Overlay wraps when the line font is height >1.0

[stfl]

In my org-mode config, my outline-* fonts have a height >1.0. Particularly, outline-1 has height 1.5.

In these lines, the blamer overlay wraps to the next line because apparently the calculation for available space is not correctly incorporating the different space required.

[Artawower]

Hmm, I remember this problem when the blamer doesn't have its own face height. In fact, I have no idea how to count the number of available characters for fonts other than 1.0 (https://github.com/artawower/blamer.el/blob/master/blamer.el#L695)

What is the size of the blamer message? Could you provide screenshot?

If the blamer size is different from the 1.0 point, could you try to setup line height for the blamer message? Something like that:

  :custom-face
  (blamer-face ((t :inherit font-lock-comment-face
                   :italic t
                   :font "Fira Code 14"
                   :height 0.9
                   :background unspecified)))

Actually I also use different font sizes for headers in the org file, and in my case it looks correct:

btw, the real cause may be in other packages, is this behavior relevant for emacs running with -Q and outline height settings?

[stfl]

Thank for having a look at this.
I already had a similar face config.

I noticed, in my config, when setting :height to a float, it is relative to the previous value.

  '(outline-1 :height 1.5)
...
  `(blamer-face :inherit font-lock-comment-face
    :italic t
    :font "JetBrains Mono"
    :height 0.9)

results in 90% of 1.5...
If I set it hard to 90. I get a consitant font size for the lens.

In your screenshot, I see that there is an arrow indicating that the line continues to the right?
Do you have word-wrap or truncate-lines enabled in the buffer?

[Artawower]

In your screenshot, I see that there is an arrow indicating that the line continues to the right?

Now I see the problem. Right now I don't have much idea how to fix it, but I'll think about it.

I'll try to play around with calculating the actual character width, but that's not available in asynchronous code subprocesses