Fixed the initial population of the epoc32 tree with rnd bins to deal with the fact that the zips are in a slightly different location. (cf the changes made for Bug 350) Simplified the way in which the setting of sf.spec.package.src.enable is handled - no need to use it to change the zipping template, just use it to control which parts of the template we invoke.

@echo off
setlocal
call %~dp0common\hlm.bat %*
endlocal