Added support for multi-cursor comments to the comment plugin (V1.1.0)

[MiguelRoldao]

To add multi-cursor support I had to re-engineer the algorithm of the plug-in quite a bit. I tried to reuse as much code that was already implemented as possible to avoid introducing new bugs. I've tested some use cases and I haven't detected any problems so far.

The new algorithm has the following workings:

1. All the selected lines are taken from all the cursors and compiled into a table.
2. The lines table will be fed to toggleCommentSelection() which will (un)comment all of the lines, keeping indentation formatting.
3. The cursors will be restored to their correct position.

Note: All cursor logic is now contained in comment().

I also sneakily added a comment type for the language I'm developing in my master's thesis :P (https://freest-lang.github.io/)

[Andriamanitra]

I played around with this a bit and everything seems to work alright even in weird edge cases like multiple cursors on the same line, selections with comments inside them etc. It's really nice that the cursor position is now updated accordingly when a line gets commented.

The only slight annoyance I found is that if the selection ends in a newline, the line immediately after the selection gets commented too. This is not how it used to work, and not how I would like it to work. To reproduce:

1. Open a file like the one below

# some comment
print("hello")

print("world")

2. Select the first two lines by holding shift and pressing arrow down twice.
3. Comment selection. Before this PR you would get comments only on the first two lines, but now you get a # on the empty line too.

[Andriamanitra]

This seems unrelated to the content of this PR. What's freest? Micro doesn't have built-in support for a file type called "freest". If it gets special treatment in the comment plugin it should probably have a file type (syntax file under micro/runtime/syntax/) as well. Whether or not it's popular enough to warrant built-in support could be a separate PR and discussion.

[MiguelRoldao]

Because the end of a selection happens to be on the new empty line it would be commented as well.

I agree that what you commented should be the correct behavior. In the last commit I added a fix for this. I added an exception to not comment the last line of a selection, if it is an empty line.

Regarding FreeST, you're right. I removed the changes. May add full language support in another PR at some point ;)

[Andriamanitra]

The latest changes introduced some bugs. I opened a PR in your fork with suggested fixes that you may use if you wish.

[Andriamanitra]

What happened here?

[MiguelRoldao]

To be honest, I have no idea how I didn't see this.

[Andriamanitra]

This can end up overwriting lines that were already set to true by another cursor.

Demo: https://asciinema.org/a/VB99WIg0Oo6HJAkqc4KVeHEPP

[MiguelRoldao]

Well caught! I have merged your suggestions. It solves the bug, and didn't seem to change any other behavior.

[Gavin-Holt]

Hi,

Sorry to be late to this contribution.

But can I re-state a strong preference for comment characters at the start of the line. #2282. This is especially useful for commenting out sections that already contain comments:

void somefunc(int a)
{
//  if (a) {
//    a += 2;
// // After adding two we can print
//    printf("a = %d\n", a);
//  } else {
// // Print "none" if not a
//    printf("none");
//  }
}

I know others feel differently, so an optional setting seems the sensible solution.

Would it be possible to use the commenttype setting to indicate that comments reside at the start of the line?
e.g.

      "*.lua": {
        "commenttype": "^-- %s"
    },

Or would another setting item be required to indicate comment location?

      "*.lua": {
        "commenttype": "-- %s",
        "commentloc": "^"
    },

Sorry I can't offer a solution, I don't think I have the knowledge to get it close to right.

Kind Regards Gavin Holt

[MiguelRoldao]

I honestly prefer the way it is implemented as of now. Having the least amount of white space between the comment characters and the text (while keeping all the comment chars of a multiline comment in the same column).

However, it should be pretty straightforward to implement a "commentindent" option:

• true (default): implemented behaviour.
• false: ignores indentation and places the comment characters in the beginning or the line.

I'd have no problem dedicating an hour to this. However, I don't know if this feature is inside the scope of this PR. Maybe @Andriamanitra has something to say about it.

[Andriamanitra]

The indentation issue (#2282) is mostly unrelated so I would fix it in a separate PR, but I don't know the preferences of the maintainers (@dmaluka and @JoeKar).

Personally I think indented comments are way more readable, but I guess there's no harm in adding an option as long as the default behavior stays the same.

[JoeKar]

The indentation issue (#2282) is mostly unrelated so I would fix it in a separate PR, but I don't know the preferences of the maintainers (@dmaluka and @JoeKar).

I like as it is by default right now including the alignment. Furthermore I support this comment.
When I remember right, then there are languages and/or parsers getting into trouble in the moment the syntactic indentation is interrupted by a comment started at the beginning of the line, but I could remember wrong.

I guess there's no harm in adding an option as long as the default behavior stays the same.

Yep. But if so then independent of this PR.