul Chen 提交于
-
- 02 12月, 2015 6 次提交
-
-
由 CyrusNajmabadi 提交于
Don't parse out a local declaration where there clearly isn't one. Fixes: #4014
-
由 CyrusNajmabadi 提交于
-
由 Andy Gocke 提交于
Remove BuildTask ref from VS Setup project
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
If we see: ```c# Identifier1 Identifier2(); ``` Don't parse that out as a declaration of a local called "Identifier2".
-