Things to be handled before 4.8 / Schedule adjustments

[Jannis Pohlmann]

Hey folks,

a part of our team is very busy working and doing other things these
days, leading to inactivity in some areas of Xfce that badly need some
love before the 4.8 release. I've been more or less absent for
about two months myself, so I don't blame anyone.

The main activity concentrates around the new panel and thunar. 

1. AFAIK, the tasklist plugin is not feature complete yet and overall
   the new panel hasn't really been tested that much. I've been using it
   on my laptop for two weeks now and I'm quite satisfied with it
   though.

Thunar has seen its first release based on GIO (1.1.0) but there are
regressions and the port to GIO has consequences for other parts of
Xfce:

2. xfdesktop still uses ThunarVFS, and even though I'm planning
   another standalone maintenance release of ThunarVFS, xfdesktop needs
   to be ported to GIO (and be it for consistency reasons only, like
   using the same volume management framework). 

3. thunar-volman, which I've written about a lot recently, is
   incompatible with the 1.1.0 release of thunar. You could argue that
   it is just a goodie but it is an integral part of the user experience
   provided by thunar. I've laid the groundwork for porting thunar and
   thunar-volman from HAL to udev and GIO by discussing the advantages
   and disadvantages on my blog:
  
     http://gezeiten.org/post/2010/01/Thunar-volman-and-the-deprecation-of-HAL

   Like the port to GIO, the port to HAL might have small consequences
   for the rest of Xfce but I fear we have no other choice.

A new component introduced with 4.8 is garcon, which is used by the new
panel, the appfinder and xfdesktop. 

4. garcon lacks file system monitoring support. This is somewhat
   important for all applications using it, so I suggest that we
   implement the same basic monitoring support we had in libxfce4menu,
   that is, no fine-grained changed/removed/added signals. Instead,
   each root menu would have to be rebuild from scratch on a change
   event. This is easy to implement and has worked ok in the past.
   Garcon is not API stable, so we can rip this out again later in
   order to replace it with something more appropriate.

These four points (there may be a few more) need to be dealt with
before we can release 4.8. Since it looks like half of the team is busy
with what we call "real life", I suggest that we delay the release. 

By delaying I don't mean to give up our release model but rather to
lengthen the 4.8 cycle to, let's say, 10 months instead of 8. That
would mean the release phase (defined in more detail on
http://wiki.xfce.org/releng/4.8/schedule) would be moved from 

  2010-02-01 - 2010-04-12

to

  2010-04-01 - 2010-06-12.

This is just loud thinking though, I'd prefer to hear some opinions on
this before we fix anything. I'd also like to mention that I wouldn't
see this as a big failure on our side but rather as an unfortunate
consequence of the fact that Xfce is a project whose pace entirely
depends on the free time of its developers.

So, yeah, what do you think about all this?

Cheers,
Jannis
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL: <http://mail.xfce.org/pipermail/xfce4-dev/attachments/20100119/1691d79b/attachment.pgp>