![]() |
![]() |
#541 |
Registered User
Join Date: Jan 2019
Location: Germany
Posts: 3,304
|
Sorry, I cannot reproduce this.
|
![]() |
![]() |
#542 |
Paranoid Amigoid
Join Date: Mar 2008
Location: Athens/Greece
Age: 45
Posts: 1,978
![]() |
Ok is this with GuiGfx enabled in iGame prefs?
Because if we disable this (NoGuiGfx checked) it works for us as well. Also can you check please how it performs under native resolutions and 256 colors as FOL said and I could reproduce? So far what we have is that: 1.if we use rtg.library from 3.3.2 installation no problem with GuiGfx exists. 2.if we use native modes with <=128 colors no problem with GuiGfx exists Could the problem be with colors >= 256 in newer rtg.library? As I said, I made a pure stock 3.2.2 installation only with P96 (3.3.3), MUI5, iGame 2.2.1 installed and I could reproduce it just fine under WinUAE. Sorry don’t have any Linux machine to test it under UAE but my A600 with V2-600 and V4SA behave exactly the same as WinUAE so I have no reason to suspect any hardware issue. I can shoot a thorough video if you want to show you exactly what it does in this stock unpatched system. Last edited by mfilos; 15 April 2023 at 17:36. |
![]() |
![]() |
#543 |
Registered User
Join Date: Jan 2019
Location: Germany
Posts: 3,304
|
|
![]() |
![]() |
#544 |
Paranoid Amigoid
Join Date: Mar 2008
Location: Athens/Greece
Age: 45
Posts: 1,978
![]() |
|
![]() |
![]() |
#545 |
Registered User
Join Date: Jan 2019
Location: Germany
Posts: 3,304
|
Thanks, so I guess I finally got it. The trouble here seems that some part of the code just biindly copies the bitmap attributes over back to the input - this has the side effect that the "magic DPaint fix flag" is fed back into the input of the bitmap allocator, which does not know this flag, and then fails.
Why that only happens with IFF and 256 colors I do not know. |
![]() |
![]() |
#546 |
Retro Freak
Join Date: Nov 2001
Location: Slovenia
Age: 51
Posts: 1,665
|
Hi Thomas.
Just tried the new P96ScreenCX utility that doesn't switch screen if a button is pressed. When using a commodity like MagicMenu which displays a right mouse button context menu anywhere on the screen and does this with just a click of right button (you don't need to hold it for the menu to be visible) - and while the menu is being shown you go to the next screen the mouse pointer freezes at the edge of the screen and you need to reboot. Here is a short video (the right mouse button is not pressed when going to the left side of the screen). [ Show youtube player ] |
![]() |
![]() |
#547 |
Registered User
Join Date: Jan 2019
Location: Germany
Posts: 3,304
|
MagicMenu cannot work and does not work - it violates principles of the intuition state machine, and as it is not part of intuition, it can not even know the internal states of intuition. Deinstall MagicMenu, and get rid of stability problems. There is really nothing I can do on my end.
|
![]() |
![]() |
#548 | |
Retro Freak
Join Date: Nov 2001
Location: Slovenia
Age: 51
Posts: 1,665
|
Quote:
|
|
![]() |
![]() |
#549 |
Registered User
![]() Join Date: Aug 2016
Location: Germany
Posts: 122
|
|
![]() |
![]() |
#550 |
Retro Freak
Join Date: Nov 2001
Location: Slovenia
Age: 51
Posts: 1,665
|
|
![]() |
![]() |
#551 | |
Registered User
![]() Join Date: Aug 2016
Location: Germany
Posts: 122
|
Quote:
Changing to Intuition Compatible for the Pull-Down /Pop-up menu solved the problem for me. |
|
![]() |
![]() |
#552 | |
Registered User
Join Date: Jan 2019
Location: Germany
Posts: 3,304
|
Quote:
If you call these functions from the input handler itself (such as what P96ScreenCx does), intuition has another special case as the functions cannot send an even to itself - intuition in that case simply "does it". Now the menu system of intuition(!) of course also requires intuition locks, it locks the screen. If a program now attempts to perform a ScreenToFront() or MoveSizeWindow() or some similar call, intuition is aware that itself holds the lock, and the corresponding activity is not taken because it is currently busy, and the same special case if the call comes from the input chain is also tested. Problem is now if *some other program* such as MagicMenu holds a screen lock, then intuition is not aware that this program actually holds the lock to render or display a menu, and thus it attempts to lock itself - and the result is a deadlock situation. This problem appears to any program that attemps to make screen re-arrangements or window arrangements with intuition calls while MagicMenu has rendered its screen. P96ScreenCx is just one of such programs - the ClickToFront commodity will have exactly the same "problems" because it also calls intution. Or any other program attempting such a modification. There is always a chance that the user has a magic menu open while a program requires an intuition lock as part of a normal intuition function - intuition would cancel itself in such a situation, but any other program it does not know. Long story short: Such programs *cannot* work. Not MagicMenu, nor any other program that attempts to display menus or provide an emulation of the intuition input system - it is outright impossible. Such programs would need to have access to the internal state machine of intuition and would need to modify it such that intuition internally(!) would be able to cancel a menu. Since that state machine is (for good reason) not documented, it cannot work. The only chance to get something like magicMenu is to make it part of the Os and integrate it into the intuition core, or expand the intuition core such that it would provide an interface to such extensions. None of that happened, which is to a major degree because nobody really knows the intuition state machine well enough. Fitting in the iconification gadget was already a major milestone (as simple as it sounds), and Os 4 for example completely botched it. Its (os 4) iconification handling is completely borken, due to lack of understanding of intuition internals... so folks there unfortunately completely misunderstood my "demo code" and took this as "good working code". It wasn't.... Don't mess with menus, you can only break it. |
|
![]() |
![]() |
#553 |
Retro Freak
Join Date: Nov 2001
Location: Slovenia
Age: 51
Posts: 1,665
|
|
![]() |
![]() |
#554 |
Retro Freak
Join Date: Nov 2001
Location: Slovenia
Age: 51
Posts: 1,665
|
Thanks for the explanation Thomas. With the "intuition mode" it works now fine. It is such an old system it is a wonder we get so much stuff working really. And with the development of 3.2.x we get more and more stuff. Also your multi-monitor support is one of these wonders, never thought I would see that working on Amiga across different monitors AND different graphic systems.
