Edit/Update Removal can eliminate too much white space.

[mogsdad]

Script: Magic™ Editor
Issue type: Bug
Link to example post: (where can the problem be reliably replicated?)

This:

Generally i need to pick up `authData` and `RequestData` elements to the one level up (or to delete `processRequestRequest` element).


**UPDATE**

> Spring WS offers "only" contract-first, starting from an XSD Schema

...gets changed to this:

Generally I need to pick up `authData` and `RequestData` elements to the one level up (or to delete `processRequestRequest` element).> Spring WS offers "only" contract-first, starting from an XSD Schema

...it should be this:

Generally I need to pick up `authData` and `RequestData` elements to the one level up (or to delete `processRequestRequest` element).

> Spring WS offers "only" contract-first, starting from an XSD Schema

...but the white space between the Block-quote and the preceding sentence has been removed.

Observed in this edit, for example.

With App.globals.showRules = true;, isolated the problem to an apparent interaction between two rules, editupdate and blanklines, shown below. Debug info didn't show any unexpected replacement, so perhaps there is a third player? (Placeholder logic, maybe?)

With either rule disabled, there's no problem (but that the rule isn't doing its job).

        editupdate: {
            // https://regex101.com/r/tT2pK6/9
            expr: /([-_*]+[\t ]*\b(edit|update)\b([\t ]*#?[0-9]+)?[\t ]*:*[\t ]*[-_*]+:*|[\t ]*\b(edit|update)\b([\t ]*#?[0-9]+)?\s*:+[\t ]*)/gi,
            replacement: "",
            reason: App.consts.reasons.noise
        },

...
        blanklines: {  // https://regex101.com/r/eA5hA2/2
            expr: /^(?: *[\n\r\f])+|(?: *[\n\r\f])+$|((?: *[\n\r\f]){2})(?:(?: *[\n\r\f]))+/g,
            replacement: "$1",
            reason: App.consts.reasons.layout
        },

This might be related to issue #39 , in that the placeholders / block replacement logic could be involved.

(Migrated from old repository #115.)

[makyen]

This appears to have been previously resolved. The edits provided as test cases no longer appear to have the problem.