On Wed, Jan 20, 2010 at 7:45 AM, Robby Findler <[email protected]>wrote:
> Let me just double check something to be sure I'm on the same page. > The planet mirror will completely avoid the current planet server, > right? So, it has to be able to make the decision: given a planet > package request (including the version number of the mzscheme loading > the package) and the information on the mirror, which package to > serve, right? > Yes - the planet proxy/mirror has the same decision logic (sans bugs) that I have observed from the current planet server externally, so once a package is downloaded it can avoid hitting the planet server again. > I'm thinking you don't want to duplicate that code, right? Do you have > a plan for that? One thing would be for you to actually run a planet > server on the mirror, no? > I agree that generally it's better not to duplicate code. In this case I built the proxy/repo (and add the mirror/crawler in the next release) as a proof of my bzlib/* framework, so the duplication is currently intentional (a part of my plan was to test the framework). The proxy/mirror is currently only for planet client consumption instead of human, so it is not a verbatim planet server clone. But building on top of bzlib/shp will allow me to extend that way as well. It would be interesting to see if the two can/should be merged/collaborated in the future. That is if you are interested of course ;) Also - thanks for exposing the archives link - that's exactly what I was looking for. Appreciated. Thanks, yc
_________________________________________________ For list-related administrative tasks: http://list.cs.brown.edu/mailman/listinfo/plt-dev
