What a fun!

[T0Mi]

In your first release that floors&ceiling flip happened on my (rather outdated) Laptop too.
But only when Truecolor was disabled in the allegro.cfg. Setting it to 1 'solved' the problem.
I may as well run your next release on my (dual cored, more up-to-date graca) Catia workstation, if it turnes out that it is really the speed that causes problems.

I plan on not (repeat and note to self: -not-) bring home any work for the weekend, so I have time to play a little with CSBuilt, which usually leads to depression, agression, suicidal tendencies and playing Conflux, so any testing done on your clone is highly appreciated, because then crashes are not my fault. *harhar*

greez.
T0Mi

[Bit]

Being too engaged with other things, I lately digged out the stuff again and found a fix for the flickerbug - though I don't really understand it, because Paul's source still contains those lines.
But well - I changed my provider now, will hopefully have some webspace soon and place newer things there.

To reveal if a thing is buggy or not, I'd need an answer for the questions:
1. In an intense battle with four skeletons - may it happen that a lvl 1-fireball slips through the enemies?
2. Those stone crawlers in level 2 - do they release everytime a tune when showing the engaged-picture?

[beowuuf]

1. a bug in the original game, the code seemed to rarely let a monster move and fireball move ensure the check for if the monster is in the same square gets made too late, and so it slips through
2. i think sometimes not...that might be csbwin's sound checking though...

[zoom]

In csbwin:
1.I think this never happened in one of my games.
Maybe to take place situation needs:
many turnings of the monsters and frightening some of them while others are not in panic,
so there is an indiscrepancy on the monsters´ square and position/movement.
(one monster wants to get to the backrow other flee other charge the player?) This is just pure feel on my side,
beo has reliable info, as always!

2.interesting question.. I don´t know.
Probably it has to do with the attack conditions(is the monster able to breach the defense of the champion,etc)
At first I thought it would "uiuiurrrg" all the time it attacks, but I recall in csbwin that sometimes the sound would miss.
In a recent post of Paul Stevens, he stated that sound issues with csbwin are queued and that direct X would work "better".
http://www.dungeon-master.com/forum/vie ... ctX#p94403

[beowuuf]

1. No, it does happen with even a single monster if they move at the right - or wrong - time when the fireball / spell goes. You won't see it alot, but it does happen, and it doens't need a confusion of monsters or turning alot or anything like that

[zoom]

oh, ok. As said, I was just assuming.
a video footage of that would be nice, never witnessed that conciously!

@ Bit
does all this help you with your testing? You seem to dislike these occurences.
I would agree if 1 and 2 happened most of the time. But 1 in a thousand, no problem?

[Bit]

It's not that I dislike it, I just worry about the code's integrity.
So I'm very thankful for beowuuf's information.

I have to go through the process of source processing again now - that was already done, but you run into two different kind of bad situations, however you make it - won't describe that, that's too long. Needs some ideas now.
As soon as I have to do some things manually, there is a big risc that I do something wrong, but I can't let lower level code stay as it is.
Priority still is, to keep the functionality redundant (whichever original bug I may carry on) - so best is, to have an automated process that leads from version A to B, having another process that does it vice versa.
For that I need to know that the current version doesn't include too much faults already, and that's why I better ask. Hope I can provide that version as soon as possible, so that the one or other may help with some testing.
There is still no special purpose what to do with a final version after all, except that this code will be less complex than CSBwin is. It's more to keep my programming skills alive

[zoom]

Hope I can provide that version as soon as possible, so that the one or other may help with some testing.

yay! I would gladly help testing!!

There is still no special purpose what to do with a final version after all, except that this code will be less complex than CSBwin is.
It's more to keep my programming skills alive

You said you would do it because of programming exercise, no hurt there.
I just wondered:

1. Is it possible to run your program(do you have a Name for it??) normally and then -as an inbetween for the graphics- use another program, like entombed?
[OS] <------> [your program] <------> [3D program]

2. Any ambitions to introduce something like Csbuild(level editor) or ADGE(item-monster-graphic-editor)?

[Bit]

zoom, that's nothing more than the original game whereby I mixed both of the sources of DM and CSB into one, seperately compilable with one define set. It would even take ages to bring it that far that CSBwin is.
At the moment, the source is nothing more than a long C-code, where I put each 68000-assembler-command into a C-function, and wrote the necessary things to handle that. Just the hardware-things are wrapped to the Allegro-library, which is pretty easy to use.

