Image Resize when available size.x > image width

[soufianekhiat]

Resize when ImGui::GetContentRegionAvailWidth() < imageData.size.x with respect of image ratio.

Tested to support image position on the line (not convicing).

float const imgPos = ImGui::GetCursorPosX();
contentSize.x -= imgPos;

TBD: support for:

![Alt](imgs/Image.png =250x50)
// And
![Alt](imgs/Image.png =250x)

[juliettef]

Thanks for the PR and apologies for the delay.

@dougbinks and I'd prefer the imageCallback to handle sizing. Indeed, although I could add a boolean to the MarkdownImageData struct to select whether to resize if not enough space, this could become unwieldy.

I think this would be more suitably added to the documentation as an example of how to auto-resize in the ImageCallback() definition in https://github.com/juliettef/imgui_markdown/blob/master/README.md#example-use-on-windows-with-links-opening-in-a-browser

Do you want to update your PR to modify the readme example or do you prefer me to do it?

[soufianekhiat]

I see, for now imageCallback do not allow use allow us to handle the resize of the image.
Currently the image callback is agnotic of the current ImWindow.
Two solutions:

• Either we change the callback which will receive as parameter let say an ImVec2 which is ImGui::GetContentRegionAvail();
  OR
• As you propose a boolean on the struct (cf. my last push)

What do you think? Personaly I prefer the first one, which fit in the essence of "The callback manage the resize".

[dougbinks]

The imageCallback function is called immediately before ImGui::Image is called, and so the ImGui context state is in the correct state to call ImGui::GetContentRegionAvail() as needed:
https://github.com/juliettef/imgui_markdown/blob/master/imgui_markdown.h#L584-L590

So the exact same code should work from within the imageCallback function, returning the size via the MarkdownImageData struct. Thus there are no changes needed to the callback.

[soufianekhiat]

I understand, it's just counter intuitive for me to have the user aware of the implementation, particularly with a callback.
Feel free to close this PR otherwise I can provide an update.

Best

[dougbinks]

Yes, I agree it's not intuitive so we need to improve the documentation & examples. Thanks for highlighting this issue!