Login
Username:

Password:

Remember me



Lost Password?

Register now!

Sections

Who's Online
155 user(s) are online (115 user(s) are browsing Forums)

Members: 0
Guests: 155

more...

Headlines

 
  Register To Post  

« 1 ... 4 5 6 (7) 8 9 10 ... 40 »
Re: Beta testing
Just can't stay away
Just can't stay away


See User information
(a) all

Also, as you can see, I cannot write at (a) in the Qt browser. For some reason the alt modifier has stopped working. Also there are other local problems with the Danish keymap: You cannot add a diaeresis (umlaut) to letters, or a grave or acute (i think they are called), because I am using RAWKEY input. I think I have to restructure the keyboard input code completely and do both RAWKEY and VANILLAKEY input. Thanks to Olaf Barthel for pointing out this solution.

Go to top
Re: Beta testing
Home away from home
Home away from home


See User information
@Alfkil

Quote:
Tooltip problem now fixed.


Great, thank you !

Please if you have time can you give a quick look also on the other pressing issues ?

- Menu (right click) is completely broken in beta 3, you can see it on the browser demo --> Issue introduced in beta 3 only

- Tabs sometimes need to be clicked twice --> An old issue

Quote:
The crash you are seeing in sharedmemory has to do with native dialogs, which are not recommended ;)


Ah understand !
Hope with the time ...


Go to top
Re: Beta testing
Just can't stay away
Just can't stay away


See User information
(a) samo79

The tab problem you are mentioning has to do with me reimplementing the mouse key handling. Sometimes, when you single click a single item twice (tab bars are interpreted as a single object, even when clicking different tabs), the single click events would be overrulled by the double click event. This is now fixed.

Also popup (as in menus) handling should be a lot better now.

Go to top
Re: Beta testing
Home away from home
Home away from home


See User information
Awesome, step by step we are getting closer to a perfect implementation !

Before you will be annoyed enough please let me to remember other 2

Quote:
4) - Double Focus on an input text: Try for example to open two separate Qt windows that contain an input type, for example Calculator
You will see that the area when you can entering a number will got the focus (cursor active and blinking) on both windows rathen then only on the one selected


This one doesn't seems a big issue, however would be cool if you can fix

Quote:
5) - As you know still the usual autoalign crap on windows and sometimes some graphical glitch when you try to reduce a window at a minimal size .. this second problem is not always visible, it mostly depens on what demo/example you are testing


I annoy you with this one many time i know, however during other test in various Qt programs i observe that this specific problem is not visible at all, for example the about window of MuseScore or ScreenGrab (that indeed like others they are normal popup windows) will be opened perfectly, fast and without any autoalign crap .. so i though, maybe you can use the same solution for any other type of windows ?

Same thing about glitch during resize, altrough here i think you can't use the same solution for any tipe of windows around ..

Go to top
Re: Beta testing
Amigans Defender
Amigans Defender


See User information
Quote:

alfkil wrote:
About URLs, I am not sure how to do it differently? Of course I could insert yet-another option in qtprefs. Or is there a more system compliant way?


Yes - check out launch-handler.
Basically if you open "URL:http://www.amigans.net" (and then close it again immediately) that will ask which browser to use from URL prefs, and open that URL.

Quote:

Also, as you can see, I cannot write at (a) in the Qt browser. For some reason the alt modifier has stopped working. Also there are other local problems with the Danish keymap: You cannot add a diaeresis (umlaut) to letters, or a grave or acute (i think they are called), because I am using RAWKEY input. I think I have to restructure the keyboard input code completely and do both RAWKEY and VANILLAKEY input. Thanks to Olaf Barthel for pointing out this solution.


Hmm... I'm not sure that's right. I use RAWKEY only in NetSurf and dead keys, qualified keys, etc, work just fine.
I basically just use MapRawKey, but if you search for ami_key_to_nskey in amiga/gui.c you'll see exactly how it works.
I keep track of the qualifiers too, but only for uses other than text input.

Go to top
Re: Beta testing
Home away from home
Home away from home


See User information
@alfkil

Just to add to the Chris post, using the same method you can also set the e-mail links to open them with a mailer

Go to top
Re: Beta testing
Just can't stay away
Just can't stay away


See User information
Quote:
4) - Double Focus on an input text: Try for example to open two separate Qt windows that contain an input type, for example Calculator<br>You will see that the area when you can entering a number will got the focus (cursor active and blinking) on both windows rathen then only on the one selected
Those are two separate apps, and there is no way I can make a global (accross apps) focus widget. Sorry.

Go to top
Re: Beta testing
Just can't stay away
Just can't stay away


See User information
@samo79: Auto align is troublesome. The X11 code, that I have used as a model for the amiga version is quite complex and pretty incomprehensible in this area. I am guessing this should be much less troublesome in Qt 5 (hoping hoping), and thus I am not going to spend a lot of time with it.

