Re: [ql-users] New WMAN (was: GD2)
??? 14/11/2002 1:12:14 ??, ?/? Wolfgang Lenerz [EMAIL PROTECTED] ??: Are the sources available? BTW, QPTR pratically doesn't need to be updated - I've done some work on that, and, provided one or two new functions are used, to reduce 24 bit colours to 16 bit and set the correct MS bit, you can use it to produce high colour PE windows from Basic. There is a distinctive problem there provided you've used just what I call bit cropping... ie removing by brute-force the highest order bits from each colour... The problem (on which I ran into while developing my all around Windoze bmp converter is that once you start chopping bits left and right you are killing the displayed picture... that's not important for sprites for example (well for me it is) as for artificial images you can have the colours pre-arranged so they won't use the whole 32 or 24 bit gamut, but nonetheless I think that we should come to an agreement on how this function should be performed... there are plenty of algorithms to go around, but the fastest (albeit not the most correct ;-) is to scan your bitmap and statistically adjust it... this can be done extremely fast if using the FPU (anybody use S.G and G.G's FPUFNs ? ) but for slow QL systems (ie Aurora - which hopefully will have it's rendition of the GD2 once I get a: The SGC and b: The sources) this is impossible to make at a decent speed :-(... Maybe we should restrict colour rounding to high colour systems only (All High-Colour systems, including QPC do have FPU support or they are sufficiently fast to perform this)... I'm trying to finish an article for QL Today for that We're all waiting :-) Just a couple of questions here: 1 - Does button mean the buttons in the button frame, or also (other) loose items? I call a button every loose item. My 2 (euro) cents (pennies) regarding WMAN/PE in general. With the major updates that happened to the PE, using a different Window Manager should become redundant I would think It would be a good idea on restructuring it a bit (without losing compatibility) to accept add-ons... Since the PE in general provides a good enough (and known enough most important) framework a new gui (for which I strive for so long) will be a lot easier to construct. AFAIK, attaching bitmaps (ie skins) and high-colour (or should I say multi-colour) icons (sprites) is already possible so by polishing it up a bit every one could build his own window manager in a sense... As for my GuiDemo (Norman, Wolfgang and others have seen it so far), once I started using FPU-Presence benefiting instructions the whole thing is flying! (and on high-resolution, high-colour as well). Even interpreted S*Basic is burning rubber... (Generated some 5 Hi colour, 3D shaded Windows in about 15 minutes which is NOT BAD at all)... (Regular speed without FPU is about 2 windows on a Q40) Now if mine can do this, imagine how fast the really optimised (and written in assembly) PE can go Phoebus
Re: [ql-users] New WMAN (was: GD2)
On Wednesday, November 13, 2002, at 11:50 PM, Phoebus Dokos wrote: My 2 (euro) cents (pennies) regarding WMAN/PE in general. With the major updates that happened to the PE, using a different Window Manager should become redundant I would think It would be a good idea on restructuring it a bit (without losing compatibility) to accept add-ons... Interesting idea, to allow for add-ons. Would it be possible to make it so that not only would it allow add-ons but also allow 'replacement' functions, just like S-Basic replaces old with like named pieces? That would allow updates or modified window routines to be written that a user could choose - talk about custimization. Also, would it be easy to add additional structures/objects within PE. I am using several custom pieces in QDT that I wrote. It would be nice to eventually make them part of the standard PE setup (I said eventually - QDT comes first!). Jim Hunkins
Re: [ql-users] New WMAN (was: GD2)
Wolfgang Lenerz wrote: I hope so. I want to tie up some lose ends like this WMAN thing before concentrating on the university project. After that I hope others are sufficiently familiar with the code to start contributing, too (it's BTW not only SMSQ/E. EasyPtr should be updated as well, for example. Any volunteers?). Are the sources available? IIRC somebody asked Albin and he said that he would give them away. BTW, QPTR pratically doesn't need to be updated - I've done some work on that, and, provided one or two new functions are used, to reduce 24 bit colours to 16 bit and set the correct MS bit, you can use it to produce high colour PE windows from Basic. Good. I've never used the QPtr BASIC interface, so I've no idea when it comes to that. Just a couple of questions here: 1 - Does button mean the buttons in the button frame, or also (other) loose items? Loose items. But another entry for the button frame thingies is a good suggestion. 2 - I often give my application subwindows different colours to my info subwindows. Ok. How would you call the entries? Marcel
