A hack to support content length and content hash

[Sean Conner]

It was thus said that the Great Emery once stated:
> Hey folks,
> 
> I've found a solution to the great content length dilema that I may have
> unwisely contributed to, magnet links!
> 
> magnet:?xs=gemini://gemini.spam.works/mirrors/lainzine/lainzine.org/txt-releases/lainzine-1.txt&xl=60132&xt=urn:sha256:9e7b437857cd3939e39e0a1c953f73da3f9045fb0b878014e9917110f8652afd

  As a concept, this is cool, but in practice, I don't think this will work
all that well.  I thought I might try to create an alternative index page
for my Gemini site (gemini://gemini.conman.org/) with magnet: links, but ...

=> http://boston.conman.org/                    Blog
=> gopher://gopher.conman.org/1phlog.gopher     Phlog
=> gemini://gemini.conman.org/boston.gemini     Gemlog
=> mailto:sean at conman.org                       Email
=> http://www.conman.org/people/spc/            Web

  Mostly non-gemini links here, and for those, I have no control over the
size or the hash, so they would have to remain as is (unless I want to
burden my server with retrieving the size/hash of each target, which could
lead to stale information in the time between serving you the link and the
time it's actually checked).

  The one gemini link I do control, but it changes size whenever I post, so
that link would have to be updated whenever that file changes else I serve
up outdated information.

=> boston.gemini                                The Boston Diaries Feed
=> /boston/                                     The Boston Diaries Archive
=> boston.atom                                  The Boston Diaries ATOM feed
=> news.txt                                     News about this server
=> /gRFC/                                       Gemini Request For Comments
=> gemini://gemini.conman.org/sourcecode/       Source Code
=> https://github.com/spc476/GLV-1.12556        Source Code on Github
=> /extensions/                                 Custom Extensions to the Source 
=> /sourcecode/                                 Alternative Link
=> /source-code/                                Old link

  Most of these pages are dynamic in nature (although they don't change all
that often).  So this section will have to be updated as well.

  I'll spare you the rest of the links, which are more of the same.  Each
change to the server would require a near complete rebuilding of the site. 
It's a similar issue to the one I have with gopher, where the file type is
part of the URL instead of being associated with the actual content.

  -spc (That's just my take---if you like it, go ahead with it ... )