Add backtracking loops, backreferences, and if-then-else constructs to Regex "simplified" code gen #61906

stephentoub · 2021-11-22T04:02:55Z

Contributes to #61902

Adds support to our "simplified" Regex codegen for backtracking loops, backreferences, and both if-then-else constructs.

cc: @joperezr

ghost · 2021-11-22T04:03:05Z

Tagging subscribers to this area: @eerhardt, @dotnet/area-system-text-regularexpressions
See info in area-owners.md if you want to be subscribed.

Issue Details

Contributes to #61902

Adds support to our "simplified" Regex codegen for backtracking loops, backreferences, and both if-then-else constructs.

cc: @joperezr

Author:	stephentoub
Assignees:	-
Labels:	`area-System.Text.RegularExpressions`
Milestone:	-

src/libraries/System.Text.RegularExpressions/gen/RegexGenerator.Emitter.cs

src/libraries/System.Text.RegularExpressions/tests/Regex.KnownPattern.Tests.cs

joperezr · 2021-11-22T22:59:52Z

...libraries/System.Text.RegularExpressions/src/System/Text/RegularExpressions/RegexCompiler.cs

@@ -74,6 +74,7 @@ internal abstract class RegexCompiler
        private static readonly MethodInfo s_stringIndexOfCharInt = typeof(string).GetMethod("IndexOf", new Type[] { typeof(char), typeof(int) })!;
        private static readonly MethodInfo s_stringLastIndexOfCharIntInt = typeof(string).GetMethod("LastIndexOf", new Type[] { typeof(char), typeof(int), typeof(int) })!;
        private static readonly MethodInfo s_textInfoToLowerMethod = typeof(TextInfo).GetMethod("ToLower", new Type[] { typeof(char) })!;
+        private static readonly MethodInfo s_arrayResize = typeof(Array).GetMethod("Resize")!.MakeGenericMethod(typeof(int));


mmm interesting, I would have expected this to trigger a linker warning since the place using this is not annotated as RequiresUnreferencedCode cc: @eerhardt

…onversion

Labels need to be followed by something other than a closing brace. Previously I was trying to opt labels in to emitting a semi-colon, but it's too error prone. Instead, we now by default emit a semicolon, and only skip it from call sites that opt-out because it's obvious they'll always be followed by code.

src/libraries/System.Text.RegularExpressions/gen/RegexGenerator.Emitter.cs

src/libraries/System.Text.RegularExpressions/tests/MonoRegexTests.cs

dotnet-issue-labeler bot added the area-System.Text.RegularExpressions label Nov 22, 2021

stephentoub mentioned this pull request Nov 22, 2021

Finish "simplified" regex codegen strategy #61902

Closed

10 tasks