- 14 9月, 2015 1 次提交
-
-
由 Dustin Campbell 提交于
-
- 12 9月, 2015 6 次提交
-
-
由 Dustin Campbell 提交于
-
由 Dustin Campbell 提交于
-
由 Dustin Campbell 提交于
-
由 Dustin Campbell 提交于
-
由 Dustin Campbell 提交于
-
由 Dustin Campbell 提交于
-
- 11 9月, 2015 4 次提交
-
-
由 Dustin Campbell 提交于
-
由 Dustin Campbell 提交于
-
由 Dustin Campbell 提交于
-
由 Tomas Matousek 提交于
-
- 10 9月, 2015 2 次提交
-
-
由 Kevin Halverson 提交于
Move context.IsPreProcessorKeywordContext to beginning of all checks in overrides of AbstractSyntacticSingleKeywordRecommender.IsValidContext.
-
由 Brett V. Forsgren 提交于
-
- 09 9月, 2015 2 次提交
-
-
由 Kevin Halverson 提交于
-
由 Andy Gocke 提交于
-
- 05 9月, 2015 2 次提交
-
-
由 Jonathon Marolf 提交于
Remove check for 'IgnoreCase In CSharpAddImportCodeFixProvider we explicitly override IgnoreCase to be false. Therefore, we will never use a case in-sesitive comparer.
-
由 Manish Vasani 提交于
1) Reduce number allocations in the analyzer driver that show up perf traces for Picasso. 2) Currently we maintain a cache of CompilationWithAnalyzers for projects being actively analyzed. This is causing the VM to rise and also forcing GC to kick in at different phases causing regressions at various scenarios. The fix is to just create a new CompilationCompilationWithAnalyzers instance for each document/project analysis request.
-
- 04 9月, 2015 1 次提交
-
-
由 Jonathon Marolf 提交于
speculative semantic model determine if we should use the fully qualified name or not. What this approach failed to account for was if the name was going to be simplified further due to the containing namespace. To Fix this we run the simplifier over our document and check if there are any changes made to the namespace. If this is the case, then we re-sort the usings according to the simplified name.
-
- 03 9月, 2015 1 次提交
-
-
由 Andy Gocke 提交于
-
- 02 9月, 2015 3 次提交
-
-
由 Dustin Campbell 提交于
-
由 Dustin Campbell 提交于
Additionally, the IsDebugger and IsImmediateWindow properties have been removed from CompletionTriggerInfo. Any code that was checking these properties now checks for options that are set by the Controller in debugger scenarios.
-
由 Dustin Campbell 提交于
-
- 26 8月, 2015 4 次提交
-
-
由 Dustin Campbell 提交于
-
由 Dustin Campbell 提交于
-
由 Dustin Campbell 提交于
-
由 Dustin Campbell 提交于
-
- 24 8月, 2015 1 次提交
-
-
由 Eyas 提交于
Fix #732. C# completion list must not suggest `public` as a potential accessibility modifier to use inside an `interface { }` block; otherwise, it produces code that does not compile.
-
- 21 8月, 2015 5 次提交
-
-
由 Charles Stoner 提交于
-
由 Charles Stoner 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
由 Cyrus Najmabadi 提交于
-
- 20 8月, 2015 2 次提交
-
-
由 Ravi Chande 提交于
Before this change, showing MAS for a type with 4000 empty methods took about 2 minutes on my machine. With this change, it now takes 2-4 seconds. There were 2 different issues causing the slow generation: 1. When formatting the generated file, the formatting engine needs to find the distance between every token and the new line (or start of file) before it. This calls TokenStream.GetOriginalColumn, which computes the original column by walking backwards through its list of tokens counting the width of tokens and trivia until it finds a newline. Though this n ^ 2 operation takes place in all invocations of the formatter, it is normally ameliorated by the fact that real code has enough newlines that GetOriginalColumn doesn't actually traverse back to the first token in the file. Generated files have no newlines, so GetOriginalColumn takes up about half the time of generating a MAS document, because GetOriginalColumn is called on every token in the file. To mitigate this, we augment MAS' custom formatting rule to not calculate anchor indentation. Anchor indentation is what requires the original columns, but no operations are produced for MAS documents, which have no newlines in them. 2. If instructed to use "auto insertion location", the code generation service computes the location to insert each member, one at a time. This computation checks that the type's members are sorted, resulting in n ^ 2 iterations over the list of members. The fix for this takes Kayle's fix for enum members a little further: When instructed not to use "auto insertion location", the code generation generates declaration syntax for each member without inserting it. It then sorts that list of declarations once, and bulks adds them (essentially appending each declaration to the end of the growing type declaration). I've fixed up some of the "Create*Declaration" methods to match the behavior of the corresponding "Add*Declaration" methods.
-
由 Jonathon Marolf 提交于
-
- 19 8月, 2015 6 次提交
-
-
由 Jonathon Marolf 提交于
-
由 Jared Parsons 提交于
This change does two actions to our public API files: - Resets them to the versions from 1.0.0 RTM - Splits them into the shipped and unshipped files
-
由 Jonathon Marolf 提交于
-
由 Balaji Krishnan 提交于
Add test for Block Context within AutoProperty Initializer. Add better explaining comments about preferring block over initializers. Addresses code review feedbacks.
-
由 Balaji Krishnan 提交于
Fixes #936. This allows introduce variable codefix to function in autoproperty initializers for both VB and C#.
-
由 Jonathon Marolf 提交于
-