Now that I know that the next version of Allegro will include OpenGL-things, I'm really tempted to bring it on that way - not knowing which problems will happen. But for that, I have to know where to find what in the program - atm there's still no readable function or variable name, I'm just proud that it works at all under those circumstances.
So the next step must be the end of the gotos, to have real C structures, for readability and continued processing. To avoid manual mistakes, I must find a way to do this automated - all I did in this is unfortunately still not satisfying. If I can't reach this, there's no point in thinking about future things.

[zoom]

bit, I was thinking out loud and was curious what you are up to maybe.
I see that the project is building its foundation still and I was pondering about couloring the walls..
first things first.
There must be a way to end the gotos!Viel Erfolg

[Bit]

Just that you say wall - there are sometimes dark buttons in the lower corner of a wall.
For those buttons it may happen, that the graphics switches top down once you pressed the button.
This is ugly, and I'm relatively sure that this didn't happen in the original (but who knows, that's ages ago) - and - in CSBwin the same thing happens. That's very interesting, because the DM-development path is independent from CSBwin. Hope we can see later what happens there.

I'm testing DM now pretty hard and are without problems near the rats room in map 08.

It's just:
- the two things being already described: slip-through fireballs, maybe missing sounds
- sleeping recovers the values pretty slow
- a first level stamina potion cures a bit much (that's a guess)
- and Wu Tse never was such hungry

That all doesn't look really buggy - and I'm in good hope that CSB will behave so brave too.

About the gotos zoom: those are more than monsters in DM, believe me...

p.s.: got one more question: what is needed to play DM PC-dungeons with my DM-Version.
Is there a lot to do to enable this?

[beowuuf]

Out of interest, what are your test characters for DM? If you have high level but low stat characters, then even a first level stamina potion will restore alot of stamina relatively speaking

Also, if you have high primary stats but low secondary stats, then the recovery rate will be in the toilet, and also food consumption I think will be raised a touch

Then again, you know better than me, you've seen the code! Perhaps Sophia or PAul S can confirm what governs these aspects...

[zoom]

maybe , bit you could test the same champions with the faithful csbwin to see a difference in the behaviours?

[Bit]

I picked them pretty randomly, just that it makes a bit of sense:
Current state:
Syra: Artisan F,Novice N, Apprentice P,Artisan W 187,146,51 48 kg
Nabi: -, Apprentice,Craftsman, Journeyman 117,90,42 45
Wu: Neophyte,Craftsman,Journeyman,Artisan 136,78,77 46
Halk: Adept, Neophyte, Novice, - 174,110,8 58

And yes, you're right, was Wu with the stamina potion, gives effectively just 13 points.
And maybe she eats that much cause she has pretty much to carry all the time
She's always low on stamina, because I'm running around pretty fast - that state is in map 08!
I also missed the jigglers in map 07, forcing the generator to spawn some more, I realized that the first ones obviously got shot by fireballs...
So - still pretty fine as it seems!
Think I just have to worry about the endgame. We'll see soon.

@zoom - yes, I should do so.
@beowuuf - seen the code: yes, working with it: yes, but - any idea what happens where: NO
(ok, in some cases I had to cheat and took a look in Paul's source - without that I probably hadn't have any chance, but I still keep it abstract to avoid that I work out some things and let the others be... forever - I know myself!)

[Zyx]

The hole switch that is at one place when off and at another when on, is easily fixable. Both wall items should belong to the same coordinate group, which you can fix in the graphics.dat with ADGE I think.

[Bit]

So I assume that bug is an original one.

[zoom]

considered the original dungeon masters had many ideas they did not implement, but left traces of these in the game and tried to cram up everything on a 512K diskette: you will face probably quite some bugs, bit!!

[Bit]

Some update:
I'm still hunting a few graphical bugs in CSB (something with the spellbox), and there is something that seems to delay every second step the party does, (but that's more a feeling).
I could play my DM version to the end and it behaved very brave - this is ready for a release, but because the whole code is mixed with the CSB-things, just give me a little time to bring it that far too.
After changing the provider, I ordered some webspace too, so will have a nice place for a lil website - will take a little time until it's setup.

