- 02 4月, 2019 12 次提交
-
-
由 David Barbet 提交于
-
由 David Barbet 提交于
-
由 David Barbet 提交于
-
由 David Barbet 提交于
-
由 David Barbet 提交于
-
由 David Barbet 提交于
-
由 David Barbet 提交于
-
由 David Barbet 提交于
-
由 David Barbet 提交于
-
由 David Barbet 提交于
-
由 David Barbet 提交于
-
由 David Barbet 提交于
selection handler.
-
- 01 4月, 2019 1 次提交
-
-
由 Charles Stoner 提交于
-
- 30 3月, 2019 6 次提交
-
-
由 Neal Gafter 提交于
* Implement pattern-matching in the nullable walker Fixes #29909 Fixes #31881 Fixes #30952 Fixes #33499 Fixes #30597 Fixes #32414 Fixes #23944 * Remove infinite recursion by using an empty struct cache. * Changes per code review comments. * Remove debugging code accidentally left behind. * Analysis of patterns-matching in the nullable walker requires valid (>0) slots. * Skip a flaky test * Patch after merge. * Make ctor private to force use of factory methods * Correct a typo. * Fixup after merge.
-
由 Jared Parsons 提交于
-
由 Sam Harwell 提交于
-
由 Heejae Chang 提交于
* move StreamJsonRpc to 2.0 * use object and collection initializer * changed to throw rather than return * share JsonRpc creation when it can share code * add linked files to share code * more moving around so that we can have linked file working * addressed PR feedback * changed name to WatsonReporter * moved to latest streamjson version * added a way for partner to add thier own json converter to roslyn service hub service base type * fixed test break. it is caused by behavior changes between StreamJsonRpc 1.x and 2.x * added comment follow PR feedbacks * updated comment following feedbacks
-
由 Joey Robichaud 提交于
-
由 Neal Gafter 提交于
Fixes #34472
-
- 29 3月, 2019 13 次提交
-
-
由 Ivan Basov 提交于
-
由 Heejae Chang 提交于
* fixed issue where we crash due to our pending async work run after VS shutdown this is another case where we have a pending async task that run after VS shutdown and it throws and our fail fast code catch that exception and crash VS. general fix will be some thing like us making our fail fast code to aware shutdown situation and ignore any exception if we are in the shutdown situation. but until we come up with proper design, this should handle one of high watson hits. * PR feedbacks
-
由 Jason Malinowski 提交于
When we completed a batch, we'd always raise a generic ProjectChanged event, even if the change could be represented more precisely: if we only added a single Document we could still raise DocumentAdded but we'd still raise the generic ProjectChanged. This now raises more precise events, when possible. This should allow for downstream optimizations: adding a new document won't have to rerun syntax analyzers on everything else, for example. There was also a second issue where if the batch was closed and nothing happened at all, we'd still raise a ProjectChanged. This was because the old code tried having a special case where if it mutated the Solution and ended up with the same Solution at the end, it'd skip the actual change. However there was a bug where if it called Solution.AddProjectReferences or Solution.AddMetadataReferences and passed an empty list, it'd make a new solution snapshot. That's not necessary to do, so we optimize that at the Workspaces layer. Fixes https://github.com/dotnet/roslyn/issues/34309
-
由 Jared Parsons 提交于
-
由 Sam Harwell 提交于
-
由 Sam Harwell 提交于
-
由 Jared Parsons 提交于
-
由 Julien Couvreur 提交于
-
由 AlekseyTs 提交于
-
由 Heejae Chang 提交于
-
由 AlekseyTs 提交于
-
由 Jared Parsons 提交于
A recent refactoring caused a number of insertion failures as we weren't properly updating the contents of VS.ExternalAPIs.Roslyn.nupkg. This adds basic verification that the contents are correct based on our build output.
-
由 Joey Robichaud 提交于
* Added CSharp.Scripting to VS.ExternalApis pacakge * Add ExternalAccess to VS.ExternalApis package
-
- 28 3月, 2019 8 次提交
-
-
由 Shyam N 提交于
-
由 Tomáš Matoušek 提交于
-
由 AlekseyTs 提交于
-
由 Charles Stoner 提交于
-
由 Joey Robichaud 提交于
-
由 Joey Robichaud 提交于
-
由 Jared Parsons 提交于
The project which deploys ILASM tools ends up bringing two runtime packages that have the same assets: - runtime.win-x64.microsoft.netcore.runtime.coreclr - runtime.win-x64.microsoft.netcore.app Specifically assets like SOS.NetCore.dll exists at different versions in these packages and end up getting copied twice to the output directory. This double write ends up breaking our build (as well as basic isolation). The short term fix here is to no longer treat this as a code project but instead a tools project. That eliminates the MS.NetCore.app package and the associated double writes Long term though the root issue needs to be addressed: making ilasm and ildasm easier to deploy. https://github.com/dotnet/coreclr/issues/15059
-
由 Jason Malinowski 提交于
If you have an STA thread that created Runtime Callable Wrappers, it's possible to end up in a deadlock where the finalizer and app domain unload code are waiting for each other. The suggested workaround from the CLR team is to wait for finalizers before letting the domain shutdown but while the thread is still running. Fixes https://github.com/dotnet/roslyn/issues/34248
-