We're using Torque on a variety of projects here at Vecna and came up with a
strategy for incorporating Torque into our build.xml script.  The general
ideas are:

- we reference a central install of torque for each project;
- we include <torque-data-model> tasks directly in our build.xml, whose
properties reference a torque install and output Java source to a directory
of temporary source files;
- overridden Torque classes get checked into source control in the project
proper; autogenerated stuff doesn't
- we have a build-src directory to copy all source files from the project
into a temp directory immediately before compiling torque-generated classes
get copied there too (have to do this to deal with co-dependencies between
autogenerated Java classes and overriden ones)
- we use the standard <uptodate> task to re-run the <torque-data-model> task
on demand when torque-schema.xml is newer than anything Torque had spit out

It has worked out pretty well so far.  How does this compare with other
people's experiences?  Would it benefit anyone for me to spam the list with
portions of our build.xml files?

-- Bill


--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to