- 14 3月, 2019 5 次提交
-
-
由 Chris Sienkiewicz 提交于
- Don't do analysis for speculative attributes - Manually call TryGetThisParameter, rather than MethodThisParameter which can throw
-
由 Chris Sienkiewicz 提交于
- Add an entry point to nullable walker for attribute analysis - Implement analyze attribute in the walker - Remove the analysis from binder_attributes and call nullable walker instead - Add extra fields to BoundAttribute that are needed for nullable analysis - Add some extra test scenarios
-
由 Chris Sienkiewicz 提交于
- Support supression - Fix params array of null - Fix arguments out-of-order
-
由 Chris Sienkiewicz 提交于
-
由 Chris Sienkiewicz 提交于
- At binding time check both constructor and value arguments for const null - When arguments are arrays, check for nullable array mismatch - When arguments are non nullable arrays check for nullable inside initializers - Add tests
-
- 13 3月, 2019 5 次提交
-
-
由 Heejae Chang 提交于
we want to move to newer version of JsonRpc (2.x) from 1.x we use. but that has some complexity since version 2.x and 1.x is incompatible and we let partner to directly access Rpc exposed from ServiceHubServiceBase type. this is first step so that partners move to new API that doesn't expose JsonRpc directly. once they moved, moving us to new versino should be easier since we will not break partner teams.
-
由 Charles Stoner 提交于
-
由 Heejae Chang 提交于
-
由 Neal Gafter 提交于
Fixes #33783
-
由 Andy Gocke 提交于
Ref and out parameters with suppression should change the nullable state for the argument being suppressed. Moreover, the changes should flow through type inference. Closes #27743
-
- 12 3月, 2019 9 次提交
-
-
由 Julien Couvreur 提交于
-
由 Julien Couvreur 提交于
-
由 Jared Parsons 提交于
Have to disable quiet restore explicitly on local dev machines when doing a bootstrap build to work around this arcade issue: https://github.com/dotnet/arcade/issues/2220
-
由 Jared Parsons 提交于
-
由 Jared Parsons 提交于
-
由 Neal Gafter 提交于
* Correct parsing of a parenthesized constant pattern expression in switch expression arm Fixes #33675 * Add a test that we properly bind and code-gen a parenthesized constant expression.
-
由 Jared Parsons 提交于
Did some profiling and all of the slowness is due to peverify. It seems to hit a wall about 700 levels of nesting. Once it's removed we can easily test our tolerance levels in the compiler
-
由 vsadov 提交于
-
由 Neal Gafter 提交于
* Implement LDM decisions on the behavior of `??` and `??=` operators Fixes #33743
-
- 11 3月, 2019 5 次提交
-
-
由 Jared Parsons 提交于
-
由 Jared Parsons 提交于
-
由 Jared Parsons 提交于
-
由 Jared Parsons 提交于
-
由 Jared Parsons 提交于
As a part of implementing nullable reference types many of our locals switched from `TypeSymbol` to `TypeSymbolWithAnnotations`. In the vast majority of cases this doesn't have a meaningful impact on compilation. They are bigger (about 3X) but it's still a relatively small `struct` (three words). The size difference is significant though in `BindNamespaceOrTypeOrAliasSymbol`. This method is used in recursive parts of binding and is mutually recursive with `BindQualifiedNam`. This method defines a large number of locals which contribute to every layer of recursion. When they moved to `TypeSymbolWithAnnotations` this pushed us outside our tolerance levels and we hit an overflow in extreme cases. Virtually none of these locals are used in the recursive case. Factored their use into local functions so we only pay the stack usage on demand. closes #33909 fixes https://github.com/dotnet/coreclr/issues/22757
-
- 09 3月, 2019 16 次提交
-
-
由 Tomas Matousek 提交于
-
由 vsadov 提交于
Bases of Nullable are reference types and thus base methods are called via boxed values which wold NRE on nulls. Fixes:#33174
-
由 Ivan Basov 提交于
-
由 vsadov 提交于
-
由 Tomáš Matoušek 提交于
-
由 vsadov 提交于
-
由 Heejae Chang 提交于
* support feature flag service in IExperimentationService this will let us to create private ring of specific groups for new features until it is ready for bigger group. once feature is ready for bigger group, this service can support VS experiment flight service which will let us to enable new features for bigger group. once we are confident on the feature, we will enable it for general public * remove extra blank line
-
由 vsadov 提交于
```cs S? x = <expr of Type S>; // works already S? x = new Nullable<S>(<expr of Type S>); // means the same as above and should work the same. ``` Fixes:#32626
-
由 Julien Couvreur 提交于
-
由 Yair Halberstadt 提交于
* Improve the error message for an 'in' extension method so it's clear why constraining it to a struct wont help.
-
由 Neal Gafter 提交于
* Eliminate unspeakable NullableAnnotation values. * Dynamic invocations result in an oblivious state Fixes #33638 * Fix up expected diagnostics Due to #33924 * Rename NullableAnnotation.Unknown to NullableAnnotation.Oblivious
-
由 Sam Harwell 提交于
See jaredpar/VsVim#2471
-
由 Sam Harwell 提交于
This reverts commit 4155c636, reversing changes made to 471c7e52.
-
由 Jason Malinowski 提交于
Co-Authored-By: Ngenlu <genlu@users.noreply.github.com>
-
由 Julien Couvreur 提交于
-
由 Sam Harwell 提交于
Cherry picked from commit aa2db962
-