Jump to content

Photo

Zelda Classic 2.53.0 Beta 10 (Official)


  • Please log in to reply
8 replies to this topic

#1 ZoriaRPG

ZoriaRPG

    ZC Contributor & Wiki Editor

  • Members
  • Gender:Unspecified
  • Location:Prydon Academy

Posted 01 October 2017 - 04:23 AM

Here is Zelda Classic 2.53.0 Beta 10 for Windows

Updated 2nd October, 2017 at 16:22GMT with --fixdrawint --fixgamecombo

 

Here is ZC 2.53, Beta 10, which fixes the following, over the last major beta build:

  • Doors on dungeon maps when using diagonal movement, nor operate more smoothly.
    • This includes some slight adjustments to ensure that if the left and right doors have a solid obstacle in front of them, that Link cannot open them around the obstacle.
  • Fixes the issues with Link in water, flying through land tiles when he is hit by his own weapons.
  • Adds additional filetyles to the Import Script dialogue (.zh, .zs, .zlib, .zscript), in addition to the extant .z extension.
    • This will make it easier to use the import dialogue for header files, and other filetypes.
    • Note: We purposely avoided showing files with an extension of .txt, as to avoid making the use of .txt for a ZScript filetype some kind of absurd, and lazy standard.
  • The use of .zs, .zlib, and .zscript, are to aid in the filetype conflict for .z, which is used for ZLIB Archive files on *nix systems.
  • Fixes an issue where scrolling warps with different DMap IDs for source->dest, do not show a map that Link does not have; and that ZC does not show an incorrect map, on the passive subscreen.
  • Fixes compatibility issues with wizzrobe and other floating/flying enemies spawning on solid combos (when they should not), or flying through walls.
  • It also adds a warning dialogue when you open the ZConsole via the menus:

DebugConsoleWarn.png



See the Changelog below for more information. As before, the Build ID for this release is '31', to distinguish between 2.50.3RC1 (Build ID '30') and earlier ZC releases.
 
This thread replaces the previous Beta thread for 2.53.0.

We would appreciate bug reports on any of the following:

  • Playing quests made in 1.92, 2.10, 2.50.x and reporting any compatibility bugs.
  • Verification that scripts made in 2.50.0, 2.50.1, 2.50.2, and 2.50.3RC1 still work.
  • Testing that scripts compiled in 2.53.0 work properly in 2.53.0.
  • Tests that scripted quests compiled and saved in 2.53.0 work proper;y on 2.50.2 (and/or 2.50.3RC1).
  • Bug reports on any changes, or new features.
  • Bug reports on joystick configuration.

Please note: The std.zh files in this package may not represent what will be included in the final release. It is likely that the std.zh changes will be pushed forward to 2.54 or later, and not included; but they will be available as a separate download. The latest versions of all common ZScript headers will be included in the final release.

This release is built on the Allegro library, version 4.4.3, with some special fixes applies, and it includes new Allegro library files.

Zelda Classic and ZQuest v.2.53.0 API Specification (PDF, Rev. 1.2)


ZLaunch Repository on GitHub

Zelda Classic Changelog for v.2.53.0 Beta 10



ZLaunch Changes v2.6.1

ROMview Changelog Version 2.6.1

Added Win32 platform guards to ZCL video driver selection, and bumped version to 2.6.1. ( ZoriaRPG, 29th August, 2017 )

Added the ability to set the allegro video mode on Windows (fullscreen, and windowed drivers).

Added 'Drivers' tab. ( ZoriaRPG, 28th August, 2017 )
Added ROMView launching button to main panel. ( ZoriaRPG, 27th August, 2017 )



Thank you for testing, and for your feedback!


Edited by ZoriaRPG, 02 October 2017 - 09:26 AM.


#2 Joelmacool

Joelmacool

    JMC

  • Members
  • Real Name:Joel
  • Gender:Female
  • Location:Somewhere...

Posted 01 October 2017 - 04:45 AM

Looks good but...

BH8GKuTKSueYOnUzWiLPSg.png
I'm pretty sure "uption" is a typo.

 


  • Neptune, Shane and Eddy like this

#3 ZoriaRPG

ZoriaRPG

    ZC Contributor & Wiki Editor

  • Members
  • Gender:Unspecified
  • Location:Prydon Academy

Posted 01 October 2017 - 06:22 AM

Looks good but...

BH8GKuTKSueYOnUzWiLPSg.png
I'm pretty sure "uption" is a typo.

 

 

Fixed.


  • Joelmacool likes this

#4 ZoriaRPG

ZoriaRPG

    ZC Contributor & Wiki Editor

  • Members
  • Gender:Unspecified
  • Location:Prydon Academy

Posted 02 October 2017 - 09:32 AM

Updated 2nd October, 2017 at 16:22GMT with --fixdrawint --fixgamecombo to include a reversion of the entire set of Game->Get/SetCombo* functions. This should un-break quests that rely on them.