Then - the next work in progress-version (will fix the former versions from that one):
The debugprocess is hard to do if you got ugly tagnumbers only. So I imported a lot of Paul's routine names - still not funny. After all, we're 20 years ahead, and there is C++. And I'd like to group things in another way than the disassembly gives the result to me (which ofc is maybe caused by some optimizing or whatever). Paul began that already in some parts - but stayed very close to the original source (which was a pretty good help for me to start that at all!). Let me say: CSBwin is a very smart solution in the middle of original and possible expansions! But - I'd like to have the source even more readable, more handable, whatever, I think those who worked with the source may have a feeling for my feelings.
So, what I want next, is a thing that has no more relation to old 68000-commands, laughs about endianess, has very clear objects with well-defined ports - but finally just changes the inner mechanics, invisible to the player. I even would like to add Paul's expansions, so that all the fine custom dungeons can be played with.
For this, I must read back if Sophia's DSB is already that what my intentions are - and if so, I better support her instead of going ahead through all that again with no better result.

Still - don't expect any exciting stuff in a near future - I still enjoy that all (even if I said that debugging is pretty hard - those are challenges ) - but I'm still running behind - ofc.

[sucinum]

I have done a bit of testing around stamina/mana reg and food consumption some time ago. It might help you or not: http://www.dungeon-master.com/forum/vie ... ina#p40746
I never looked into the code, so i can't say exact values, but since you can, having the idea might be enough. Sorry to bother if you knew that already and the problem is a different one

[ChristopheF]

Bit, in case you are interested, I can create an account for you on the Encyclopaedia so you can upload your program and post about it there, in which case you would not need to create a dedicated web site. However if you prefer creating your own web site (which I can perfectly understand) I'll be happy to add a link to your own site.

[Bit]

@sucinum:
At the moment in my CSB the characters regenerate (by sleeping) much faster than in my DM.
In fact I believe they do in CSB somewhat faster than in the original, in DM much slower.
Playing to the end being impatient (ofc), with not too much sleep, it was pretty hard to survive some situations and I died a couple of times (shouldn't it be that way?). Finally I ended having not one master title for any character (just Wu was close to master wizard, but she always does, she is and was never a ninja ). She and Syra could do a level 6 fireball, but with a high risc of sizzling. Both hadn't that much mana to cast a higher level lightning. I will find out all formulas, and once I got them, check against the original disassembly again. One thing I can say that I'm pretty sure about: We had a smaller discussion about the fireshield - seems that there is indeed a bug in the original.

I hope I can catch up fast now in understanding what happens where and why, so that I stand a chance to find the reasons for that behaviour. Currently I don't have a real vertical blank interrupt, and I'm not sure about the accuracy of the timing (and how it's finally happening at all!) - I just inserted a 70 Hz-timing, poll the vbl-routine often and execute it, if the timer has done a step.
Because each operation is still an original 68000-command that is performed in an (unoptimized) C-function, and each memory access is controlled and the original 68000-endianess is kept, also the graphic becomes always decoded from original ST-format to a 640*400-screen, the overall performance is even more worse than on an original ST if you play it on a P3 with less than 1 Ghz speed...
And - Paul already pointed on it - it still steals all CPU time. I'll probably have a solution for that once I fixed the CSB-problems.

Have also to investigate those event-things, some events seem to take a little time. i.e. there is that foot-plate in level 6(? under the firetomb), with that 'endless' passage - that opens the wall to the cave where the Delta can be found. There is some delay until that wall vanishes - and I doubt that this is original. And - as I said - in CSB there's a small delay after each second step (mirrorized floor? - why in CSB only then?!?). Also, for my taste, fireballs fly too slow there, compared to DM. There *is* something.

@Christophe:
That's a generous offer! Both executables aren't more than 750 kb, one source (including both, DM and CSB) has about 1 MB - unzipped. Then the allegro.dll is needed.
I wanted to release different sources of some milestones in the development process - this could be of interest, like a tutorial. Not that I'd support 'reverse' engineers - in fact the outcoming is: look, 140 kb original code explode into 50000 lines, a thousand routines and some variables - so - let it be!
Your encyclopaedia is the real home of any DM/CSB-stuff, so I'd be proud to present it there.
I still can use my webspace then for more instable development branches.
Just those few things to fix (and one CSB-game played until the end), and I will contact you by PM.

[zoom]

One thing I can say that I'm pretty sure about: We had a smaller discussion about the fireshield - seems that there is indeed a bug in the original.

