[Thread Prev][Thread Next]   >Date Index >Thread Index

[wmx] New tarball, and some questions

cannam - Thu May 11 15:26:54 2000

Right, I think we're getting a few too many "here's my own release,
sorry no patch" tarballs -- Henri's GNOME+group code, Sven's multihead
stuff et al.  So I've attempted to merge them and tidy things up a bit.


This tarball contains wmx-6pre1+gnome+group from Henri, wmx-6mh1 from
SvOlli, Lasse's keyboard-menu tweak, an 8-bit-charset-clean Rotated.[Ch],
my own channel-button patch and a new thing of mine that attempts to
make the circulation order a bit more useful (as per recent discussion;
see below for more).

If you guys are still working on the bits that interest you, please
could you use this release as a base and post patches against this?
It'd make things much simpler for everyone else.  Similarly, if you're
the author of one of the things mentioned above, or if you've put
forward another patch that should be in here, please test whether it
works in the above tarball and post a fix if it doesn't.

Now, problems that I'm having with this version:

(1) I think the keyboard menu is now broken; when I try to use it,
which I don't normally, it hangs wmx.  Maybe one of the patches has
interacted badly with another one, or maybe I've applied something
wrong.  Please could someone who uses and/or understands the
keyboard menu test and fix?

(2) The circulation-order code works like this: If you circulate
(using Alt-Tab) repeatedly, without releasing Alt in between, then
you get the same circulation order as before; but every time you
release Alt, the order reverses.  So if you always release Alt
between circulation attempts, you'll always flip between the top
two windows; if you circulate through and then release Alt and
circulate again, you'll see the same windows in the reverse order.

I don't know whether this is "ideal", but I like it a bit better
than before.  BUT there is a big problem with the implementation:
Because we now want to grab KeyPress and KeyRelease events
for the Alt key as well as for other keys that use Alt as modifier,
we now need to know which keysym Alt uses as well as which modifier
it is.  That means you need to code Alt twice in Config.h (as
CONFIG_ALT_KEY_MASK and CONFIG_ALT_KEY_SYM) and the circulation
code will only work if the two happen to correspond to the same
key.  Obviously this is pretty grim; can anyone suggest a fix?

(I guess 6pre3 will happen when both of the above problems have
been resolved.)