Re: --enable-gtk-doc

[Owen Taylor <otaylor redhat com>]

To follow up to myself, I though of one disadvantage -
initially one reason why I wanted this was to make
'make distcheck' run a quicker, but on second thought -
this isn't a good idea, since it could mean not catching
missing files needed for building docs.

In fact, since 'make distcheck' actually runs 'make dist'
inside the new tarball, it might break make distcheck,
since --enable-gtk-doc is required for 'make dist'

So, I'd probably need to either

 - Make enabling docs the default when the build directory
   is called =build. 

 - Have a gtkdistcheck target is like distcheck but
   configures the tarball with --enable-gtk-doc.

Both ugly. still, I think I'm in favor of the change
in general.

Regards,
                              Owen


Owen Taylor <otaylor redhat com> writes:

> I'd like to make --disable-gtk-doc the default and add
> --enable-gtk-doc to autogen.sh.
> 
> Reasons for this:
> 
>  - Somebody building from a tarball has no reason to rebuild
>    the docs in almost all cases.
> 
>  - Rebuilding the docs is slow. Really slow.
> 
>  - Getting Jade setup can be a nightmare if you don't already
>    have a setup.
> 
>  - We are bad about putting out gtk-doc realeases in sync
>    with GTK+ releases, so you often have to get gtk-doc
>    from CVS to build with --enable-gtk-doc.
> 
> [ Note that --enable-gtk-doc does check for gtk-doc, so if
>   you don't have gtk-doc installed at all, you are fine 
>   currently. ]
>   
> Disadvantages:
> 
>  - If you are developing gtk+, you'll have to remember
>    to add --enable-gtk-doc, unless you use autogen.sh.
> 
>    (How many people develop GTK+ and don't use autogen.sh? 
>    Windows developers don't count since I'd imagine that
>    getting the gtk-doc stuff to build on windows would be
>    as much fun as having your beard plucked out hair by
>    hair.)