Snip markers containing commas leads to unpredictable behaviour

[jonas]

Given a snippet marker like the following:

diff --git i/plugin/src/sbt-test/paradox/snippets/src/main/paradox/multiple.md w/plugin/src/sbt-test/paradox/snippets/src/main/paradox/multiple.md
index c8b7cb3..9c9b082 100644
--- i/plugin/src/sbt-test/paradox/snippets/src/main/paradox/multiple.md
+++ w/plugin/src/sbt-test/paradox/snippets/src/main/paradox/multiple.md
@@ -1 +1,3 @@
 @@ snip [multiple snippet](../../test/scala/Multiple.scala) { #multiple }
+
+@@ snip [comma separated snippet](../../test/scala/Multiple.scala) { #parseint-imports,parseint-def }

An error is logged in the wrong file:

[info] com.lightbend.paradox.markdown.Snippet$SnippetException: Label [parseint-imports] not found in [.../sbt_4232344d/snippets/src/main/paradox/../../test/resources/reference.conf]

Expected behaviour would be to fail with "invalid marker" or read multiple sections (as in #21)

[ktoso]

Agree, failing on missing things is very important