AGS 3.0 Final - it's been a long road

[Rui 'Trovatore' Pires]

That's why Alt-X always quits any game, and can't be changed.

Actually, it can. But thankfully, no one bothers to... but it can. I love the current system, but Alt-X as an example shows that people have more sense than we may think.

Of course, it's not immediately obvious that you *can* change it. That might be related.

[Shane 'ProgZmax' Stevens]

...and I've disabled alt-x before .  Being able to design your own wrapper in a sense is one thing, being able to design your own setup that adds functionality is something else, something a wrapper gui for the setup can't do.

[SupSuper]

I don't know if this question has come here before. But I have Windows Vista Premium and it can't read no longer chm help file because Windows Vista don't support it any longer.

Erm, that can't be right, CHM is Microsoft's latest help format, every major application uses it. Vista only dumped support for the old Win9x HLP format (probably to help push CHM forward), and that can still be restored from Microsoft's site.

Wonkypedia says: "Microsoft Compiled HTML Help is a proprietary format for online help files, developed by Microsoft and first released in 1997 as a successor to the Microsoft WinHelp format. It was first introduced with the release of Windows 98, and is still supported and distributed through Windows XP and Vista platforms. In 2002, Microsoft announced some security risks associated with the .CHM format, as well as some security bulletins and patches. They have since announced their intentions not to develop the .CHM format further, and will be moving to a new generation of Windows Help called Microsoft Assistance Markup Language in the Windows Vista operating system."

MAML isn't even widespread though. In any case, my Vista opens CHM, so

Anyways, I think it's about time I ask the big question here, now that the editor is near-final: what's the new plugin API like? I suspect ingame-plugins will still work the same (specially for backward compatiblity), but what about editor plugins? Given everyone's ranting about the interaction editor one way or another, I feel itchy to come up with a replacement.

[Pumaman]

Issues with DirectDraw mode: Room fades don't match the room resolution (in my case, the rooms switch from 320*200 to 320*240 while fading)

I can't reproduce this problem? Are you using the latest beta -- there were some problems with fades in earlier betas.

What exactly is happening when you say it's switching from 320x200 to 320x240? Does the screen all stretch out or something?

[SupSuper]

Issues with DirectDraw mode: Room fades don't match the room resolution (in my case, the rooms switch from 320*200 to 320*240 while fading)

I can't reproduce this problem? Are you using the latest beta -- there were some problems with fades in earlier betas.

What exactly is happening when you say it's switching from 320x200 to 320x240? Does the screen all stretch out or something?

Exactly. Though I just noticed AGS switches to 640x480 when I try to run on 320x200 (although it looks different from actually running on 640x480), so maybe it's a side-effect of that? It only happens on fade-outs, too. Here's a screenshot.


Another bug is that moving the cursor along the bottom or right edges of a dialog selection screen reveals the background below.

[Pumaman]

Exactly. Though I just noticed AGS switches to 640x480 when I try to run on 320x200 (although it looks different from actually running on 640x480), so maybe it's a side-effect of that? It only happens on fade-outs, too. Here's a screenshot.

Hmm, I still can't reproduce this. Does it only happen with the DX5 driver? Would it be possible to upload a sample of this problem for me to take a look at?

Another bug is that moving the cursor along the bottom or right edges of a dialog selection screen reveals the background below.

Thanks, I've reproduced this and will get it fixed.

[subspark]

Another bug is that moving the cursor along the bottom or right edges of a dialog selection screen reveals the background below.

That bug was present in the Demo i sent you Chris. However I thought it was related to the GUI stuff I mentioned.
Just FYI.

Cheers,
Paul.

[Recluse]

I just tried this out with both DX5 and D3D and it was fine. What problem are you getting? Is the screenshot corrupt, or is it simply not creating the file? Did you remember to look in the save game folder?

I'm running everything from the "compile" folder and it's not creating the screenshot.

I hadn't checked the save game folder, because I didn't see it in the compile folder, is it stuck over in My Documents? I'm not at my machine right now or I'd have found it. I'll look for it once I get home.

