Prev: Redirecting System.out and exotic characters
Next: How would I rewrite this to satisfy the code checker?
From: Lew on 5 Nov 2009 09:31
Knute Johnson wrote:
> Lew wrote:
>> Knute Johnson wrote:
>>> Window is an AWT component. No requirement to use the EDT on AWT
>>> components that I know of.
>>
>> From what I've read, all GUIs, even those not of Java, really need to
>> be single-threaded to work right. I've been under the impression that
>> this applies to AWT, too.
>>
>> I need strong evidence that AWT is thread safe, not evidence that it
>> isn't.
>>
>
> You won't find a single example on Sun's web site of AWT GUI creation on
> the EDT. You won't find any mention of AWT components needing to be
> created on the EDT in the docs or that AWT isn't thread safe.
>
> It is hard to prove the negative when there is no mention of the
> positive, or is it the other way round?
>
> But find me a mention in some Sun doc, tutorial or anywhere that says
> that it isn't and I will quietly fade into the background :-).

You're the one assuming an unprovable negative.

The literature says that all GUIs share the need to run GUI actions on the GUI
thread. You assume safety where danger might lurk without evidence that you
are correct. I assume danger with some evidence that it might exist. Absent
an authoritative assertion of safety, prudence is the superior strategy.

The burden of proof is on the assertion of safety.

--
Lew
From: Knute Johnson on 5 Nov 2009 11:59
Lew wrote:
> Knute Johnson wrote:
>> Lew wrote:
>>> Knute Johnson wrote:
>>>> Window is an AWT component. No requirement to use the EDT on AWT
>>>> components that I know of.
>>>
>>> From what I've read, all GUIs, even those not of Java, really need
>>> to be single-threaded to work right. I've been under the impression
>>> that this applies to AWT, too.
>>>
>>> I need strong evidence that AWT is thread safe, not evidence that it
>>> isn't.
>>>
>>
>> You won't find a single example on Sun's web site of AWT GUI creation
>> on the EDT. You won't find any mention of AWT components needing to
>> be created on the EDT in the docs or that AWT isn't thread safe.
>>
>> It is hard to prove the negative when there is no mention of the
>> positive, or is it the other way round?
>>
>> But find me a mention in some Sun doc, tutorial or anywhere that says
>> that it isn't and I will quietly fade into the background :-).
>
> You're the one assuming an unprovable negative.
>
> The literature says that all GUIs share the need to run GUI actions on
> the GUI thread. You assume safety where danger might lurk without
> evidence that you are correct. I assume danger with some evidence that
> it might exist. Absent an authoritative assertion of safety, prudence
> is the superior strategy.
>
> The burden of proof is on the assertion of safety.

So you say but Sun doesn't and they do for Swing. Sun never shows
invokeLater() on AWT code examples and they do for Swing. Sun never
mentions thread safety issues with AWT and they do for Swing. I think
you are assuming there are alligators in the swamp without getting in to
look :-).

--

Knute Johnson
email s/nospam/knute2009/

--
Posted via NewsDemon.com - Premium Uncensored Newsgroup Service
------->>>>>>http://www.NewsDemon.com<<<<<<------
Unlimited Access, Anonymous Accounts, Uncensored Broadband Access
From: Jim Janney on 5 Nov 2009 12:16
Knute Johnson <nospam(a)rabbitbrush.frazmtn.com> writes:

> Jim Janney wrote:
>> Anabolik <bumsys(a)gmail.com> writes:
>>
>>> I made in Timer. Here the code:
>>>
>>> int delay = 30000;
>>> Timer timer = new Timer();
>>> timer.schedule(new TimerTask() {
>>> public void run() {
>>> myDialog.toFront();
>>> myDialog.repaint();}
>>> }, delay);
>>>
>>> but my dialog did not appear on the front of all windows.
>>
>> Your code is running on the wrong thread. See SwingUtilities.invokeLater.
>>
>
> Probably not necessary. repaint() certainly not and toFront() is a
> method of Window and probably doesn't need to be called on the EDT.

Looking at

http://java.sun.com/products/jfc/tsc/articles/threads/threads1.html

I find

The following JComponent methods are safe to call from any thread:
repaint(), revalidate(), and invalidate(). The repaint() and
revalidate() methods queue requests for the event-dispatching
thread to call paint() and validate(), respectively. The
invalidate() method just marks a component and all of its direct
ancestors as requiring validation.

so yes, repaint() is safe. A quick search doesn't turn up anything
one way or the other for toFront(). Call me old-fashioned, but
"probably safe" doesn't seem like a good basis for making coding
decisions.

--
Jim Janney
First  |  Prev  | 
Pages: 1 2 3 4
Prev: Redirecting System.out and exotic characters
Next: How would I rewrite this to satisfy the code checker?