The following discussion from !44332 should be addressed: - [ ] @justin_ho started a [discussion](https://gitlab.com/gitlab-org/gitlab/-/merge_requests/44332#note_425519851): (+3 comments) > Question (non-blocking): I might not have the full context but just wanted to check if we have considered normalizing the input before it is sent to this function? Asking because `insertMarkdownText` is already quite complex as it is and adding extra responsibility of normalizing its own params doesn't really help. > > Here is something I tested that works the same: > ``` > diff --git a/app/assets/javascripts/lib/utils/text_markdown.js b/app/assets/javascripts/lib/utils/text_markdown.js > index f4c6e4e3584..64ae9866c25 100644 > --- a/app/assets/javascripts/lib/utils/text_markdown.js > +++ b/app/assets/javascripts/lib/utils/text_markdown.js > @@ -290,7 +290,7 @@ function updateText({ textArea, tag, cursorOffset, blockTag, wrap, select, tagCo > const $textArea = $(textArea); > textArea = $textArea.get(0); > const text = $textArea.val(); > - const selected = selectedText(text, textArea) || tagContent; > + const selected = (selectedText(text, textArea) || tagContent).toString(); > $textArea.focus(); > return insertMarkdownText({ > textArea, > @@ -378,7 +378,7 @@ export function addEditorMarkdownListeners(editor) { > blockTag: mdBlock, > wrap: !mdPrepend, > select: mdSelect, > - selected: editor.getSelectedText(), > + selected: editor.getSelectedText().toString(), > text: editor.getValue(), > editor, > }); > ``` > Not saying the above approach is better per-se but just wanted to raise the idea. > I agree with @justin_ho that any type casting required for `insertMarkdownText` to work properly should happen outside