Re: Graphical help browser

[John Swensen]

On Nov 24, 2008, at 3:20 AM, Søren Hauberg wrote:

> man, 24 11 2008 kl. 09:11 +0100, skrev Rafael Laboissiere:
> > * Søren Hauberg <address@hidden> [2008-11-23 20:20]:
> >
> > >  These days it's quite easy to create special purpose web browsers
> > > (using webkit), so I figured I'd give it shot. At
> > > http://hauberg.org/wiki/doku.php?id=help_browser you guys can get  
> > > the
> > > source code to a graphical help browser, and see a screen shot. I've
> > > been using it for the last month or so, and it's quite stable.
> >
> > I tried your code, it compiles okay, but fails to run:
> >
> > octave:1> autoload ("__help_browser_add_path__", "help_browser.oct");
> > octave:2> __help_browser_add_path__ ("core", "core_html");
> > error: help_browser.oct: undefined symbol: gtk_dispatcher
>
> You shouldn't need to call 'autoload' manually, as that command is in
> the PKG_ADD file. Somehow it seems that the PKG_ADD file is not  
> sourced.
> Before calling any of the help_browser functions, try running all the
> commands in the PKG_ADD file manually from Octave. These commands are
>
>  autoload ("__help_browser_add_path__", "help_browser.oct");
>  mark_as_command ("help_browser")
>  __init_gtk__
>
> The important one you're missing is '__init_gtk__'. I don't understand
> why PKG_ADD isn't sourced, but that is the cause of your problems.
>
> > I adapted your code for Octave 3.0.1 (patches attached below)
> > and this may be the cause of the problem.
>
> No, you're changes seem just fine. I have used the code under 3.0.x  
> with
> similar changes.
>
> Søren

On a somewhat related note, do you know if the current implementation  
of the 'doc' function allows us to insert a different handler for doc  
requests?  Just like there can be a different pager, or different  
graphics backend, doesn't it make sense to allow different document  
viewers to be plugged in?

Also, I really like your package for generating the HTML pages.  A  
while ago I had attempted to modify the document generating process to  
get rid of all the Sourceforge headers and such, but couldn't get it  
to work.

John Swensen