Re: [Iup-users] Some mixed bugs and questions

[Antonio S. <sc...@te...>]

   On the buttons you have those problems, are you using IMAGE and IMPRESS?

 

Thanks,

scuri

 

From: Robert Persson [mailto:r.k...@gm...] 
Sent: sexta-feira, 28 de maio de 2010 14:46
To: IUP discussion list.
Subject: Re: [Iup-users] Some mixed bugs and questions

 

Hello Scuri.

There has been a few places where the sluggish visual feedback of buttons
has annoyed me. My current project is a CAD/drawing system which has
unlimited undo/redo capabilities through an incremental spatial database.
Oftenly the user wants to go 'back' many steps, let say 10-20 by pressing
the undo toolbar button. Maybe i'm very quick but the visual feedback only
shows the button has been pressed 2-3 times on my computer, but the button
ACTION callback is called as many times as it's pressed (except for the
FOCUSONCLICK bug I reported earlier). I think I can live with this but it's
really nice whe the user interface is 'responsive'. This behaviour of
buttons has not been in any toolkit I've worked with before, for example
wxWidgets, Borland VCL, WindowsForms, FLTK, FOX toolkit and more.

/Regards
Robert Persson




2010/5/28 Antonio Scuri <sc...@te...>

  Hi,

 

  About the button problem, are you using it to build an increment/decrement
button set? Like in IupSpin? If so, I may have an alternative. To use a
scrollbar control. Notice that its buttons are much more simpler that a
standard button. And it can be reduced to show only its increment buttons.
What do you think?

 

  BTW, did you notice that IupText has the SPIN attribute?

 

Best,

scuri

 

From: Robert Persson [mailto:r.k...@gm...] 
Sent: quarta-feira, 26 de maio de 2010 10:39
To: IUP discussion list.
Subject: Re: [Iup-users] Some mixed bugs and questions

 

Hello Scuri!

*  On my windows 64 box, when a button has FOCUSONCLICK=NO, and is clicked
two times quickly (like 'double click'), the ACTION callback is called not
two, but three times! I suspect this has something to do with double click
handling in button control. Point 9 in my last mail (se below) was about a
delay when a button is clicked several times quickly. The ACTION callback is
called correctly, but the visual feedback has a really sluggish feeling. I
bet this has to do with the double click handling in the button control as
well. (note that I run without visual styles). It would be really nice to
have this fixed.

*  About point 1 below, binary incompability between ubuntu and the linux on
your support machines (CentOS???), is it known why this incompability
exists? Diffrent GTK versions? 

* In a future version of IUP, a nice utility function would be a way to
convert a K_ANY key converted to the actual output of that keypress.

/Regards
Robert.P., Sweden





2010/5/11 Antonio Scuri <sc...@te...>

 Hi Robert,

 Thanks for all the reports.


> 1) On a linux2.6 64 bit (ubuntu 9.10), the code below does not show the
> title text of the checkbox. The natural size of the checkbox seems to
> be correctly computed. The console also shows the following error
> running this program:
> "(iupctest:10430): Gtk-CRITICAL **: gtk_label_set_text: assertion
> 'GTK_IS_LABEL (label)' failed"

 That is a binary compatibility problem between the machine I built the
library and your system. I use Ubuntu myself, but the machine I use the
build the libraries are managed by the Tecgraf technical support, and they
are not Ubuntu. While I convince them to change the OS, the only solution is
to rebuild the library on your system. Only the main library is need to be
rebuilt. Let me know if this is a problem for you. In Ubuntu you will have
to install the gtk2-dev package and download the IUP source code, in the
iup/src folder type make.

 Tip, to download the source code from CVS type:

cvs -z3 -d:pserver:ano...@iu...:/cvsroot/iup co -P iup
(once)
cvs -q -z3 -d:pserver:ano...@iu...:/cvsroot/iup update
-P      (to keep it up to date)



> 2) In the tree control DELNODE=MARKED cannot delete the node with ID 0
> even if ADDROOT=NO

 Sorry, this is a bug. I removed the constrain in 3.1, but a "for" was
starting at 1. It is fixed on CVS. It affected all drivers.



> 3) PARENT attribute of tree return 0 for nodes that has no parent. This
> is a misbehaviour if ADDROOT=NO, because there is no way to know if a
> node is in the branch with id 0 or directly in the tree with no parent.

 Well, in fact it returns NULL. If you use IupGetInt to check the attribute
then the NULL will be converted to 0. You must check if it's NULL before
using IupGetInt.