Again, thanx for all the effort you put into this. |
![]() |
![]() |
#555 |
Registered User
Join Date: Oct 2012
Location: Germany
Posts: 585
|
|
![]() |
![]() |
#556 | |
Registered User
Join Date: Nov 2015
Location: Italy
Posts: 192
|
Quote:
I think it would be easy to keep having sticky menus work, by implementing the stickiness in MagicMenu differently. Make the actual right mouse button sticky, not the menus. So basically swallow the right mouse button release event in the inputhandler while menu is open. So to the OS and apps/commodities it looks like button is held done the whole time, while in reality it is not. |
|
![]() |
![]() |
#557 | |
Paranoid Amigoid
Join Date: Mar 2008
Location: Athens/Greece
Age: 45
Posts: 1,978
![]() |
Quote:
The application with previous version of rtg.ibrary from P96 (3.3.2) worked just fine as you can see in the screenshot bellow ![]() After replacing rtg.library with the new one from P96 (3.3.3) then the application gives 16 errors upon running which are the following: Code:
01. Out of memory! - Error triggered by Function: LoadBrush - Line: 211 02. Out of memory! - Error triggered by Function: LoadBrush - Line: 150 03. Could not find brush 1! - Error triggered by Function: SelectAlphaChannel - Line: 151 04. This function cannot be used with layers enabled! - Error triggered by Function: Plot - Line: 157 05. Wrong usage/parameters for this command! Read the documentation - Error triggered by Function: EndSelect - Line: 158 06. Could not find brush 1! - Error triggered by Function: GetAttribute - Line: 669 07. Could not find brush 1! - Error triggered by Function: GetAttribute - Line: 670 08. Could not find brush 1! - Error triggered by Function: DisplayBrush - Line: 288 09. Could not find brush 1! - Error triggered by Function: DisplayBrush - Line: 294 10. Could not find brush 1! - Error triggered by Function: DisplayBrush - Line: 265 11. Could not find brush 1! - Error triggered by Function: DisplayBrush - Line: 266 12. Width/heigh values must be greater than 0! - Error triggered by Function: CreateBrush - Line: 267 13. Width/heigh values must be greater than 0! - Error triggered by Function: CreateBrush - Line: 268 14. Width/heigh values must be greater than 0! - Error triggered by Function: CreateBrush - Line: 271 15. Could not find brush 1! - Error triggered by Function: DeleteAlphaChannel - Line: 308 16. Could not find brush 1! - Error triggered by Function: DisplayBrush - Line: 309 ![]() ...giving the following error: Code:
Out of memory! - Error triggered by Function: LoadBrush - Line: 246 Just thought to give a head's up in case you find anything wrong with it (as it might be affecting other Hollywood applications as well). Last edited by mfilos; 21 April 2023 at 09:55. |
|
![]() |
![]() |
#558 |
Registered User
Join Date: Jan 2019
Location: Germany
Posts: 3,304
|
May I suggest a different approach? Please report the issue at the iComp forum. In such a case, Jens can ask you about your order number, and I can provide a beta of 3.3.4 for testing.
|
![]() |
![]() |
#559 |
Paranoid Amigoid
Join Date: Mar 2008
Location: Athens/Greece
Age: 45
Posts: 1,978
![]() |
Sure thing Thomas.
I just did, so waiting to hear from Jens now... Last edited by mfilos; 21 April 2023 at 15:08. |
![]() |
![]() |
#560 | |
Paranoid Amigoid
Join Date: Mar 2008
Location: Athens/Greece
Age: 45
Posts: 1,978
![]() |
Quote:
I already gave my Order Number in that post (to save Jens time from asking) but still no joy... |
|
![]() |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
Thread Tools | |
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
P96: What's the right way to do X? | Warty | Coders. General | 2 | 21 December 2020 00:00 |
Providing 2 fire button support / cd32 joypad support | amigapd | request.Other | 0 | 13 July 2015 17:20 |
Portaudio support (was: WinUAE support for ASIO drivers) | Amiga1992 | support.WinUAE | 57 | 28 March 2009 21:15 |
Classic WB P96 | Anubis | project.ClassicWB | 5 | 08 May 2006 14:30 |
amiga-news.de: Collected software-news | Paul | News | 0 | 14 November 2004 15:50 |
|
|