Detect md image link

[flurmbo]

Fixes #49238. I added an option to the regex to have an image instead of text for the link description. Then when iterating over regex matches providerInlineLinks will now check to see if an image resource was also matched, and if so add it to results. All the tests pass, as well as a new one I've added.

Feedback on the regex would be appreciated, also I think perhaps the code I copied starting here could be refactored into a separate function.

[mjbvz]

Thanks for taking a look! Changes look good, just a few tweaks.

Let me know if you have any questions about my comments or the code

[mjbvz]

Thanks for taking a look! Changes look good, just a few tweaks.

Let me know if you have any questions about my comments or the code

[mjbvz]

Regex looks good but I think it does need to be tweaked to be less greedy:

• Instead of using . for matching the contents between the brackets, match on what that content cannot contain, such as [^\]] so that we match any character that is not a square bracket.
• Use a non-greedy quanitfier: +? instead of just + so we consume the minimum number of characters that still meet the regexp

The reason is that for markdown files like:

[![alt text](image.jpg)](https://example.com) other text [![alt text](image.jpg)](https://example.com)

if the matching is greedy, the detected link could end up spanning both links.

[mjbvz]

Probably worth adding a test for that example md as well

[mjbvz]

Thanks! This change will be in the next VS Code insiders build and is scheduled to be included in VS Code 1.31

[flurmbo]

Awesome, thanks for the fast response