BTW: Good job on adding the "Remember Last Folder Used" option, that was getting to be quite annoying for me as well.

[mchammer]

I'm running everything from the "compile" folder and it's not creating the screenshot.

I hadn't checked the save game folder, because I didn't see it in the compile folder, is it stuck over in My Documents?

Both of them, saved games and screenshots are in My Documents.

[Recluse]

Well don't I look silly

Thanks for the help... 

[monkey0506]

I don't think it's been brought up, and at this late stage it probably wouldn't make the 3.0 cut anyway, but the idea's crossed my mind several times about static extender methods.

A specific example of why one might need this would be the current implementation of the Room structure which is completely static, making it impossible to implement extender methods for the structure.

I figure it probably wouldn't be something easy to implement, however I think it surely must have at least equal importance to the request for extender methods in the first place (due to the implementation of certain AGS structures like Room and Game).

I understand it probably won't be put in this version, but consideration for a future version would be much appreciated.

All the interactions were imported, but the events list is empty.

That's bizarre, has anyone else experienced anything similar?

Actually yes, I just deleted one of my cursor modes and following that my event list no longer showed any events. They all showed prior to deleting the cursor, but not after. In fact, even saving and loading the game in the editor didn't restore it. I could only see the events again after restoring the cursor.

I was trying to figure out how events now register the cursor modes. Idealistically we should be able to have as few or as many cursor modes as we would like and all event panes should reflect this by making events available for each mode. About the only hard-coded cursor that really makes sense (to me) being hard-coded would be the wait cursor, though I think if an option were made available to set up the wait cursor (with the ability to set it to "no cursor") it would be even better.

This of course is something that has been hard-coded into AGS for some time, though it would make games requiring fewer than the hard-coded 7 cursors easier to script (without having to force/emulate modes) while allowing games requiring more cursors easier to implement (having specialized events available instead of having to go through the catch-all of "Any click").

This is just my two cents on the matter and further opinions are welcome (though if too many opinions show up the mods may have to take the discussion to a separate thread).

Thanks.

[khnum]

Sorry if i always bring up the same kind of problems, but shouldn't ttf antialiasing work even without a solid background now that we have D3D? or it is a different problem from the alpha blended antialiasing?

[Pumaman]

I don't think it's been brought up, and at this late stage it probably wouldn't make the 3.0 cut anyway, but the idea's crossed my mind several times about static extender methods.

That's a reasonable idea, but yes it's probably too late for 3.0 now.

Actually yes, I just deleted one of my cursor modes and following that my event list no longer showed any events. They all showed prior to deleting the cursor, but not after. In fact, even saving and loading the game in the editor didn't restore it. I could only see the events again after restoring the cursor.

Aha!! Thanks, I've reproduced the problem and will get it fixed. In the meantime the workaround is, as you suggest, to create cursor modes 8 and 9 again which will cause the events list to reappear.

I was trying to figure out how events now register the cursor modes. Idealistically we should be able to have as few or as many cursor modes as we would like and all event panes should reflect this by making events available for each mode. About the only hard-coded cursor that really makes sense (to me) being hard-coded would be the wait cursor, though I think if an option were made available to set up the wait cursor (with the ability to set it to "no cursor") it would be even better.

Yes, I tend to agree. This is something which I am working towards (it's supposed to already do this with modes 8 and 9, only making the events available if the modes exist -- but due to a bug with that at the moment, all the events disappear if you don't have a cursor mode 9 )

Sorry if i always bring up the same kind of problems, but shouldn't ttf antialiasing work even without a solid background now that we have D3D? or it is a different problem from the alpha blended antialiasing?

No, this is a different issue. Due to TTF rendering being a relatively slow thing to do, AGS renders the text to an internal dynamic sprite once, and then blits that sprite to the screen each game cycle. This means that the text gets antialiased against the sprite background rather than the room background, thus causing the issue that we have at the moment.

[Pumaman]

Ok, beta 13 is now up which addresses various issues.

I'm not planning on adding anything more to AGS 3.0 from this point, only bug fixes and very minor enhancements will be considered. The next release should be a Release Candidate.