Go to top
Re: Beta testing
Just can't stay away
Just can't stay away


See User information
@Chris: Thanks for the explanation, I had completely overlooked the URL pref app :) Sounds like it should be easy-peasy to implement!

Go to top
Re: Beta testing
Home away from home
Home away from home


See User information
@Alfkil

Quote:
Those are two separate apps, and there is no way I can make a global (accross apps) focus widget. Sorry.


Well, mine was just an example and honestly not the best if i think more about, in reality we can see the same problem also with a single window

I'm not a coder at all but I think you may just "disabled" the cursor once the window will be unselected and then re-active the cursor again when the user click again on that window

At the end it's just a very basic behaviour common on any framework around, including Qt on other platforms or MUI/Reaction on Amiga

Go to top
Re: Beta testing
Not too shy to talk
Not too shy to talk


See User information
Chris: Do I remember that you have checked Calligra suite sources already? Now QT starts to be usefull and Calligra is at least advertiset to be "relatively easy to port", would it be possible to implement?

Up to date Office suite would be usefull.

Go to top
Re: Beta testing
Amigans Defender
Amigans Defender


See User information
Quote:

utri007 wrote:
Chris: Do I remember that you have checked Calligra suite sources already?


Nope, not me, but I've had a look at the build instructions (far more difficult to find than they should be), and it appears to need kdelibs - Qt alone is not enough.

Go to top
Re: Beta testing
Not too shy to talk
Not too shy to talk


See User information
For noncoder it is same thing ;)

Go to top
Re: Beta testing
Home away from home
Home away from home


See User information
@Chris

How about try to port Arora browser now ?

https://github.com/Arora/arora

It's no more developed, but it's better than our current instable Qt browser, can be a good candidate for deep testing and real usage of our Qt port !

Go to top
Re: Beta testing
Just can't stay away
Just can't stay away


See User information
@samo79
You mention Arora and stability. Well, the stability issues (or rather just bugs) in the current Qt browser come from libQtWebKit, and thus a port of Arora would have exactly the same issues.

About auto align "crap": I figured out, that the problem is in intuition. I have delayed the showing of windows until the autoalign has happened. But sadly, when trying to resize a hidden window, it is automatically showed on screen... Doh! I think I can cook up a solution, though.

Quote:
Calligra

Calligra needs A LOT more modules than just Qt and KDE. Most of them are probably pretty easy to port (I have done some already (untested)), but the sheer amount of code that has to be processed makes it quite impossible for me to go into it at the moment. Also it seems to need libQtDbus, which we don't have in the current Qt port.

Go to top
Re: Beta testing
Home away from home
Home away from home


See User information
@Alfkil

Quote:
You mention Arora and stability. Well, the stability issues (or rather just bugs) in the current Qt browser come from libQtWebKit, and thus a port of Arora would have exactly the same issues.


Ah ok then, I realized that the problem was only into the included test browser, not that the entire libQtWebKit needs some more love

Quote:
About auto align "crap": I figured out, that the problem is in intuition. I have delayed the showing of windows until the autoalign has happened. But sadly, when trying to resize a hidden window, it is automatically showed on screen... Doh! I think I can cook up a solution, though.


Thanks Alfkil
Sometimes ago i open a thread on the Hyperion forum and more or less the result of the discussion was that the best thing is to implement the necessary fix directly on the third parties software (in this case Qt) rather than Intuition

http://forum.hyperion-entertainment.biz/viewtopic.php?f=14&t=884

Personally i have noticed that this auto-align issue does not occur with any types of windows, for example the "About Qt..." window suffer of this problem and it's not the only one, but other Qt's windows not (the about window of MuseScore is ok for example)

Perhaps you could use the same code of that specific "good" window as a basis for any other windows of the framework ?

Don't know, just an idea

Go to top
Re: Beta testing
Home away from home
Home away from home


See User information
Some new issues:

1- Flick client
(Started from qtdemo) always crashed when you try to close it with ESC

Quote:
Crash log for task ""
Generated by GrimReaper 53.5
Crash occured in module at address 0x4F9149EC
Type of crash: ISI (Instruction Storage Interrupt) exception

Register dump:
GPR (General Purpose Registers):
0: 5FFAA000 4F9076C0 00000000 4F9076C8 55A3A818 511D3E9C 00000006 00000000
8: 6561645F 4F9149EC 00002E4D 00000003 00000007 00000000 FFFFFFFF 55E4A7F0
16: 02270000 7FD8DD10 02270000 4D0BBDC0 001E8FF8 00000000 4E5286B8 4F9611D0
24: 4E928B30 00000000 5494F800 4F907980 4F9079A8 511D3E9C 4F95B588 55A3A818


FPR (Floating Point Registers, NaN = Not a Number):
0: nan 79 0.329 0.64
4: 0.33 0.3 0.329 0.3127
8: nan 0.06 0.15 0.6
12: 0 0 0 -0
16: 0 0 0 0
20: 0 0 0 1.61895e-319
24: 4.94066e-324 0 0 -1.20437e+36
28: 0 1.35808e-312 8.8275e-311 -0.234375

