Side by side images

[wabiloo]

Hi,

I'm trying to find a reliable way to display 3 images side by side (horizontal grid) - they are all of the same size / aspect-ratio.
When the pixel mode is SYMBOLS (which seems to be the default that's decided by detect_capabilities() on my mac, both in VS Code (integrated terminal) and iTerm2 (not quite sure why it doesn't detect support for iTerm2 protocol), that's easy enough as I can extract the lines from the canvas, concatenate them, and print them line by line.

However when I force ITERM2 pixel format, that obviously cannot work.
I resulted to write some really clunky code that uses ANSI codes to move the cursor back up after printing each image, which kinda works.

However, I'd like to find a way to automatically determine:

• What pixel mode was used
• On that basis, determine what mechanism to use to place the images side by side
• And in particular, determine how many up-cursor characters to print to get to the top of the previous image.

Any suggestion?

[GuardKenzie]

Hiya! This is quite a late reply, sorry for that.

I think the best way to achieve this is using ANSI codes, as you stated. You can now use the new TermInfo.emit function to construct those.

You can check which pixel mode was used by just comparing CanvasConfig.pixel_mode to the enumerator values in PixelMode.

Since when you specify the canvas config height and width, you know the size of the area your image is going to fit into.

The only tricky part is that the amounts you need to move to the right and upwards after printing an image with each graphics protocol varies. For sixels, you need to move up and to the right CanvasConfig.height and CanvasConfig.width respectively, however, when using ITERM2 (and kitty), you need to move up CanvasConfig.height - 1 lines and you don't need to move right.

There is currently a bug in 1.2.0 where TermInfo.detect_capabilities doesn't detect ITERM2 image protocol support (my bad, I forgot to check for it in the code). I have fixed this in eefcb1f. You can remedy this by checking TermInfo.have_seq(TermSeq.CHAFA_TERM_SEQ_BEGIN_ITERM2_IMAGE) manually.