I thought these bugs were:
normal shields and anti fire stat,
both having effectively no effect
apart from style(hey look I am wearing a big nice shield and
cool, 80 in a-fire stat.
Nevertheless they were cool to help fleshing out more the champions..
Lizars got higher A-fire than elves etc)

what is the bug with fireshields? are they much too powerful?

Have also to investigate those event-things, some events seem to take a little time. i.e. there is that foot-plate in level 6(? under the firetomb), with that 'endless' passage - that opens the wall to the cave where the Delta can be found. There is some delay until that wall vanishes - and I doubt that this is original.

what you describe as being probably faulty, is however the way how I know it.. there is/was definitely some delay, don't remember exactly though. So it seems your version works correct here!

And - as I said - in CSB there's a small delay after each second step (mirrorized floor? - why in CSB only then?!?). Also, for my taste, fireballs fly too slow there, compared to DM. There *is* something.

Does this help to shed some light?
In the Atari version there are some differences to the amiga version.
e.g Atari>yes, you can run into your projectiles and spells(fireball) because you move faster than the object/spell flies! Amiga :no((pretty sure, then again not really )
use hand to let leader drink from fountain: atari no
operate door switch with item in hand(atari: need empty hand to operate switch)
..probably many other differences, from scroll placement and item colour in viewport(blue for atari)

maybe you dissassembled a broken version? just kidding , good luck in sorting it out, bit!!!

At the moment in my CSB the characters regenerate (by sleeping) much faster than in my DM.
In fact I believe they do in CSB somewhat faster than in the original, in DM much slower.
Due to Champions having different stats? 1% regeneration "per second" can be different for increased health values and I guess with vitality etc it can get quite complicated!
No way to compare different champions.. at least this is my guess.. your version could be wrong
there are differences in actions for dm and csb.. action consume different stamina some gain different subskills(warcry)
read it at encyclopaedia....items and actions

[sucinum]

@sucinum:
At the moment in my CSB the characters regenerate (by sleeping) much faster than in my DM.
In fact I believe they do in CSB somewhat faster than in the original, in DM much slower..

Did you try Atari DM or PC DM? I think latter is as fast as CSB, but the early versions of DM used to "tick" slower (in terms of regeneration, both awake and sleeping - might affect also AI upgrades and stuff, but don't nail me on that). PC DM was V3.4 (my multilanguage one at least). There should be a "break" (greater update) somewhere, maybe the amiga (2.0) or PC (3.0?) version (where you could also drink from fountains or detect secret doors by clicking). Might or might not be related to that.
Might be wrong, though, was a long time ago and i didn't compare both next to each other. But i'm quite sure there was a difference for me after swapping to a PC (and PC DM). Back then, i thought it was because the PC had a much faster engine, but could have been coded differently as well.
You can get PC DM 3.4 (and several amiga versions) on the encyclopedia if you want to investigate further. I don't think there is a full version history of all changes somewhere, at least no official one.

Sorry for that much speculation, i hope i am not misleading you. Translate all my babble into "you might be up to something, better find out by yourself".

[Bit]

My DM version is an old Atari one (1.1).
The CSB code is from version 2.0.
I tried UAE-memory-snapshots from the Amiga version, but after a short look into it, it this doesn't make sense to me. The Atari code has about 500 returns in it, so one should expect them in that code too - but they aren't! Either they unscramble each part of code which lies ahead, or they use another technic to return from a subroutine - or that code is nothing but Basic tokens
So CSB-Amiga stays in the dark for now - and with it the automapping

I am still sorting out the stuff now, we will see what the difference makes.
One's for sure - it's all pretty sensible. I cannot simple clean out already known variables from the great data-array. This can lead to immediate crashes. Means: Even simple integer-variables may belong to higher structures that then are cleaned packed by simple memcpys.

I just imagine Paul smiling on all of this, getting some more popcorn, already knowing the next trap I will run into and wondering how I will solve it - if at all. But - Bits can byte!

[Antman]

It's not only monsters that can be missed by fireballs if they move at the right time, the party can also 'walk through fireballs' this way if you move foward a square at the right time or spin at the right time or move a characters position at the right time. Tricky though, it's probably only a 6th of a second window or something.

I also remember those weird switches flicking from the bottom of the wall to the top of the wall in DM, I can't remember exactly where, I think it was only the switches that can only be used once, not repeatedly.