I'm still unable to reproduce the crash loading a room that has been reported by a few people, if anyone manages to get this consistently please let me know.

[OneDollar]

When I try to play my FLIC animation using the run command in the editor under DirectDraw 5 I just get flickering bands of the animation. When I try it under Direct3D 9 I get the following error...

An exception 0xC0000005 occurred in ACWIN.EXE at EIP = 0x00000000 ; program pointer is +1101, ACI version 3.00.953, gtags (0,5)

AGS cannot continue, this exception was fatal. Please note down the numbers above, remember what you were doing at the time and notify CJ on the Tech forum.

in "room5.asc", line 9

Line 9 is:

Code: ags

PlayFlic(0, 1);

When I was using AGS 2.72 it all worked fine.

<<EDIT>>
Oh, and I am using Beta 13

<EDIT 2>>
And another issue - when running under Direct3D 9 and using the 'disolve' for room transitions I get a mismatched colour depths error, though it works under DirectDraw 5. My game is 16-bit

[GarageGothic]

* Added DynamicSprite.CopyTransparencyMask
* Added extra parameter to DynamicSprite.CreateFromExistingSprite to allow you to specify whether the alpha channel is copied or not
* Added COLOR_TRANSPARENT option to DrawingSurface.DrawingColor property to allow you to draw transparent areas onto the image

CJ, will you marry me?

[subspark]

NO! Take me CJ! I'm not a gold digger like GarageGothic! 

[Kweepa]

Awesome progress on this! Debugger! D3D! RawDrawing to sprites!

I'm still having the problem where Exiting the editor from the "Welcome To AGS!" dialog leaves the AgsEditor.exe process running using 100% CPU. This time I'm using Win2k on an AMD Sempron (single core). It's quite intrusive - I tried running Process Explorer to see what threads were still alive but getting properties on the process locked up Process Explorer!

The Ags3d plugin works fine. It does highlight a workflow problem - I have to copy external files from the Compiled directory into _Debug. Any chance of that being done automatically? Or of running a post-build or pre-debug command (so I can run a batch file)?

Wish list:
* multi file search (and replace)
* a property pane that allows checkboxes (True/False is much harder to read)
* a drop down selection box that drops down on first selection, wherever you click (rather than requiring a selection, then a second click on the little arrow on the right) - like the old editor
* variable display in the debugger
* F10 to step over

[monkey0506]

I don't think it's been brought up, and at this late stage it probably wouldn't make the 3.0 cut anyway, but the idea's crossed my mind several times about static extender methods.

That's a reasonable idea, but yes it's probably too late for 3.0 now.

Fair enough. I figured it was probably too late to mention it for this version. Thanks for the consideration though!

I was trying to figure out how events now register the cursor modes. Idealistically we should be able to have as few or as many cursor modes as we would like and all event panes should reflect this by making events available for each mode. About the only hard-coded cursor that really makes sense (to me) being hard-coded would be the wait cursor, though I think if an option were made available to set up the wait cursor (with the ability to set it to "no cursor") it would be even better.

Yes, I tend to agree. This is something which I am working towards.

Glad to hear that I'm not just being stupid. Unless you're just being equally stupid, but if that's the case I'd glad to be thought of as having been as stupid as you! 8)

Thanks for the consideration on this request as well. Oh, and thanks for the FontCount/MouseCursorCount properties. Those should be useful.

[subspark]

Fixed Out of Memory crashes if you used a non-text window dialog options GUI

AGS now shifts the bottom row of pixels of a Dialog GUI background image 1 line above the top.

Also perhaps for those of us who wish to save space on background art, could we perhaps have an option to turn off all elements of a GUI including the background image. That way we don't have to paint in an ugly black square on our background art to emulate the GUI going black. (The background color and border should be the only elements visible).

EDIT: Could the developer choose what color to 'Grey' the inventory out with? I can't find an easy way to 'Brown' the inventory out without having all my inventory items disappear due to using the GUI background as the brown-out result.

Cheers, Paul.