On 06/27/2012 06:48 PM, Makarius wrote:
The answer also depends on the HTTP technology. E.g. on CTAN you see
plain directory views by default, and then can ask to download zip etc.
which might be dynamically generated on the spot, or just taken
statically from somewhere in the backhand. Another example is hg, where
the archived formats can be requested dynamically, somehow in Python.
I wouldn't want to put much magic into the web server setup, which is
much harder to maintain than a few standalone scripts. In particular, it
seems that there is no simple ready-to-use solution out there. Of course
one can hack everything using cgi of some form, but I don't think it's
worth it.
The open directory layout might have a slight advantage here, because we
could put etc/nonfree files into each nonfree package, and the HTTP
service script we observe that accordingly.
Simple Alternative via directories:
/home/isabelle/components/public
contains tarballs of freely distributable components
(served via http)
/home/isabelle/components/nonfree
contains tarballs of non-free components
(not available publicly, only via TUM NFS)
/home/isabelle/contrib
automatically maintained via script: contains untarred versions of
the above (both public and nonfree)
Alex
_______________________________________________
isabelle-dev mailing list
[email protected]
https://mailmanbroy.informatik.tu-muenchen.de/mailman/listinfo/isabelle-dev
