[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[MiNT] XaAES 0.901
Hi,
I spent quite some time playing with the XaAES 0.901. First of all, I
had big problems getting it to work at all. It works when you start
it from mint.cnf (but only if you cd to the dir where XaAES binary is
just before), and as login shell, but it does not work when started
directly from bash. The result of that last is just a bus error (the
last text the XaAES says is "auto program"). By the way, Henk,
couldn't you catch SIGBUS and shutdown the AES gracefully? Nothing
can be loaded after that, i.e. no AES, without previous reboot.
The bad news is that to run it, I had to switch the memory protection
off. With mp my best success was that the computer rebooted when it
was about to load the desktop. I would think, that the programming
discipline imposed by memory protection could have good influence on
the XaAES stability, no?
Well, at the other hand I exspected something much worse. At last, I
am writing this in qed running under XaAES, and the last crash was
about a hour ago. The only bigger problem I am having is, that the
keyboard sometimes looses characters, and when I start typing too
quickly, some portion of spurious characters (not keypresses!) is
generated. Looks like a virus, really :-)
Things I like are nonmodal alerts, small amount of memory taken (ps
reports below 300 kilobytes, so it is a half or even 1/3 of the
amount of memory taken by N.AES or Atari AES 4.10, respectively), log
window. Also I like that it is free and almost works.
Things I don't like is the unpronounceable name (next one to "oAESis"
and "Thing"), the barocco look and of course bugs. And here are my
observations about bugs:
1. There are quite lot of redraw errors. The popup displayed by the
explode.prg causes the desktop to be redrawn improperly, when the
popup is closed. The "drive list" popup in the file selector goes all
black when the mouse pointer touches an empty field (empty fields are
there if someone hasn't an even number of partitions). The desktop
icons can overdraw alert boxes, and so on.
2. A window still can be positioned over the menu bar, just must be
fully extended vertically (try with winview.app, drag a larger file
on it, then click fuller and try to drag the window onto the menu).
Also, a window still can be opened so that it overdraws a part of
the menu.
3. The form_center() seems to center objects relative to the top left
corner of the screen, I am not sure if this shouldn't be the top left
corner of the desktop window instead (i.e. everything except the
menu). I am almost sure the wind_get(0,WF_WORKXYWH) should return
coordinates for the desktop window (excluding the menu), unfortunately, it
seems it always returns coordinates for the entire screen, even in N.AES,
though I guess in N.AES it should make a difference whether the menu is
set visible or hidden. It doesn't seem so.
4. When you iconify and uniconify the same window several times in a
row, the iconified instance never takes the same position as before,
but rather is placed at the next position each time. This is a bit
funny.
5. Hierarchical menus are fanaberious. It is rather difficult to
select something from the "Clients" menu. I also failed to open a
submenu in Profile 2.2.
6. The fileselector works only from the XaAES system menu, when you
invoke it from any other program (qed for example), the XaAES locks
up and there is nothing more to do than a reboot.
7. Radiobuttons don't deselect themselves automatically when some
other radiobutton of the group is clicked.
8. Seems that menu_popup() returns incorrect value when the user
clicks something outside the menu (should return 0 in this case).
9. Weird stuff happens when you select checkboxes in the "Configure"
tool of Taskbar 2.50. Also see Atari's Procalc, this is a very good
test for AESes, actually oAESis had more problems with it ;-)
Anyways, playing with the XaAES allowed me to find and fix some bugs in my
slb library, so that e.g. the Crypt! proggie (on my page) should now work
with XaAES - as soon as I upload new version of the slb it uses ;)
Oki, that's all for now,
--
Konrad M.Kokoszkiewicz
mail: draco@atari.org
http://draco.atari.org
** Ea natura multitudinis est,
** aut servit humiliter, aut superbe dominatur (Liv. XXIV,25)
*************************************************************
** Taka to juz natura pospolstwa, ze albo sluzy ono unizenie,
** albo bezczelnie sie panoszy.