Re: [anjuta-devel] RFC: Killing Anjuta-Extras

[Johannes Schmid <jhs jsschmid de>]

Hi!

> I tried looking at this, porting shouldn't be too difficult. I already
> ported ~200 lines in plugin.c, this port would have an advantage of
> having a real plugin written in python (as opposed to my unpublished toy
> plugins that I often end up rewriting in vala) and will thus force us to
> "fix" non bindings-friendly API. For now, the only problem I saw is that
> pygobject doesn't support passing more than a callback to a function (in
> this case anjuta_plugin_add_watch).
>
> Another problem is that the plugin uses libbfd, and thus cannot be
> rewritten entirely in python, but since these changes are restricted in
> symtab.c, I guess it would be possible to have it as a python module.
>
> I'll post again when I have something to show off.

I don't think a 1:1 port is of any use. It would be better to redesign
the plugin a bit, like not using libbfd and possibly using the xml mode
of valgrind. It also has far too many options and the preference system
is really broken (doesn't use anjuta-preferences/gsettings).

Regards,
Johannes