Re: [Nautilus-list] harder tasks

[Darin Adler <darin bentspoon com>]

on 11/1/01 10:42 PM, Maciej Stachowiak at mjs noisehavoc org wrote:

>> OK everybody, Darin needs assistance.  Some general harder porting
>> tasks I cataloged while digging around in the code:
> 
> It would be great if we could get nautilus to actually run long enough
> to display a window without crashing on startup, even if the window
> contents are wrong initially. The idea of doing significant rewriting
> blind makes me nervous. I don't know which of your points below are
> needed to get to that point.

That's the path that I'm personally on. I am running Nautilus and fixing
each thing that I run into. I've only done that for a few hours so far -- a
lot more to do. But a few more days of this, and we'll have a window up, I
bet.

Doing these tasks would be a big help, though, even if the code is written
blind. Because the code hasn't been tested yet in most of these cases, so
it's not going to hurt much to have newly written untested code there. Of
course, once Nautilus runs we can do even better.

But at the moment, the only patches have been from Havoc and Thomas Canty
(thanks guys). So I'm not really having a problem with people doing too much
work too early :-)

>>  - fix the icon cache to not need gdk_pixbuf_set_last_unref_handler
> 
> Did that get removed for 2.0? That would be ironic, since it got added
> specifically for the benefit of this Nautilus code.

True. But the gtk maintainers didn't exactly remove that specific feature
for 2.0. They changed GdkPixbuf to be a GObject, so all the old unref code
was removed. Sadly, it's not clear if GObject has a suitable feature, so the
irony remains. I think that timj never quite approved of the whole thing :-(

    -- Darin