Download component doesn't work when + character is in filename

[juw177]

1. Upload a file to DAM with a + character in the file name
2. Use this in the Download component
3. The link comes up as broken

[ky940819]

For reference.

This issue is caused by line 180 here:

aem-core-wcm-components/bundles/core/src/main/java/com/adobe/cq/wcm/core/components/internal/link/LinkBuilderImpl.java

Lines 177 to 188 in d2a7b02

	private @NotNull Link buildLink(String path, SlingHttpServletRequest request, Map<String, String> htmlAttributes) {
	if (StringUtils.isNotEmpty(path)) {
	try {
	path = LinkUtil.decode(path);
	} catch (Exception ex) {
	String message = "Failed to decode url '{}': {}";
	if (LOGGER.isDebugEnabled()) {
	LOGGER.warn(message, path, ex.getMessage(), ex);
	} else {
	LOGGER.warn(message, path, ex.getMessage());
	}
	}

The path is decoded by the LinkBuilder. In your case the + symbol is decoded into a space.
Later, the path is re-encoded and the space is converted into %20 (as would be proper for a space).

The addition of this pre-decoding was done here:
87a8c2a

If I understand the change correctly, then the purpose was to handle situations where the link may come from author-input data and may already be encoded (i.e. a button where an author copy/pasted a link - which is reasonable for them to do). If the URL were not to be decoded, then it would end up double-encoded.

Ideally this wouldn't happen when the link is coming from a resource path (i.e. a page, asset, etc) because we know it's not already encoded.

This specific case could probably be eliminated by using a URL decoding method that adheres to RFC2396.
The java.net.URLDecoder utility states This class contains static methods for decoding a String from the application/x-www-form-urlencoded MIME format (see https://docs.oracle.com/javase/8/docs/api/java/net/URLDecoder.html).