Re: Object hierarchy generation problems

[Matthias Clasen <maclas gmx de>]

Am Fr, den 17.10.2003 schrieb Robert Varga um 23:17:

> I have to deal with a source tree that has functions/files/objects
> added/deleted about weekly, and  don't want to maintain a list myself. I
> hoped that gtkdoc-scan.c finds the objects and their declarations are
> put in the right places (-sections.txt and -types.txt, right?).

Hmm, if the functions are added and deleted within a week, they're
probably not worth documenting anyway...

> All that I want to know is that if I set up $MODULE-types and use
> gtkdoc-scangobj and I don't get object declarations in
> $MODULE-sections.txt and $MODULE-types.txt, can there be a problem in
> the C code, and can someone help me with this.

$MODULE.types is the input for gtkdoc-scangobj, which should produce 
$MODULES.hierarchy (containing the object hierarchy)
$MODULES.prerequisites (containing the prerequisite relation)
$MODULES.interfaces (containing the interface-implementation relation)
$MODULES.signals (containing signal descriptions)
$MODULES.args (containing properties descriptions)

$MODULE-sections.txt is an input for gtkdoc-mkdb. For each object listed
in a section, it takes the information from the files generated by
gtkdoc-scangobj to insert object hierarchy, signals, etc in the xml file
generated for that section.

At what step does this process fail for you ?

Matthias