On Tue, 27 Oct 2015 at 10:03:02, Christoph Gysin wrote:
> > Yes, this behavior is intended and I personally think it makes sense.
> > The semantics of sources the form filename::url is: Obtain the file from
> > the given URL but change the source name to the given file name. The
> > best way to replicate this semantics in HTML is to use the URL as link
> > target and use the name as link label.
>
> Yes, I guess in the case of filename::url this makes perfect sense. In
> the example of Peter though, it's more obfuscating the git url.
>
I don't quite understand this sentence because in Peter's example, the
source is of the form filename::url.
> One way to fix this of course is to not specify an alternative name:
>
> source=("git+https://github.com/CDrummond/cantata.git";)
>
> But this is still an invalid url. If the url starts with "git+", that
> should be stripped at least from the href.
>
URLs referring to a protocol other than HTTP/HTTPs are not invalid.
> I would suggest to use:
>
> <a
> href="https://github.com/CDrummond/cantata.git";>git+https://github.com/CDrummond/cantata.git</a>
>
So you are suggesting to replace the link target with a link to another
URL? That is not a good idea because we cannot simply assume that they
mean the same thing in general. Also, if your browser does not know how
to handle the git+https protocol, that is a client configuration issue
and not a server issue (if you consider that an issue at all).
> Chris
> --
> echo mailto: NOSPAM !#$.'<*>'|sed 's. ..'|tr "<*> !#:2" org@fr33z3
