Craig Bradney wrote: > Its a plan, and I have done a lot of work to separate code wherever > possible, however this won't continue now until we move to Qt4 in 1.3.5! > where we get the benefit of the separation between core and GUI code in Qt4.
To explain further, Qt 3 is a single unified library that requires a GUI to be available in order to do some many basic operations (such as creating an off-screen pixmap or doing most other graphics operations) even when the application has not created any windows. Qt4 lifts that restriction and allows an application to start up in a pure console mode that doesn't require a GUI. Scribus could still support command line scripting in Qt 3, it'd just have to be run from a session that has a DISPLAY (under UNIX) or is enabled to interact with the user (Windows). Note that on UNIX you can do this with the Xtest server or the Xvfb server for VNC, so the output need not actually _go_ anywhere. Of course, getting Scribus to the point where it can do useful things on the command line is a lot of work too. Craig B. has been working on that for a while amongst other stuff, and there's still more to do. Batch PDF export should not actually be _too_ hard (last time I looked anyway) if someone wants to get in there and focus on that. It'll still involve some significant refactoring of the PDF output path to avoid direct access to the options dialog controls by the PDF export code. I've done bits and pieces toward this, but last time I checked there was still quite a bit more left. Anyhow, Qt4 is _extremely_ nice (I've been using it quite a bit lately) and it'll make this stuff easier, to there's probably not too much incentive to leap onto the automation stuff until that transition is done. -- Craig Ringer
