No there is technically no impact in removing them from the file.
However, the use-case in which these are being generated and used at a Discourse site is that
-
We have Wiki post (Discourse post changed to Wiki so that any Basic user can edit them) used as Wiki pages (pages that are meant to be a source of common input and general knowledge; think Wikipedia). Thus one word with two different meanings here.
-
I frequently generate content using Graphviz - DOT and export it to SVG for use with such Wiki pages.
While I can program with DOT, SVG, and HTML, some to many users are not as skilled. The site is for SWI-Prolog and some of the users are high school/college students just learning programming. While us advanced programmers would consider learning any of these as just learning another programming language and just get it done, the site is to help others learn Prolog. I can’t expect them to edit DOT, SVG, and much less HTML.
- The pages being Wiki pages for editing by and for everyone, expecting them to also figure out that once they create an SVG they have to convert the CR/LF to LF and then find and remove the offending comment line is a bit to much to ask. My goal is to get them to help create useful Wiki pages, not pull their hair out.
Anyway, the bug is now noted and ways to work around it for experienced programmers is demonstrated either by changing the text or removing the offending comment.
Thanks.
PS
If the offending comment is a copyright, CC or other license, then that would be a problem. Also while I don’t know of or use any such add-on, add-ons sometimes extend languages by using specific patterns in comments as part of a language and removing them would be damaging.