TMI: Another month, another office
Jul. 1st, 2013 07:16 pmToday was my first day back in Mountain View, from which I've been gone since mid-March. Everything is pretty much the same. Except the new people. And there are gummy bears in the kitchens now.
I re-submitted #7419, my pull request to make rustpkg build with no argument work, after some frustrating fumbling around trying to figure out a problem with a test case hanging that turned out to be due to a bad merge that I did. I accidentally un-ignored a test I didn't mean to enable; oops. But with any luck, this change will make it into the impending 0.7 release!
I also worked on getting the package database to work. I decided to start small, by creating the package database either in ~/.rust or (if there's a RUST_PATH) the first directory in the path. The package database, for now, is just a text file that contains a list of package IDs separated by newlines. Simple as can be. Of course, we'll want to store more information about each package, and probably use JSON or something, but I just want to get something working first. I took the "just write the test!" advice to heart -- somehow writing a test is almost always more manageable than writing the non-test code -- and wrote a test that installs a package, then calls rustpkg list and checks that the package ID is in the output.
I re-submitted #7419, my pull request to make rustpkg build with no argument work, after some frustrating fumbling around trying to figure out a problem with a test case hanging that turned out to be due to a bad merge that I did. I accidentally un-ignored a test I didn't mean to enable; oops. But with any luck, this change will make it into the impending 0.7 release!
I also worked on getting the package database to work. I decided to start small, by creating the package database either in ~/.rust or (if there's a RUST_PATH) the first directory in the path. The package database, for now, is just a text file that contains a list of package IDs separated by newlines. Simple as can be. Of course, we'll want to store more information about each package, and probably use JSON or something, but I just want to get something working first. I took the "just write the test!" advice to heart -- somehow writing a test is almost always more manageable than writing the non-test code -- and wrote a test that installs a package, then calls rustpkg list and checks that the package ID is in the output.