> Jeff, I've added you to the page. Cool! Thanks, Daniel. Does anyone know when it will go live? I want to send an email around the office.
I went into some detail back when I posted that, especially related to the apache mod_jk setup and load balancing. Check the archive (search on the subject or poster). We deploy on RedHat Linux 7.2, Tomcat 4.0.3, and Apache 1.3 with mod_jk (not mod_webapp). We use the Turbine 2.1 release. I was conservative and chose the 'officially supported' version. I developed from the TDK because it was easy to get running. I would love for an official TDK2.2 to come out, or I'll just wait for 3.0. I agree with Alex about turbine being hard to learn (I already knew Velocity). It took me awhile to get used to the 'turbine way', especially related to actions vs screens, and how to use the 'doRedirect', 'setTemplate', etc, methods. Combining setPage with setAction is so great - the form processing is separated from the link destination, which is set by the front-end code (the less doRedirect or servlet forwarding, the better). Programmers no longer control page flow! We eventually began using torque (built-in) with MSSQL 2000, and we use the connection pool for raw SQL. I don't generate the DB from torque, or the tsite-schema.xml from the DB - I do it manually. I tried using the jdbc2xml and jdbc2sql tasks but was unsuccessful. So I had SQL server dump the database to sql, and then hacked that into xml. When I change/add tables, I have to double maintain the xml and run rebuild-om, but it's pretty easy. I actually check my OM code into CVS even though most of it is generated, because it changes so rarely (I had to hack the 'generated by turbine on xx/xx/xx' out of the torque templates to make this work better). The generated OM code works great! Hopefully when we get rid of our stored procedures and some of the more 'proprietary' raw sql, Torque will make it easy to switch to postgres or something on Solaris or Linux. I suppose I could switch to Sybase easily enough tho... I don't use all the capabilities of Turbine - no intake, no UI, no security/fulcrum/flux. I use User only to store session variables (we have 3 kinds of users, and I haven't tackled integrating that yet). I rely heavily on the cookie/parameter parsers and the action classes, and of course, Velocity. I also use the Email classes a lot (VelocityEmail is particularly cool). At first I used only screen classes and raw SQL (I ported straight from ASP), but now I'm refactoring the code into more Action classes, pull tools, and Torque OM's. You wouldn't believe how much nicer (smaller) our code base is! It's fun to use Turbine because I always find/learn new things as I work on new features - like the other day I was looking through the utility package to find something to convert %20, %2C, etc in strings back into normal characters (I ended up writing it myself), and found the CSVParser class. An upcoming project is a CSV import tool for franchise data, so I'll be looking into that. I hate re-inventing the wheel... All in all a good experience. -Jeff > > > > We would be honored to have www.franchise.com listed on the 'powered by > > > > turbine' page. > > > > > > What turbine's version are you deploying? > > > > > > Are you using Torque? with which Database? > > > > > > In your experience what are the do's and dont's of using turbine? > > > > > > Anything else that you'd like to share? > > > > > > For the rest of us, will be really interesting that the people of 'powered > > > by turbine' list provide some inside information when they become hall of > > > famers! ;-) -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
