> The idea is the viff.dk will display the latest code and documentation > whereas each release will contain a copy of the code and documentation > as it was at the point of release. That way people should be able to > always obtain API docs for their code, whether it being new or old. I > would of course love if anybody thinks this is a bad idea...
This is all great. You would probably also want to either fork the repository or at least set some kind of marker (I don't know the correct term) when a release is done. In this way it is possible to go back and fix bugs in the release and release that without upgrading or waiting to the next release. -- Janus Den 25/02/2008 kl. 13.26 skrev Martin Geisler: > Good evening everybody, > > I just saw that someone had pushed the 'Force Build' button on the > BuildBot page -- and that had triggered an error in the 'link trial' > step. > > That step simply does > > ln -s $(which trial) > > which means that we can run trial afterwards as ./trial. We do this > since to test with different versions of python, we can then simply > run > > python2.4 ./trial viff > > or > > python2.5 ./trial viff > > Using a symbolic link to the real Trail binary makes this work both > when > Trial is installed in /usr/bin (that is the case on my home > computer) or > in ~/opt/bin (which is the case on DAIMI). > > It turns out that if you ask for a build of the tip revision, then > BuildBot is smart enough to avoid deleting the build directory, and so > the trial link was left there -- that made the second call to ln fail. > > I have fixed it now -- thanks to whoever forced the build by pressing > the button! > > > I'll try and make a third builder now which runs Epydoc whenever the > source changes and maybe even uploads the API docs to viff.dk. > > The idea is the viff.dk will display the latest code and documentation > whereas each release will contain a copy of the code and documentation > as it was at the point of release. That way people should be able to > always obtain API docs for their code, whether it being new or old. I > would of course love if anybody thinks this is a bad idea... > > -- > Martin Geisler > _______________________________________________ > viff-devel mailing list (http://viff.dk/) > [email protected] > http://lists.viff.dk/listinfo.cgi/viff-devel-viff.dk _______________________________________________ viff-devel mailing list (http://viff.dk/) [email protected] http://lists.viff.dk/listinfo.cgi/viff-devel-viff.dk
