On Tuesday 11 March 2003 10:07 am, Andre Poenitz wrote: > On Tue, Mar 11, 2003 at 09:48:01AM +0000, Angus Leeming wrote: > > > The problem here is that we use direct access from the GUI to the > > > kernel to extract information instead of using the 'pipe interface'. > > > One reason for that is, of course, that our current 'pipe interface' > > > does not allow to retrieve things (dispatch only returns different > > > levels of success, no data). I think this is a fundamental flaw, and > > > we've hit that problem before e.g. when talking about getStatus()... > > > > No we don't. We pass a string that is used to populate a frontend > > variable. > > Ok, so I did not understand that. The string interface is all I want. > I'll have a look again later.
The current comment to 'show' in frontends/Dialogs.h is a little out of date. I'm going to modify it so: /** \param name == "bibtex", "citation" etc; an identifier used to launch a particular dialog. \param data is a string representation of the Inset contents. It is often little more than the output from Inset::write. It is passed to, and parsed by, the frontend dialog. \param inset is _not_ passed to the frontend dialog. It is stored internally and used by the kernel to ascertain what to do with the FuncRequest dispatched from the frontend dialog on 'Apply'; should it be used to create a new inset at the current cursor position or modify an existing, 'open' inset? */ void show(string const & name, string const & data, InsetBase * inset); Angus