aboutsummaryrefslogtreecommitdiff
path: root/doc/forum/public-web-frontend/comment_2_0026d7be6b17e50d86b3b54985882f80._comment
blob: bfb60ede1e59023d6e8bce61cd125c3390946263 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
[[!comment format=mdwn
 username="https://www.google.com/accounts/o8/id?id=AItOawmdLmVxigDvxiGhnQ34xhyAaxl0CRRe9gs"
 nickname="Bryan"
 subject="Alternate version - smarter integration with other remotes/web special remote?"
 date="2013-03-03T14:19:12Z"
 content="""
I love the idea of pushing files to \"visible\" using annex, but I'm wondering how difficult to would be to make this even fancier.

For instance, if there are files that were originally imported from the web special remote (and have http:// prefixes, since that's not guaranteed), and they're not actually in the pushed annex, generate a redirect for all of them and write a special remote-local .htaccess file.

Or, similarly, allow each remote to have some metadata which specifies what, if any, the URL prefix for a directory in annex is. Then, whenever you update a web remote, the auto-generated redirects list could contain not only files not present but with web-remote locations, but also other remotes which have the file under a web-visible path prefix. Bonus for allowing URL prefixes for S3-style remotes that are not encrypted.

My use case for this sort of fancy is that I have a large server with gobs of space but on a personal (slower) internet connection, and a much smaller hosted instance with much faster bandwidth. With this smarts, I could push and pull large/expected high traffic files to the higher bandwidth/smaller machine ahead of, say, announcing a new feature or sharing a link to a video via e-mail.. or even reactively if the lower-bandwidth server is getting overwhelmed. 
"""]]