 

The patch also fixes the output of DrawInteger() with zero decimal places; and both the array size, and the permitted values of Game->LKeys[] and Game->LITems[]. The new array sizes are [512] and the assigned values have a valid range of 0 to 255.


Edited by ZoriaRPG, 02 October 2017 - 09:35 AM.


#5 Lüt

Lüt

    Germanize

  • Members
  • Real Name:Steve
  • Gender:Male
  • Location:Chicago

Posted 05 October 2017 - 06:37 PM

OK, late as always.

 

Regardless, wrapping up prior issues:

 

NES Locked Door Sensitivity: I believe this is fully fixed. I tried all arrangements of the test quest with both 8-way and 4-way movement enabled, in both 2.53 B10 and 2.50.2, and their performance is entirely consistent in every arrangement.

 

Debug Console Note: Great, exactly what it needs to be. As long as everybody's aware of precisely how it functions, because that one can be a rude awakening.

 

Minimap Updating: I see it briefly disables during the transition between DMaps when in the form of a default dungeon map item. Fair enough, I think. From what I've tried, transitions between pre-drawn custom maps work as smooth as always (no disappearing), and that applies to multi-segmented overworlds, as well as transitions between overworld and interior minimaps. I'll try custom drawn minimaps between NES dungeons just to be sure, but given everything else so far, I don't imagine any trouble.

 

I also noticed B9 had an error compiling std.zh, but B10 seems to have fixed that.

 

And now, the last of the new issues:

 

Ladder Remains After Farore's Wind: [Demo Quest]

 

If you use the ladder to cross water, then use Farore's Wind while standing on the ladder, the ladder will remain displayed in its position from the previous screen until the opening screen wipe on the warp screen has completed, then it will disappear directly afterward.

 

Open the demo, grab the ladder, cross up to the next screen, grab Farore's Wind, stand on any of the 3 pools of water to bring out the ladder, then warp back to the first room. The ladder will temporarily display in its previous position. It also works on the same screen, so try it on the river in the first screen as well.

 

Far as I can tell, this seems to affect only the ladder. Leaving out bait, then warping back to the start room, doesn't cause a lingering sprite. I guess it's because the ladder is "active" or "in use" while the bait is just sitting there? I tried it with other stuff like the boomerang or arrows or bombs, but the Farore's Wind animation takes long enough that they all return or disappear or blow up before it completes, so I'm not sure if any other "active" sprite may be subject to the same lingering sprite issue.

 

Trap Passthrough: [Demo Quest]

 

This has been brought up before, but I don't believe it's ever been submitted as an issue. Basically, traps ignore combo barriers where the top half is solid, but the bottom half isn't, thus passing through what should be a combo that blocks them.

 

Actually, all enemies do this, so I probably should have named this "Enemy Passthrough." However, I first noticed it with traps, and since they don't obey Flag 96/97 (No [Ground] Enemies), they're the primary concern.

 

In the demo, there's 2 rooms with half-solid rectangular combo borders built around a set of 4 traps (one in each corner). The first room has the combo borders on the outside of the inner tiles; walk around the entire rectangle and see the traps pass through only the upper barrier. The second room has the combo borders on the inside of the outer tiles; walk about the entire rectangle and see the traps pass through only the lower barrier. If you head down, you'll walk through the same 2 rooms again, this time with enemies in addition to the traps; see the enemies pass through the upper and lower barriers along with the traps. If you head down again, you'll walk through these 2 rooms once more, this time with Flag 96 on the upper and lower barriers; see the enemies stay within the barriers but the traps pass through.

 

It's popped up in various quests and LPs where it seems obvious that the author didn't intend for such behavior.

 

Moreover, I found an interesting entry in the Zelda Classic v1.90 Development History page from April 2001:

 

"I hope to fix the enemy walking bug (an enemy can walk through a block unless the bottom left and right corners have walk flags in them) soon."

 

Looks like, 16 years later, this bug is still waiting to be fixed.

 

Shop Stays On Item Message (Script): [Demo Quest]

 

I'm not even sure if this is an issue with ZC. It happens while using Joe123's Item Pick Up Message script, so maybe it's the script that should take account for this behavior.

 

In the demo, you get 2 shops. The one on the left has 2 items that play pickup messages. The one on the right has 2 items that don't use the pickup message script at all. A message and message-less wooden boomerang is in each shop, along with a different item. If you buy an item in the message shop, the shop will start to flicker out as usual, but then stay on the screen once the flicker completes. By comparison, if you buy an item in the message-less shop, it flickers out and disappears as usual.