It is true that some things in the amiga version are slower, mainly monsters making it easier, but you can still walk into you own projectiles and fireballs, with and without the boots of speed.

My 5 cents anyway. Good luck with the project!

[Paul Stevens]

I just imagine Paul smiling on all of this

How did you know?

The endianess will not go away if you want to read standard
save files and you want your code to have some chance of
working on big-endian machines.

[Bit]

Yes, that's right - but like you, I will try to hit it with a club as soon as it's coming in.
And all that leaves again, will get a kick
But for this, I must learn a lot more about the formats - will not happen soon. Until then I have to carry a lot...

But I'm in a good mood this morning, because I found a simple solution for the last obvious bug.
That was also due to the polling of the vbl - was about highlighting (and especially de-highlighting) rectangles. The delay problems vanish with the release code on a not too slow machine.

I then was shocked testing all that, when I shot a fireball in the very first chamber of CSB against a teleporter (from a distance). Just do it, then step aside and look what's happening - never saw that before - but - it's original - CSBwin does the same

Will provide a first pack of files tonight - if I manage ftp-upload to my new site. Looks promising.

[Bit]

Available here.
It's a 2.0-MB-file with two source-folders (lvl 1 and 3 of development) and play-folders for DM and CSB.
Ofc you have to add graphics.dat and dungeon.dat (and copy the allegro.dll to the folders, that's in an extra folder). See the notes in info.txt.
Have fun!

Edit: had to make another update - recent zip-file is now from 11.april 09, 7:39. 2.069.324 bytes.
Vbl-timer simulation set back to 70 Hz. Much smoother now.
Some odd behaviour when loading the sample CSB-datafile: two colors are modified the wrong way.
I can't say if it is my csb-code, or if there got something screwed during the conversion by the csbwin-utility.
We'll see later. For now I simply ignore the modifypalette-routine and return immediately - don't see side effects so far.
Weblink changed - nicer name now

[Bit]

I have uploaded another pair of current exes, just update them in the other folder
work in progress

Except the stack, the global registers have done their work, all parameters of procedures and functions are well defined now - not in their true structures (which still have to be revealed), but the work can continue now more local. That way I can enhance the automatic goto-elimination step by step. That one already worked, but the result was even more unreadable - there's still a lot to do!
In that WIP the lzw-decompression does the work independently, and so the reading of the save game happens in a second.

So far I spotted no problems, except the already known thing with the restart after death in CSB (DM works!), and that the endsequence of DM is too fast for my taste.

Next is cleaning up the procedure's local stack, and for that I probably have to setup the first original data-structures. That already worked well in another development-branch for some graphical things, a nice class is growing there handling dm's graphic memory and lists. In fact I dislike their handling with variable sized structures, that's not good for C++. Will surely rewrite that in another style. First trials using new/delete instead of their malloc/free-system for temporaray things worked pretty good.
I fully agree to Paul that graphic's caching isn't needed anymore - and if, this should be done in another style. I also hate that most graphical things are limited to 16-bit-sizes, should be possible to enhance this to use larger graphics of other formats too.

Found another two bad commands that don't harm the program as long as it works in 68000-emu-environment, but crash immediately if being translated - surely nothing to do with the protection stuff, those are simply bugs, no idea what the compiler did there. CSBwin not harmed by them, Paul must have met them too.
Having learned now about how that compiler worked, I must say: there were reasons, why the atari dropped a lot of bombs...

Any testing comment welcome!

EDIT: newer version uploaded, because I've been told about problems with the mouse on very fast machines. poll_mouse-routine used now as well as the acquire/release-screen commands, which are meant to fix such problems - never had to use them for my own, just pressing thumbs now...

[Bit]

Just a quick note about the current standing:

Last week I was progressing fine, structuring the data-field more and more and reflecting it into the source,
also adjusting load&save to 'unendian' the main program. Then I met that field that Paul named 'DBCOMMON' - and that spreads like hogweed throughout everything. I'd say I'm in the middle of adjusting that now.
Until yet, I could keep code stability the whole time by checking often diff-files of versions, but that one is now a free fall - it's simply too much. Hopefully I got the right ideas by forcing compiler errors step by step.
More than 100 zones left to workout...
If the code survives that without too much problems - I'm in good hope to reach the end.