List continuation

[spoptchev]

Hello, is there a way to configure a formatter to insert a list continuation on a blank line when the previous list item is not blank?

* first item
* # continuation

* first item
* 
# no continuation

It would be even better if the empty list item is deleted once there is no continuation.

[vsch]

@spoptchev, what do you mean by list continuation?

Do you mean that an empty list item which is followed by a blank line should be deleted?

For example:

* list item 1
* 

* list item 2
* 
list item 3

would be formatted as:

* list item 1
* list item 2
* list item 3

[spoptchev]

Lets say I want to build a markdown editor:

If the user is starting to write a markdown list and hits enter on a list item I want to be able to automatically insert a opening marker for the next line item. If the user hits enter on an empty list item I want to delete it.

As an example:

* list item

When the user hits enter after list item I want to add a new opening marker for the list on the next line. The result would be:

* list item
* 

If the user tabs enter again on the empty list item I want to have the original state:

* list item

[vsch]

@spoptchev, I have the behaviour you describe implemented in my plugin: enter on list item inserts a list item marker on the next line, if enter is hit on an empty item then the marker is deleted.

However, I am doing this by deleting the list item marker text in the editor without using formatter or the markdown AST. I am guessing you are using the formatter to achieve this effect.

I will take a look at the formatter. I think adding an option to delete empty list items would achieve this effect.

[spoptchev]

I can think of several hacks to achieve this functionality 😄 . But it would be perfect if it could be done with a formatter. Then I could separate formatting from syntax highlighting.

Is your plugin open source? 😄 Can you share your implementation?

Thanks for helping me out 👍

[vsch]

@spoptchev, there is an open source version but I have not updated it in over a year. I diligently kept it updated for over a year and received exactly zero PRs or other participation beyond issues. After heavy refactoring of the licensed plugin the update would take a lot of effort and I did not see any justification to spend the time since no one is interested in contributing.

The OSS does not cover this part of the code and I am not sure it would help since it is based on JetBrains internal syntax tree which they call PSI and uses they internal PSI to implement functionality.

I am looking at the formatter to implement an option to delete empty lines, however, it will delete all empty items in the file not just the item on which enter is hit.

[spoptchev]

I understand. Yes, its a completely different environment.

Can you point me in the right direction where I could extend the formatter to implement this kind of functionality?

[vsch]

@spoptchev, almost every extension implements a NodeFormatter, however the closest to list items is TaskListNodeFormatter.java

The formatter extensions use the same concept as the Renderer extensions of the main library except the context and interface is formatter specific.

The formatter code handling list items is found in the formatter module's core node formatter implementation: CoreNodeFormatter.java

[spoptchev]

Thank you, I will take a look at it

[vsch]

@spoptchev, new version released.

• Fix: List continuation #232, List continuation, add Formatter.LIST_REMOVE_EMPTY_ITEMS, default false. If
  true then empty list items or list items which only contain a BlankLine node are removed
  from the output.
• Fix: formatter output of empty list items
• Fix: formatter output of empty loose task list items

[spoptchev]

Hello again, I know now that I need to use a NodePostProcessor to get the additional * on the next line. And I got a partly working solution, but I do not understand how to add a completely new BulletListItem at the end of a BulletList. All the code examples from the other projects just create new items from existing parts of the AST.

As a very simple example I have in the process method of a NodePostProcessor:

val firstChild = node.firstChild as BulletListItem

if (firstChild.childChars.contains("\n") && !node.chars.contains("insert")) {
    val item = BulletListItem(CharSubSequence.of("${firstChild.chars}* insert", firstChild.endOffset))
    firstChild.insertAfter(item)
    state.nodeAdded(item)
}

When I execute the code the AST looks like this:

Document[0, 7]
                            BulletList[0, 7] isTight
                              BulletListItem[0, 7] open:[0, 1, "*"] isTight
                                Paragraph[2, 7]
                                  Text[2, 6] chars:[2, 6, "test"]
                              BulletListItem[7, 15] isTight

As you can see the last BulletListItem is not created like the test element. What do I have to do to create a new BulletListItem that is matching the AST from those of the list and has the correct line numbers and offsets?

[vsch]

@spoptchev, most node elements do not parse their text to set sub-sequences. This is done by the element parser since it already has this information.

Here is a sample code from the ListItemParser:

        myBlock = myListData.isNumberedList ? new OrderedListItem() : new BulletListItem();
        myBlock.setOpeningMarker(myListData.listMarker);
        myBlock.setMarkerSuffix(myListData.markerSuffix);

The markerSuffix is used by the to do list item to set the [ ]. So you don't need to set it.

Also keep in mind that a ListBlockParser traverses all the list items and computes the isTight setting depending on parser settings and list items. This happens during parsing when the list block item is closed. When you are adding the list item the list block element has already resolved its child items. ListBlockParser.java