 

However, this only happens with the "Messages Freeze All Action" quest rule enabled. If you open the quest and disable this rule, the message shop will flicker out and disappear as usual when you buy an item.

 

Again, I have no idea if it's a ZC inconsistency or an oversight on behalf of the script author, but given that 2.54 (or was it 2.55?) is aiming to optimize script compatibility/functionality/etc., I just thought I'd put this out there, if only for future reference.

 

OTHER: I've managed to identify a crash, created twice in a row, while moving from one room to the next in a specific interior DMap. This never happened in over a year testing this map in 2.50.2, then suddenly two times in 2.53 B10. However, I'll have to dig deeper to confirm what the issue may be. I think I have an idea though (this room has some specific FFCs [non-scripted] that no other rooms have).

 

Anyway, that wraps up my list of bugs/issues/suggestions for ZC. I know you're not dealing with 8-way water issues in this build, so all that leaves for me is testing and confirming everything reported so far when the next beta hits.



#6 CSD

CSD

    Newbie

  • Members
  • Real Name:Rob
  • Gender:Male
  • Location:Canada

Posted 06 October 2017 - 12:01 AM

"Ladder Remains After Farore's Wind"

I think this occurs with more than just FW? Probably any method of warping away while the ladder is active. Also, if you spawn on the new screen at the same location as the ladder, the ladder doesn't go away and is entirely functional (so it's potentially a softlock if you spawn in a place where you would normally need to move perpendicularly to the ladder direction). Since it's not just a graphical bug, it's conceivable I think that somebody could be unknowingly relying on it. I think also, if you remove the ladder from the inventory with a script, the graphic disappears, but again it remains functional until you finally step off of where it was. It was a pain to script around these bugs.

 

"Trap Passthrough"

IIRC this occurs for all enemies and is consistent with Zelda 1. You can see the enemy on the level 1 entrance sometimes walk through the door into the tree.


Edited by CSD, 06 October 2017 - 04:46 AM.


#7 Lüt

Lüt

    Germanize

  • Members
  • Real Name:Steve
  • Gender:Male
  • Location:Chicago

Posted 06 October 2017 - 03:14 AM

Maybe more than FW, but I didn't have a way to test it, at least that I could think of. I guess you could put a warp gate on the other side of a river? I wonder if the warp type makes a difference too. I should check it eventually. Doesn't seem like a huge priority if it's all related to the same processing method though.

 

And I'm not sure we can really know exactly how Z1 was constructed in regard to the L1 door, or any other cave entrances for that matter. ZC only came about from estimating Z1's behavior, not deconstructing its original source code and replicating it precisely, so it's possible to mimic Z1's behavior using a different approach. That is to say, we can't load up the Z1 ROM and check the walkability grid for the cave entrances. ZC has the top half flagged as solid, and with the current code that allows this upper-half enemy passthrough, it faithfully recreates Z1 behavior. But maybe the cave entrances should have been fully walkable, and the current code "fixed," as the guy who wrote the change log implied way back then? The end result would be the same, and you wouldn't have enemies passing through solid barriers either. At the least, as you walk around the borders in the demo, you'll see it's inconsistent at best, and with no clear reason for any designer to expect different behavior along one border than they would any of the other 3 borders. But, with the possibility of having to edit cave entrance walkability to account for a change in this behavior, I suppose this is heading toward quest rule territory.



#8 Lüt

Lüt

    Germanize

  • Members
  • Real Name:Steve
  • Gender:Male
  • Location:Chicago

Posted 07 October 2017 - 02:52 PM

OTHER: I've managed to identify a crash, created twice in a row, while moving from one room to the next in a specific interior DMap. This never happened in over a year testing this map in 2.50.2, then suddenly two times in 2.53 B10. However, I'll have to dig deeper to confirm what the issue may be. I think I have an idea though (this room has some specific FFCs [non-scripted] that no other rooms have).

OK, isolated it.

Angled Walk Corner Crash: [Demo Quest]

This is directly related to 8-way movement.

If you exit the bottom left corner of an open screen by walking diagonally downward-left, and you walk into the precise corner of the screen - like, the exact corner pixel - ZC can't seem to make up its mind whether to scroll sideways or downward, and simply decides to crash instead of scroll at all.

This only affects downward-left screen exits - downward-right, upward-left and upward-right all work fine.

DMap type is irrelevant - it happens on Overworld, NES Dungeon, Interior and BS Overworld types. Link's hitbox size is irrelevant. Combo type is probably irrelevant - I tried it with standard ground (none), slow walk and shallow water, all to the same effect.

It requires a bit of precision to duplicate, but fortunately, it was easy to set up for recreating in the room that truly matters.