Re: [ql-users] New WMAN (was: GD2)
??? 14/11/2002 3:46:51 ??, ?/? James Hunkins [EMAIL PROTECTED] ??: Interesting idea, to allow for add-ons. Would it be possible to make it so that not only would it allow add-ons but also allow 'replacement' functions, just like S-Basic replaces old with like named pieces? That would allow updates or modified window routines to be written that a user could choose - talk about custimization. I can't see why not :-) Then again, I've yet to see the sources :-) Nonetheless first: all guesswork from Desktop developers (HIIN) would be eliminated and second: Why re-invent the wheel? Also, would it be easy to add additional structures/objects within PE. I am using several custom pieces in QDT that I wrote. It would be nice to eventually make them part of the standard PE setup (I said eventually - QDT comes first!). That would be possible if using the add-on idea... Maybe it's already possible... I can't tell (yet). and not only that, that would shorten your QDT development by months :-)! Phoebus
Re: [ql-users] New WMAN (was: GD2)
On 14 Nov 2002, at 2:50, Phoebus Dokos wrote: There is a distinctive problem there provided you've used just what I call bit cropping... ie removing by brute-force the highest order bits from each colour... err no, the lowest oder bits of course... Anyway, the problem you describe is, IMHO, irrelevant here, since you are designinbg a wiondow, and can thus choose any colour scheme you like. It's not a question of displaying a bitmap in this respect... I'm trying to finish an article for QL Today for that We're all waiting :-) You mean, you are all programming with QPTR...? With the major updates that happened to the PE, using a different Window Manager should become redundant I would think It would be a good idea on restructuring it a bit (without losing compatibility) to accept add-ons... Since the PE in general provides a good enough (and known enough most important) framework a new gui (for which I strive for so long) will be a lot easier to construct. AFAIK, attaching bitmaps (ie skins) and high-colour (or should I say multi-colour) icons (sprites) is already possible so by polishing it up a bit every one could build his own window manager in a sense... Actually, it is pretty easy to do so, within some limits. I have here a small progs I use for keeping my dvd colection. This displays the dvds (I scanned the sleeves) and uses, for example, a green triangle as a button (that lights up when the pointer is over it). this was done entirely in Basic - only, it uses RPTR instead of RD_PTR(T) and thus is responsible itself for handling things like pointer in/out of the window/loose item etc... I was able to display higher colours even before Marcel changed WMAN (it also works on the Q60, for example) since it doesn't use WMAN, just the Pointer interface... Wolfgang - www.wlenerz.com
Re: [ql-users] New WMAN (was: GD2)
On 14 Nov 2002, at 13:54, Marcel Kilgus wrote: 2 - I often give my application subwindows different colours to my info subwindows. Ok. How would you call the entries? Info window colours... Appsub window colours... Wolfgang - www.wlenerz.com
Re: [ql-users] New WMAN (was: GD2)
In message [EMAIL PROTECTED], Marcel Kilgus [EMAIL PROTECTED] writes Wolfgang Lenerz wrote: I hope so. I want to tie up some lose ends like this WMAN thing before concentrating on the university project. After that I hope others are sufficiently familiar with the code to start contributing, too (it's BTW not only SMSQ/E. EasyPtr should be updated as well, for example. Any volunteers?). Are the sources available? IIRC somebody asked Albin and he said that he would give them away. I would like to suggest that the sources be given to Rich Mellor who could work on them and maintain a cheap commercial release for them. -- Roy Wood Q Branch, 20 Locks Hill Portslade. Sussex. BN41 2LB. UK Tel : +44 (0)1273 386030 Fax : +44 (0)1273 430501 (New number!) Mobile +44(0)7836 745501 Web : www.qbranch.demon.co.uk
[ql-users] New WMAN (was: GD2)