> 4) I dont know if this is a 'bug' but TITLEFONTid in a tree returns
> nothing if the TITLEFONTid attribute was never set.

 That's correct.



> 5) VALUE attribute of tree returns 0 when no node is focused, just as
> the documentation tells. This was not a problem before, but now when
> ADDROOT=NO there is no way to tell if the first node is selected or
> not.

 Almost what you said. VALUE is focus, not selection. For backward
compatibility VALUE must return 0 when no node is focused, independent of
the selected node(s). It returns -1 if there are no nodes in the tree (this
is new in 3.1).

 To check if the first node is selected, use "MARKED0".



> 6) Setting MARK=CLEARALL in tree control does the same as setting
> MARK=MARKALL, ie. it selects all nodes in the tree. It seems like
> CLEARALL is misconnected. This bugs shows om my 64 bit windows7 box.

 It is fixed now on CVS. It affected only the Windows driver.



> 7) Question: RENAMESELECTION in tree control returns NULL when I try to
> use it. How is it supposed to be used? Sample?

 I forgot to document its usage. It is used only once after SHOWRENAME_CB
is called and before the text box is shown, so it is write-only, cannot be
consulted in RENAME_CB.



> 8) Question: How is the new attribute EXPANDWEIGHT for vbox and hbox
> children supposed to be used? In my test, i tried to add three buttons
> to a vbox, setting "EXPANDWEIGHT"="0.5" and "EXPAND"="YES" assuming the
> buttons would expand only through the half width of the vbox, but

> nothing happened compared to the usual behavior.

 EXPANDWEIGHT is used relative to the possible expansion of the control.
The expansion for each control is done if there is room in the box, but it
will be done equally for all controls than can expand inside that box.
EXPADWEIGHT allow you to control that allowing a control to expand more or
less than another. Usually you will set something like 0.7 in one control
and 1.3 in another control, so the second control will expand more than the
first. But notice that this will only work if both controls have EXPAND in
the direction of the box.

 But I just noticed a stupid error, in the implementation I used
"EXPANDWEIGTH" (noticed the misplaced TH at the end). It was wrong also in
my test application... What a shame... Although in the documentation is
spelled right. It is fixed on CVS.



> 9) This is a question/bug that has been up before but is not cleared.
> The button control in windows has a delay after it's clicked, making it
> impossible to click (send ACTION callbacks) several times a second.
> This can be easily tested by counting the times per second a button is
> clicked. It is impossible to more than about 1-2 times a second. This
> is no problem for toolbar buttons but it is really annoying for buttons
> that increases a value in a text box for example.

 Yes, that's correct. To accept a repeating click in a button, instead of
implementing the ACTION callback, use the BUTTON_CB callback. This is a
lower level callback and the results are much better for that application.

 Another interesting trick is that if you just want to enable a click and
hold behavior, start a timer in BUTTON_CB and stop it when the button is
released. See the implementation of the IupSpin control in the
"iup/src/iup_spin.c" module.



> 10) The new sizer control is very nice. One thing a noticed is that it
> was possible to drag the bar below the childs natural size. Maybe some
> way to set a limit would be nice?

 On the child you want to limit, set the MINSIZE attribute. But I just
noticed that IupSplit is not respecting MINSIZE and MAXSIZE, its children
are limited but the bar moves over the control. It is fixed in CVS. I also
added a new attribute for IupSplit, MIXMAX that allows to constrain its
interaction.


> 11) Finally some future feature request.

> >utf-8 is on the way and this is a major need for me. Will this be in
> like IUP3.2 or further inte future, IUP4.0??

 I think this will be in 3.2.



> >Images in lists has been discussed. Will thoose also work when the
> list is a drop down list?

 Yes, it will also work for drop down.



> >Setting a grid in ROWSELECT or COLUMNSELECT mode would be a nice
> feature as well as custom painting in the cells and/or images for the
> cells. This would make the grid a powerful listview control.

 Don't get which control you are talking about. But IupMatrix is our
current control most similar to listview. In IupMatrix you can draw in the
cells using CD. We have planned some improvements for IupMatrix, like
support for toggle marks, images or a progress indicator, instead of text in
a cell. This is planned also for 3.2.

Best,

Scuri



----------------------------------------------------------------------------
--

_______________________________________________
Iup-users mailing list
Iup...@li...
https://lists.sourceforge.net/lists/listinfo/iup-users

 


----------------------------------------------------------------------------
--


_______________________________________________
Iup-users mailing list
Iup...@li...
https://lists.sourceforge.net/lists/listinfo/iup-users