In the demo, you start in the room with the corner crash (marked by the red x). Walk left into the gray angled barrier, press down while still holding left, and that diagonal walk should crash you when you hit the bottom left corner of the screen. (It would probably be good to load the quest, exit ZC, then start the program again so that you don't have to keep reloading the quest after the crash.) The other rooms have adjusted angled barriers for testing, but you won't need to use those, so just ignore them (they each take some F4 frame-advance movement to position Link perfectly anyway).

Here are the 2.5x program versions I tested this on, including package name and date, and which ones crashed:

ZC 2.50.0 (official): All good.
ZC 2.50.2 (official): All good.
ZC 2.50.3 RC1 (zc-2-50-3-rc1-win.zip | Dec. 04, 2016): All good.
ZC 2.50.3 RC2 (2.50.3_Beta_2.zip | Dec. 04, 2016): All good.
ZC 2.53.0 B1 (2.53_Win_Beta1.zip | Jul. 10, 2017): Crash introduced.
ZC 2.53.0 B2 (2.53_Win_Beta_2.zip | Jul. 14, 2017): Crash.
ZC 2.53.0 B4 (2.53_Win_Beta_4.zip | Aug. 17, 2017): Crash.
[...]
ZC 2.53.0 B10 (2.53_Win_Beta_10--fixdrawint--fixgamecombo.zip | Oct. 02, 2017): Crash.
ZC 2.54 B52 (ZC_2.54_Beta_52_Public-VeryLate4.zip | Jan. 20, 2017): All good.
ZC 2.55 A (ZC_2.55_Alpha_28JUNE_2017.zip | Durrr): Crash.

I didn't think it was worth checking every beta between 4 and 10, but I could, if it's absolutely necessary.

For the record: the behavior before the crash was for the screen to scroll downward.

Edited by Lüt, 07 October 2017 - 03:08 PM.


#9 ZoriaRPG

ZoriaRPG

    ZC Contributor & Wiki Editor

  • Members
  • Gender:Unspecified
  • Location:Prydon Academy

Posted 16 October 2017 - 04:31 AM

OK, isolated it.

Angled Walk Corner Crash: [Demo Quest]

This is directly related to 8-way movement.

If you exit the bottom left corner of an open screen by walking diagonally downward-left, and you walk into the precise corner of the screen - like, the exact corner pixel - ZC can't seem to make up its mind whether to scroll sideways or downward, and simply decides to crash instead of scroll at all.

This only affects downward-left screen exits - downward-right, upward-left and upward-right all work fine.

DMap type is irrelevant - it happens on Overworld, NES Dungeon, Interior and BS Overworld types. Link's hitbox size is irrelevant. Combo type is probably irrelevant - I tried it with standard ground (none), slow walk and shallow water, all to the same effect.

It requires a bit of precision to duplicate, but fortunately, it was easy to set up for recreating in the room that truly matters.

In the demo, you start in the room with the corner crash (marked by the red x). Walk left into the gray angled barrier, press down while still holding left, and that diagonal walk should crash you when you hit the bottom left corner of the screen. (It would probably be good to load the quest, exit ZC, then start the program again so that you don't have to keep reloading the quest after the crash.) The other rooms have adjusted angled barriers for testing, but you won't need to use those, so just ignore them (they each take some F4 frame-advance movement to position Link perfectly anyway).

Here are the 2.5x program versions I tested this on, including package name and date, and which ones crashed:

ZC 2.50.0 (official): All good.
ZC 2.50.2 (official): All good.
ZC 2.50.3 RC1 (zc-2-50-3-rc1-win.zip | Dec. 04, 2016): All good.
ZC 2.50.3 RC2 (2.50.3_Beta_2.zip | Dec. 04, 2016): All good.
ZC 2.53.0 B1 (2.53_Win_Beta1.zip | Jul. 10, 2017): Crash introduced.
ZC 2.53.0 B2 (2.53_Win_Beta_2.zip | Jul. 14, 2017): Crash.
ZC 2.53.0 B4 (2.53_Win_Beta_4.zip | Aug. 17, 2017): Crash.
[...]
ZC 2.53.0 B10 (2.53_Win_Beta_10--fixdrawint--fixgamecombo.zip | Oct. 02, 2017): Crash.
ZC 2.54 B52 (ZC_2.54_Beta_52_Public-VeryLate4.zip | Jan. 20, 2017): All good.
ZC 2.55 A (ZC_2.55_Alpha_28JUNE_2017.zip | Durrr): Crash.

I didn't think it was worth checking every beta between 4 and 10, but I could, if it's absolutely necessary.

For the record: the behavior before the crash was for the screen to scroll downward.

 

This may be fixed now. I sent you a link on Discord to a test version.




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users