fix: pass image alt text to SVG images for accessibility

[afonsograca]

As a screen reader user, I should hear the alt text of an SVG image. Raster images already do this. The markdown ![Company logo](logo.svg) must announce "Company logo". It must not announce only the word "image". This must also work when a link contains the image.

In this PR we send the alt text to MDSvg:

• Renderer.image() now gives alt={alt || title} to MDSvg. The MDImage branch below it does the same.
• MDSvg already accepts this prop, and sets aria-label and accessibilityLabel. No code gave the prop a value. Each .svg URI used the default label "image".
• Renderer.linkImage() calls image(). Linked SVGs get the same correction.
• We add two tests next to the image test. One test checks the alt text. The other test checks the fallback to title.
• Both tests replace fetch with a mock. MDSvg then loads the file and shows SvgFromXml. Without the mock, the component stops in the error state, and the element with the label does not appear. This is why the old SVG snapshots show an empty View.
• We ran the two tests against the old code, and both tests failed. A snapshot test alone does not find an incorrect label.
• Two SVG snapshots in Markdown.spec.tsx.snap get the new alt prop. No other snapshot changes.
• We record the behavior in the "Images & SVG" section of the documents. The text covers both components, because the raster case also had no record.

We believe it is better to use the existing prop than to change the default label of MDSvg. This keeps the order of alt and title in one place, and makes the two image paths the same.