FPSCR (Floating Point Status and Control Register): 0xA2002100


SPRs (Special Purpose Registers):
Machine State (msr) : 0x0002F030
Condition (cr) : 0x48444084
Instruction Pointer (ip) : 0x4F9149EC
Xtended Exception (xer) : 0x00000003
Count (ctr) : 0x4F9149EC
Link (lr) : 0x6E191DE0
DSI Status (dsisr) : 0x00000000
Data Address (dar) : 0x4E66444C



680x0 emulated registers:
DATA: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
ADDR: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
FPU0: 0 0 0 0
FPU4: 0 0 0 0



Symbol info:
Instruction pointer 0x4F9149EC belongs to module "" (HUNK/Kickstart)

Stack trace:

libQtCore.so:_ZN7QString9fromAsciiEPKci()+0x3c (section 10 @ 0xE8EC0)
libQtCore.so:_ZN28QEventDispatcherAmigaPrivate8doSelectE6QFlagsIN10QEventLoop17ProcessEventsFlagEEP7timevalPm()+0x51c (section 10 @ 0x2B34B4)
libQtCore.so:_ZN21QEventDispatcherAmiga13processEventsE6QFlagsIN10QEventLoop17ProcessEventsFlagEE()+0xd8 (section 10 @ 0x2B487C)
libQtCore.so:_ZN10QEventLoop13processEventsE6QFlagsINS_17ProcessEventsFlagEE()+0x164 (section 10 @ 0x247020)
libQtCore.so:_ZN10QEventLoop4execE6QFlagsINS_17ProcessEventsFlagEE()+0x1f8 (section 10 @ 0x2480B8)
libQtCore.so:_ZN7QThread4execEv()+0x29c (section 10 @ 0x4255C)
libQtDeclarative.so:_ZN24QDeclarativePixmapReader3runEv()+0x198 (section 9 @ 0xDE6BC)
libQtCore.so:_ZN14QThreadPrivate5startEPv()+0x100 (section 10 @ 0x49D0C)
pthreads.library:run()+0x278 (section 1 @ 0x4110)
pthreads.library:ThreadCode()+0x35c (section 1 @ 0x44B8)
native kernel module dos.library.kmod+0x00022d9c
native kernel module kernel+0x0003bd20
native kernel module kernel+0x0003bda0


PPC disassembly:
4f9149e4: 503f5998 rlwimi r31,r1,11,6,12
4f9149e8: 4dd4bcd0 .word 0x4DD4BCD0
*4f9149ec: 4e6ceeb8 .word 0x4E6CEEB8
4f9149f0: 4e663680 .word 0x4E663680
4f9149f4: 4e664440 .word 0x4E664440

System information:

CPU
Model: AMCC PPC440EP V1.3
CPU speed: 799 MHz
FSB speed: 133 MHz
Extensions:

Machine
Machine name: Sam440EP
Memory: 1048576 KB
Extensions: bus.pci


2- Easing curves

Vertical scrollbar doesn't scroll in right direction if you use the mousewheel

Go to top
Re: Beta testing
Just can't stay away
Just can't stay away


See User information
@samo79

The problem is, that those apps, that do not suffer from this behavior, are those that explicitly sets the widget size *before* creation (for instance in the constructor you will see something like this: "mywidget->resize(blabla);"). The apps that suffer from "auto-align crap" all rely on automatic alignment to determine the size and position of the widgets. Also I found out, that the problem is not, that the showing of widgets come at the wrong time, the problem is (simply put), that intuition has weird behavior with hidden windows: When actually showing a hidden window, intuition will first create a window with the *original* size and position and then right after it will "auto align" it to the *actual* (new) size and position. In my eyes it looks like a bug, but I will probably have trouble convincing system devs of this view...

Go to top
Re: Beta testing
Home away from home
Home away from home


See User information
I think the problem is that you display it too fast, before the OS has managed to resize the window.

Try waiting for IDCMP_CHANGEWINDOW before unhideing the window, if your not already doing that.

It might be that need to read the event que to get it to resize, I don't know but worked for me.


Edited by LiveForIt on 2013/6/12 1:37:59
(NutsAboutAmiga)

Basilisk II for AmigaOS4
AmigaInputAnywhere
Excalibur
and other tools and apps.
Go to top
Re: Beta testing
Just can't stay away
Just can't stay away


See User information
@LiveForIt

Yes, that happened to be the consensus on the mailinglist too. Thanks anyway :)

Go to top

  Register To Post
« 1 ... 4 5 6 (7) 8 9 10 ... 40 »

 




Currently Active Users Viewing This Thread: 1 ( 0 members and 1 Anonymous Users )




Powered by XOOPS 2.0 © 2001-2023 The XOOPS Project