Login | Register
My pages Projects Community openCollabNet

Discussions > dev > [gef-dev] Fwd: svn commit: r1186 - trunk/src/org/tigris/gef/base/SelectionManager.java

gef
Discussion topic

Hide all messages in topic

All messages in topic

Re: [gef-dev] Fwd: svn commit: r1186 - trunk/src/org/tigris/gef/base/SelectionManager.java

Reply

Author penyaskito
Full name Christian López Espínola
Date 2008-12-16 07:39:09 PST
Message I know this is old now, but for reference.

On Tue, Dec 9, 2008 at 10:44 PM, Dave Thompson <dave at davet dot org> wrote:
> Hi Bob,
>
> Ok, if you like and you are sure that the change is ok, please revert my
> revert...!
>
> Christian: is there any other reason we should be aware of?

It's fine. I learnt something with this :-)


> Dave
>
>
> Bob Tarling wrote:
>>
>> meant to go to list
>>
>>
>> ---------- Forwarded message ----------
>> From: Bob Tarling <bob dot tarling at gmail dot com>
>> Date: 2008/12/9
>> Subject: Re: svn commit: r1186 -
>> trunk/src/org/tigris​/gef/base/SelectionM​anager.java
>> To: Dave Thompson <argouml at davet dot org>
>>
>>
>> The original commit looked fine to me
>>
>> Presumably the confusion was
>>
>> for (int editorIndex = 0; editorIndex < editorCount;
>> ++editorIndex) {
>>
>> }
>>
>> as opposed to
>>
>> for (int editorIndex = 0; editorIndex < editorCount;
>> editorIndex++) {
>>
>> }
>>
>>
>> There is no difference. The editorIndex is set to zero for the first
>> iteration. After an iteration the second part of the for statement
>> checks the condition to loop again. If we loop again the third part of
>> the for loop states what to increment. That increment only happens at
>> the end of the loop
>>
>> So it makes no difference whether ++ is prefix or postfix. The prefix
>> style is a habit of mine from C++ days where it gives some
>> performance. With Java it almost certainly doesn't matter which is
>> used.
>>
>> See - http://java.sun.com/​docs/books/tutorial/​java/nutsandbolts/fo​r.html
>> - the increment only happens after the first iteration.
>>
>> Regards
>>
>> Bob.
>>
>>
>> 2008/12/9 Dave Thompson <argouml at davet dot org>:
>>>
>>> Hi Bob,
>>>
>>> Sorry, the explanatory email exchange was not on the dev list, and I
>>> probably should have made my revert message clearer. Full details below.
>>>
>>> Regards, Dave
>>>
>>> Bob Tarling wrote:
>>>>
>>>> I'm just curious. What was the problem with this?
>>>>
>>>> Bob.
>>>>
>>> Dave Thompson wrote:
>>>>
>>>> Oh crap. You are right. I've reverted my change just in case. I don't
>>>> know a way of specifying the iteration range using the 'cleaner' for
>>>> loop nomenclature, so best stick to the old one for now.
>>>>
>>>> Thanks for noticing!
>>>>
>>>> Dave
>>>>
>>>> Christian López Espínola wrote:
>>>>>
>>>>> Hi Dave!
>>>>>
>>>>> On Tue, Dec 9, 2008 at 9:07 PM, Dave Thompson <argouml at davet dot org>
>>>>> wrote:
>>>>>>
>>>>>> Author: dthompson
>>>>>> Date: 2008-12-09 12:07:58-0800
>>>>>> New Revision: 1185
>>>>>>
>>>>>> Modified:
>>>>>> trunk/src/org/tigris​/gef/base/SelectionM​anager.java
>>>>>>
>>>>>> Log:
>>>>>> Swapped over to generics inside one method, to simplify the code.
>>>>>>
>>>>>> Modified: trunk/src/org/tigris​/gef/base/SelectionM​anager.java
>>>>>> Url:
>>>>>>
>>>>>>
>>>>>> http://gef.tigris.or​g/source/browse/gef/​trunk/src/org/tigris​/gef/base/SelectionM​anager.java?view=dif​f&pathrev=1185​&r1=1184&r2=11​85
>>>>>>
>>>>>>
>>>>>>
>>>>>> ====================​====================​====================​==================
>>>>>>
>>>>>> --- trunk/src/org/tigris​/gef/base/SelectionM​anager.java (original)
>>>>>> +++ trunk/src/org/tigris​/gef/base/SelectionM​anager.java 2008-12-09
>>>>>> 12:07:58-0800
>>>>>> @@ -772,11 +772,9 @@
>>>>>> }
>>>>>>
>>>>>> if (layer != null) {
>>>>>> - List editors = layer.getEditors();
>>>>>> - int editorCount = editors.size();
>>>>>> + List<Editor> editors = layer.getEditors();
>>>>>> Rectangle dirtyRegionScaled = new Rectangle();
>>>>>> - for (int editorIndex = 0; editorIndex < editorCount;
>>>>>> ++editorIndex) {
>>>>>> - Editor editor = (Editor) editors.get(editorIndex);
>>>>>> + for (Editor editor : editors) {
>>>>>
>>>>> Have you seen that it was ++editorIndex? We were not using the first
>>>>> editor in the list.
>>>>> Your works different, so I'm not sure if it was a bug, or we are
>>>>> introducing a bug.
>>>>>
>>>>>
>>>>>> double editorScale = editor.getScale();
>>>>>> dirtyRegionScaled.x = (int) Math.floor(dirtyRegion.x
>>>>>> * editorScale);
>>>>>>
>>>>>> --------------------​--------------------​--------------
>>>>>>
>>>>>>
>>>>>> http://gef.tigris.or​g/ds/viewMessage.do?​dsForumId=5948&d​sMessageId=981833
>>>>>>
>>>>>>
>>>>>> To unsubscribe from this discussion, e-mail:
>>>>>> [commits-unsubscribe​@gef.tigris.org].
>>>>>>
>>>>>
>>>>>
>>>
>>
>> --------------------​--------------------​--------------
>> http://gef.tigris.or​g/ds/viewMessage.do?​dsForumId=429&ds​MessageId=981864
>>
>> To unsubscribe from this discussion, e-mail:
>> [dev-unsubscribe@gef​.tigris.org].
>



