On Tue, Sep 16, 2008 at 6:10 PM, Evan Gilbert <[EMAIL PROTECTED]> wrote: > The code uses the persistence layer and users from Google app engine > currently, but these look easy to swap out if we want - there's only about > 150K of code in the whole project, and a lot of this is for smart diffing, > etc. See http://code.google.com/p/rietveld/source/browse/. >
I think people is working on the django model layer, so that it will be able to work with Google AppEngine and also with local DB, but things are not really there yet. > Possibly a good next step is to try it out for a few code reviews and see > how it works with our dev processes? Since it sends email with the comments, > usage would be completely optional. > > Evan > > P.S. Still need to figure out the URL to get to a specific version of a file > in SVN (for programmatic access, not just for browsing) - please email me if > you have this. > There is a dirty, undocumented trick involving some magics !vc... etc, but I think using the client and the libs is the reccomended way to go. Regards Santiago > On Fri, Sep 12, 2008 at 9:53 AM, Santiago Gala <[EMAIL PROTECTED]>wrote: > >> Answer via cellular. >> >> Is'ot the soft designed to work only with google app engine? Not >> tracking so closely... >> >> 2008/9/12, Evan Gilbert <[EMAIL PROTECTED]>: >> > On Thu, Sep 11, 2008 at 6:25 PM, John Hjelmstad <[EMAIL PROTECTED]> >> wrote: >> > >> >> On Thu, Sep 11, 2008 at 5:59 PM, Evan Gilbert <[EMAIL PROTECTED]> >> wrote: >> >> >> >> > Wanted to revive the thread on using code review tools with Shindig. >> >> There >> >> > is a really good tool up at http://codereview.appspot.com, and I'd >> like >> >> to >> >> > discuss whether it is OK to use for reviews. >> >> > - Sample diff: http://codereview.appspot.com/5245/diff/1/2 >> >> > - Full review: http://codereview.appspot.com/5245 >> >> > >> >> > Here's my quick list of pluses and minuses... >> >> > >> >> > Good: >> >> > - Ability to email cc on reviews, so that the official record is still >> >> > on >> >> > shindig-dev@ >> >> > - Threaded comment support, very clean inline display of diffs >> >> > - Open source, so we can run our own instance if we need to customize >> >> > - But hosted for now, so we can easily try it out >> >> > - Doesn't require a login to view patches and comments >> >> >> >> >> >> > >> >> > Bad: >> >> > - Hosted on appspot, which requires a Google login to comment. We can >> >> > change >> >> > this if we run our own hosted version. >> >> >> >> >> >> How difficult would it be to maintain reasonably current code, were we >> to >> >> switch out a different set of IDs? Do you have any sense for how >> >> extensible >> >> the tool is, particularly for integration with JIRA, shindig-commit@, >> et >> >> al? >> > >> > >> > As long as we have python developers, should be easy. I imagine that it >> > would be a good enhancement to the core code to support all of these >> > features via config. >> > >> > If not, the code base seems to be about 150K, so I don't think we'll have >> a >> > problem keeping it up to date. >> > >> > http://code.google.com/p/rietveld/source/browse/#svn/trunk/codereview >> > >> > >> >> >> >> >> >> In general I'm all for it, particularly since it already works (per your >> >> example).. at least for patches to start. It's far better than reading >> >> manually through diffs in email. >> >> >> >> --John >> >> >> >> >> >> > >> >> > >> >> > How do people feel about trying this out on an experimental basis? >> >> > >> >> > Evan >> >> > >> >> > P.S. Does anyone know what the URL format is to access for revisions >> of >> >> > files in svn? It looks like the diffs only work against the most >> recent >> >> > version in SVN, as it is using >> >> > >> >> > >> >> >> http://svn.apache.org/repos/asf/incubator/shindig/trunk/java/server/src/test/java/org/apache/shindig/server/endtoend/EndToEndServer.java >> >> > ?*rev=690667*< >> >> >> http://svn.apache.org/repos/asf/incubator/shindig/trunk/java/server/src/test/java/org/apache/shindig/server/endtoend/EndToEndServer.java?*rev=690667* >> >as >> >> the URL format, which doesn't actually get the bolded >> >> > revision. >> >> > >> >> >> > >> >