TonyTebby wrote: He started getting more into the concept of GD2 !! How can anyone do this - has this man a brain the size of a planet? I can't understand it and I wrote it! I just said I'm getting more into it. Not how much :-) I gather information here and there and it's often enough to do some changes (like the new buffer-less background colour code or the 16bit shadow), but that's it. I wouldn't even dare thinking about other proposals from the list, like background redrawing. If I remember right, the sprite cache routines check the address of the sprite (which is likeley to be re-used either if you are editing the sprite or if you are loading the sprite into the chank of memory) and the sprite cache version number (pto_vcch in the sprite header) Argh, I have totally overlooked that you introduced a cache version (so much for my knowledge). Of course, this way it's easy. And it even works! Thanks for the hint. WMAN should really be updated to GD2 - any volunteers? That'd be me again, then... In fact I'm almost done. Since yesterday all routines accept the new colour word definition, only a few more lines for the system palette are needed. And some sensible defaults for that palette (which'll probably be the hardest task). While I'm at it: last call for comments! I'll attach below the new WMAN specification the way I did implement it. Any suggestions for changes should be done exactly NOW before it gets official! Until then would someone like to modify the WMAN scroll sprite generation code to use a new cache version number instead of 0 each time? There must surely be a byte spare somewhere in the (a2) block to store the last value used. Yes, there is, I called it ws_scach. And yes, I'm not good at names. Will this open source software idea ever catch on? I hope so. I want to tie up some lose ends like this WMAN thing before concentrating on the university project. After that I hope others are sufficiently familiar with the code to start contributing, too (it's BTW not only SMSQ/E. EasyPtr should be updated as well, for example. Any volunteers?). Marcel - | | | Vector $7C WM.SETSP | | | | Set system palette entries| | | | Call parameters Return parameters | | | | D1.w start index D1 preserved| | D2.w number of elements D2 preserved| | D3+ all preserved| | | | A0 A0 preserved| | A1 pointer to palette entries / 0 A1 preserved| | A2 A2 preserved| | A3 A3 preserved| | A4 A4 preserved| | A5 not used by any routine | | A6 not used by any routine | | | | Error returns: | | IPARIllegal index number / invalid number of elements | | | Set the entries of the system palette to the values in the buffer, beginning with the index in D1 (counting from 0) and ending with the index D1 + D2 - 1. If A1 = 0 then the entries are taken out of the default table. Otherwise the buffer must hold an array of words with the colour values of the different items. The colour format is the standard WMAN colour format as described elsewhere. | | | Vector $80 WM.GETSP | | | | Read system palette entries | | | | Call parameters Return parameters | | | | D1.w start index / -1D1.w preserved / item count | | D2.w number of elements D2 preserved
Re: [ql-users] New WMAN (was: GD2)
On 14 Nov 2002, at 1:35, Marcel Kilgus wrote: WMAN should really be updated to GD2 - any volunteers? That'd be me again, then... In fact I'm almost done. Since yesterday all routines accept the new colour word definition, only a few more lines for the system palette are needed. And some sensible defaults for that palette (which'll probably be the hardest task). Three cheers for Marcel! Please be reminded that his work will benefit all platform physically capable of it! Yes, there is, I called it ws_scach. And yes, I'm not good at names. You're pretty good at everything else I hope so. I want to tie up some lose ends like this WMAN thing before concentrating on the university project. After that I hope others are sufficiently familiar with the code to start contributing, too (it's BTW not only SMSQ/E. EasyPtr should be updated as well, for example. Any volunteers?). Are the sources available? BTW, QPTR pratically doesn't need to be updated - I've done some work on that, and, provided one or two new functions are used, to reduce 24 bit colours to 16 bit and set the correct MS bit, you can use it to produce high colour PE windows from Basic. I'm trying to finish an article for QL Today for that System palette entries (...) Just a couple of questions here: 1 - Does button mean the buttons in the button frame, or also (other) loose items? 2 - I often give my application subwindows different colours to my info subwindows. I'm not sure whether I'm alone in that. If not, we (you!) should, perhaps, provide for it. Wolfgang