--
Cheers,

Christian López Espínola <penyaskito>

[gef-dev] Fwd: svn commit: r1186 - trunk/src/org/tigris/gef/base/SelectionManager.java

Reply

Author bobtarling
Full name Bob Tarling
Date 2008-12-09 13:40:53 PST
Message meant to go to list


---------- Forwarded message ----------
From: Bob Tarling <bob dot tarling at gmail dot com>
Date: 2008/12/9
Subject: Re: svn commit: r1186 -
trunk/src/org/tigris​/gef/base/SelectionM​anager.java
To: Dave Thompson <argouml at davet dot org>


The original commit looked fine to me

Presumably the confusion was

           for (int editorIndex = 0; editorIndex < editorCount;
++editorIndex) {

           }

as opposed to

           for (int editorIndex = 0; editorIndex < editorCount;
editorIndex++) {

           }


There is no difference. The editorIndex is set to zero for the first
iteration. After an iteration the second part of the for statement
checks the condition to loop again. If we loop again the third part of
the for loop states what to increment. That increment only happens at
the end of the loop

So it makes no difference whether ++ is prefix or postfix. The prefix
style is a habit of mine from C++ days where it gives some
performance. With Java it almost certainly doesn't matter which is
used.

See - http://java.sun.com/​docs/books/tutorial/​java/nutsandbolts/fo​r.html
- the increment only happens after the first iteration.

Regards

Bob.


2008/12/9 Dave Thompson <argouml at davet dot org>:
> Hi Bob,
>
> Sorry, the explanatory email exchange was not on the dev list, and I
> probably should have made my revert message clearer. Full details below.
>
> Regards, Dave
>
> Bob Tarling wrote:
>>
>> I'm just curious. What was the problem with this?
>>
>> Bob.
>>
>
> Dave Thompson wrote:
>> Oh crap. You are right. I've reverted my change just in case. I don't
>> know a way of specifying the iteration range using the 'cleaner' for
>> loop nomenclature, so best stick to the old one for now.
>>
>> Thanks for noticing!
>>
>> Dave
>>
>> Christian López Espínola wrote:
>>> Hi Dave!
>>>
>>> On Tue, Dec 9, 2008 at 9:07 PM, Dave Thompson <argouml at davet dot org> wrote:
>>>> Author: dthompson
>>>> Date: 2008-12-09 12:07:58-0800
>>>> New Revision: 1185
>>>>
>>>> Modified:
>>>> trunk/src/org/tigris​/gef/base/SelectionM​anager.java
>>>>
>>>> Log:
>>>> Swapped over to generics inside one method, to simplify the code.
>>>>
>>>> Modified: trunk/src/org/tigris​/gef/base/SelectionM​anager.java
>>>> Url:
>>>>
>>>> http://gef.tigris.or​g/source/browse/gef/​trunk/src/org/tigris​/gef/base/SelectionM​anager.java?view=dif​f&pathrev=1185​&r1=1184&r2=11​85
>>>>
>>>>
>>>> ====================​====================​====================​==================
>>>>
>>>> --- trunk/src/org/tigris​/gef/base/SelectionM​anager.java (original)
>>>> +++ trunk/src/org/tigris​/gef/base/SelectionM​anager.java 2008-12-09
>>>> 12:07:58-0800
>>>> @@ -772,11 +772,9 @@
>>>> }
>>>>
>>>> if (layer != null) {
>>>> - List editors = layer.getEditors();
>>>> - int editorCount = editors.size();
>>>> + List<Editor> editors = layer.getEditors();
>>>> Rectangle dirtyRegionScaled = new Rectangle();
>>>> - for (int editorIndex = 0; editorIndex < editorCount;
>>>> ++editorIndex) {
>>>> - Editor editor = (Editor) editors.get(editorIndex);
>>>> + for (Editor editor : editors) {
>>>
>>> Have you seen that it was ++editorIndex? We were not using the first
>>> editor in the list.
>>> Your works different, so I'm not sure if it was a bug, or we are
>>> introducing a bug.
>>>
>>>
>>>> double editorScale = editor.getScale();
>>>> dirtyRegionScaled.x = (int) Math.floor(dirtyRegion.x
>>>> * editorScale);
>>>>
>>>> --------------------​--------------------​--------------
>>>>
>>>> http://gef.tigris.or​g/ds/viewMessage.do?​dsForumId=5948&d​sMessageId=981833
>>>>
>>>>
>>>> To unsubscribe from this discussion, e-mail:
>>>> [commits-unsubscribe​@gef.tigris.org].
>>>>
>>>
>>>
>>>
>>
>
>
Messages per page: