Re: repaint method and design question
Peter Duniho wrote:
On Tue, 01 Jul 2008 19:02:17 -0700, Lew <lew@lewscanon.com> wrote:
[...]
I read through that page and was unable to find a clear statement to
the effect of either type of "thread safe". I'd appreciate it if you
could direct me to the portion of the page that you believes does so.
You're absolutely right. It says that the repaint() schedules the
painting to happen on the EDT, not that the action is thread safe in
general terms.
But the latter is the question at hand. We're not concerned with
whether the actual painting happens on the EDT. That can be taken as
granted. The question is whether calling repaint() itself must happen
on the EDT. I've seen nothing to suggest that it's not required to be,
and the general rule described to me is that _all_ calls to AWT or Swing
methods _must_ be done on the EDT.
According to the document cited,
JComponent.repaint() registers an asynchronous repaint request to the
component's RepaintManager, which uses invokeLater() to queue a Runnable to
later process the request on the event dispatching thread.
That seems pretty clear to me, that all GUI actions occur on the EDT. I don't
see how whether repaint() being called off the EDT will hurt that.
The ambiguity that you dislike doesn't bother me, although I see what you
mean. The important things that must be on the EDT are, according to the
article. The question of whether anything of repaint() other than that
happens on the EDT is immaterial.
Again, I see the distinction that you describe, but I don't think it matters
from a practical standpoint. It's pretty clear that the document at least
intends to show that JComponent#repaint() can safely be called from off the
EDT, at least for the implementation described.
To me, the problem isn't whether the document is clear - it seems sufficiently
clear from here. The problem is whether it's reliable, the fact that it's on
java.sun.com notwithstanding. My distrust of the article stems from the fact
that it is not presented as a normative standard - I don't get from it that
Java promises that repaint() is implemented in an EDT-safe manner, only that
the particular implementation actually does so.
I would like to know of any documentation that commits to JComponent#repaint()
being an EDT-safe method, by which I mean it can be called from off the EDT
without harm.
--
Lew