- 05 12月, 2018 13 次提交
-
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Julien Couvreur 提交于
-
由 Heejae Chang 提交于
* Set working folder for the solution added. persistent service requires working folder location to find out where to save its persistent file * added integration test to verify working folder exist
-
- 04 12月, 2018 6 次提交
-
-
由 Fredric Silberberg 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
由 maca88 提交于
Merged on behalf of @maca88 * Added UseOriginalEndOfLine option on SyntaxRemoveOptions to preserve the EOL from the original node (#22924). * Grammar correction. * Removed the normalizer parameter * Removed UseOriginalEndOfLine from shipped api * Added UseOriginalEndOfLine to unshipped public api * Added a null check in AsRootOfNewTreeWithOptionsFrom method * Removed HasEndOfLine method in favor of GetEndOfLine to avoid duplicate calls. * Modified KeepEndOfLine option to add an EOL for a directive only if there is one in the original document * Modified KeepEndOfLine option to preserve the EOL from the original node and removed UseOriginalEndOfLine option. * Modified AddEndOfLine method in order to avoid code duplication * Minor modifications * Added more tests and removed unneeded GetEndOfLine calls * Added a static method to test both Windows and Unix EOL and used where it matters. * Added a fallback EOL when requiresNewLine is set to true
-
由 Tomáš Matoušek 提交于
* Use Arcade testing targets * Skip CodeStyle tests on Mono * Include net472 in test target frameworks for Mono * Remove incorrect file existence check. * Update ADO integration test yaml
-
- 02 12月, 2018 1 次提交
-
-
由 Andy Gocke 提交于
This change simplifies our data flow analysis by removing an obsolete subclass and renaming some of the components to standardized names used in the literature on lattice data flow analysis. Fixes #23207
-
- 01 12月, 2018 7 次提交
-
-
由 AlekseyTs 提交于
-
由 Fredric Silberberg 提交于
-
由 Heejae Chang 提交于
* some code clean up and added some code instruments to make perf investigation easier * changed lambda to local function and addressed feedback
-
由 Joey Robichaud 提交于
* Classify malformed XML Doc comments as XmlDocCommentText * Added test for malformed xml doc comment classification
-
由 Heejae Chang 提交于
-
由 Manish Vasani 提交于
-
由 Spencer Lynn 提交于
* Updating classification for discard designation and underscore tokens to be Keywords rather than Identifiers. * Removing underscore from the IsKeywordKind check, since it is covered by isContextualKeyword * Reverting change to SyntaxFacts and moving discard check back to ClassificationHelpers. * Reverting all changes to SyntaxFacts. * Add a comment clarifying how _ is classified.
-
- 30 11月, 2018 10 次提交
-
-
由 AlekseyTs 提交于
-
由 Sam Harwell 提交于
-
由 Heejae Chang 提交于
make sure to VS and OOP to use same file to get mvid of analyzer file reference when creating checksum for both VS and OOP (#31438) the problem is that the OOP always used the actual file that got used for analyzer (in other words, shadow copied file in VS), but VS used original file to get MVID. based on shadow copying logic, 2 (with same assembly identity) might actually be 2 different files by sharing a shadow copied file between 2 original files with same assembly identitis causing MVID to be different. this make sure all use same file for MVID so that regardless of host implementation of analyzer loader (who is the actual thing that control what file will be used to get analyzers), we get consistent MVID
-
由 Tomas Matousek 提交于
-
由 Fredric Silberberg 提交于
-
由 Jason Malinowski 提交于
These methods are called any time we add a new project and we have a new output path; this was causing us to realize a Project object for every project in the solution, which was O(n^2) during solution load. This was adding a few hundred milliseconds of overhead in the immutable data structure, which in this particular case we don't need. Fixes dotnet/roslyn#31306.
-
由 Jared Parsons 提交于
This changes RunTests to always use procdump when tests timeout. Previously there was a single option, procdumppath, which controlled both monitoring xunit processes with procdump and using procdump on time out. The former causes stability issues at times with test execution and has to be disabled in a couple of places (spanish and integration tests). The latter has no stability issues and hence should always be used.
-
由 Shen Chen 提交于
Related issue: https://github.com/dotnet/roslyn/issues/25145 This PR add the function to VS and let use to pull members up via quick action
-
由 Manish Vasani 提交于
Address feedback from https://github.com/dotnet/roslyn/commit/dacca3d1bd6509935e4d01cf3f0e242075c217f2#r31467643 // "x is int _" => "x is int"
-
由 Manish Vasani 提交于
Address feedback from https://github.com/dotnet/roslyn/commit/dacca3d1bd6509935e4d01cf3f0e242075c217f2#r31467539 // "M(out var _)" => "M(out _)" // "M(out int _)" => "M(out _)"
-
- 29 11月, 2018 3 次提交
-
-
由 Manish Vasani 提交于
Add unit test for https://github.com/dotnet/roslyn/commit/dacca3d1bd6509935e4d01cf3f0e242075c217f2#r31465616 - unable to repro it locally
-
由 Manish Vasani 提交于
Fix code fixer for replacement inside a switch section. Addresses feedback from https://github.com/dotnet/roslyn/pull/31387/files#r236553196.
-