Added support for multi-cursor comments to the comment plugin #3543

MiguelRoldao · 2024-11-21T06:04:53Z

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:

All the selected lines are taken from all the cursors and compiled into a table.
The lines table will be fed to toggleCommentSelection() which will (un)comment all of the lines, keeping indentation formatting.
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:

Open a file like the one below

# some comment
print("hello")

print("world")

Select the first two lines by holding shift and pressing arrow down twice.
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 · 2024-11-21T06:29:51Z

runtime/plugins/comment/comment.lua

 ft["dockerfile"] = "# %s"
 ft["elm"] = "-- %s"
 ft["fish"] = "# %s"
+ft["freest"] = "-- %s"


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.

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 ;)

So 05a6870, f794609 & a023a00 can be removed from the PRs history.

So 05a6870, f794609 & a023a00 can be removed from the PRs history.

These commits are still lurking around in the history.
Also the...

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

...can then be removed from the PR's description.

Andriamanitra

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

runtime/plugins/comment/comment.lua

Andriamanitra · 2024-11-24T06:05:11Z

runtime/plugins/comment/comment.lua

+            if cursor.CurSelection[endSel].X == 0 then
+                lines[cursor.CurSelection[endSel].Y] = nil
+                staticEnd = true
+            end


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

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

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

Gavin-Holt · 2024-11-28T21:22:54Z

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 · 2024-11-28T22:49:16Z

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 · 2024-11-28T23:56:13Z

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 · 2024-11-29T19:20:59Z

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.

…compatibility)

runtime/plugins/comment/comment.lua

MiguelRoldao · 2025-11-11T21:27:03Z

Sorry I have been away. Sometimes life gets busy, and suddenly its been almost a year.

Anyways, the version is back to how it's supposed to be. Also merged with the master branch to resolve conflicts.

I have been using this PR since and haven't found any further issues.

Anything else that needs to be dealt with before merging @Andriamanitra @JoeKar ?

JoeKar · 2025-11-19T10:12:11Z

runtime/plugins/comment/comment.lua

+    local curData = {}
+    -- gather cursor data and lines to (un)comment
+    for i = 0,#bp.Buf:getCursors()-1 do
+        local cursor = bp.Buf:getCursor(i)        


This line introduces trailing whitespaces, which are removed in 0165e41.
Looks like the whole history needs a cleanup.

JoeKar · 2025-11-19T10:23:10Z

runtime/plugins/comment/comment.lua

 local config = import("micro/config")
 local buffer = import("micro/buffer")
 local micro = import("micro")



Instead of merging the master branch into the PR branch it is more preferred to rebase the PR branch against the actual upstream master.
This allows us to extract a clean set of patches, which can be applied easily.

This also helps to cleanup the PR's history.

Andriamanitra reviewed Nov 21, 2024

View reviewed changes

Andriamanitra suggested changes Nov 24, 2024

View reviewed changes

Andriamanitra approved these changes Nov 26, 2024

View reviewed changes

MiguelRoldao and others added 7 commits December 16, 2024 19:20

add support for multi-cursor comments

59ee5d9

added comment type for the FreeST language

05a6870

added comment type for the FreeST language pt.2

f794609

aesthetics

f896faa

removed freest

a023a00

added exception for selections that end in a new line (for backwards …

0165e41

…compatibility)

bugfixes

1c1a7d9

MiguelRoldao force-pushed the master branch from c789c28 to 1c1a7d9 Compare December 16, 2024 19:28

JoeKar reviewed Mar 1, 2025

View reviewed changes

runtime/plugins/comment/comment.lua Outdated Show resolved Hide resolved

Neko-Box-Coder mentioned this pull request Oct 3, 2025

fix(plugin): Fixed multi-cursor commenting #3873

Closed

Merge branch 'master'

274e35d

MiguelRoldao changed the title ~~Added support for multi-cursor comments to the comment plugin (V1.1.0)~~ Added support for multi-cursor comments to the comment plugin Nov 11, 2025

fixed versioning and removed unused code

fb2f136

JoeKar reviewed Nov 19, 2025

View reviewed changes

Added support for multi-cursor comments to the comment plugin #3543

Are you sure you want to change the base?

Added support for multi-cursor comments to the comment plugin #3543

Uh oh!

Conversation

MiguelRoldao commented Nov 21, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Andriamanitra left a comment

Choose a reason for hiding this comment

Uh oh!

Andriamanitra Nov 21, 2024

Choose a reason for hiding this comment

Uh oh!

MiguelRoldao Nov 21, 2024

Choose a reason for hiding this comment

Uh oh!

JoeKar Mar 1, 2025

Choose a reason for hiding this comment

Uh oh!

JoeKar Nov 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Andriamanitra left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Andriamanitra Nov 24, 2024

Choose a reason for hiding this comment

Uh oh!

MiguelRoldao Nov 26, 2024

Choose a reason for hiding this comment

Uh oh!

Gavin-Holt commented Nov 28, 2024

Uh oh!

MiguelRoldao commented Nov 28, 2024

Uh oh!

Andriamanitra commented Nov 28, 2024

Uh oh!

JoeKar commented Nov 29, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

MiguelRoldao commented Nov 11, 2025

Uh oh!

JoeKar Nov 19, 2025

Choose a reason for hiding this comment

Uh oh!

JoeKar Nov 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

MiguelRoldao commented Nov 21, 2024 •

edited

Loading

JoeKar Nov 19, 2025 •

edited

Loading

JoeKar commented Nov 29, 2024 •

edited

Loading

JoeKar Nov 19, 2025 •

edited

Loading