CASA ::

After reading Canalboy's comments on Thief's Adventure I started playing the game out of curiosity, and one thing led to another, I ended up thoroughly playtesting it. I can guarantee that nobody ever has solved this game, because it turns out that it is impossible to solve. And I know exactly why: the same flag is used twice for two different conditions on separate puzzles that should be totally independent. If you take the ladder rung at the beginning of the game, then you can't ever find the ivory carving in the Palace attic, and vice versa. Since you need both the rung and the carving to finish the game, you won't be able to. It's not a matter of timing in getting to the carving, as it was initially surmised: it only depends on getting the rung first or not.

This is very likely an unintended bug in the code, but it's 100% reproducible. So the game is impossible to finish in its original form.

However: It might be possible to patch the binary and fix the bug, now that I'm aware of what it is. I'll try to dust off my ancient debugging kung fu. We'll see what can be done. Even if successful, it'll be no assurance of finishing the game as there may be other fatal bugs (I have experienced the snake bug mentioned in the comments too, but that one seems to be harmless) and, of course, there's the minor issue of actually solving the puzzles. I'll report on progress.

The issue with a lot of these homegrown games is that we may well have a bugged version archived, when there was later a fixed version made. A shareware publisher could certainly use a bugged version as a tactic for people to get in contact with them and potentially sell them a licence.

John Elliot did mention that he'd updated unQuill to work on the PC versions of Quill/AdventureWriter games, but I'm not sure if I've got a copy where that is working.

Ah, it's the advlist (or advlistw32) files and/or adv2txt (or adv2txtw32) in the latest unQuill folder that will extract the game code.

EDIT: Yep, just run the game file through that and you can view all the Quill code, which should make the process of debugging and solving the puzzles a lot easier.

There's also auraes' similar tool, which they've posted about here: https://solutionarchive.com/phpBB3/viewtopic.php?t=1942 [I just haven't got around to playing about with that one yet]

OK, I have been very lucky. I have a fixed binary of the game (no unquill used, but it would certainly have saved me a lot of hassle). Fortunately the opcodes are well documented so I only had to look for sequences of the form 0B <flag> 1B <flag> way apart with the same flag (these correspond to ZERO and SET respectively) and try all of them, but there were only a dozen or so possibilities so trial and error did it. My main fear was that there might have been no ordinary flags left over, but there were. I then reduced the initial patch to just changing two bytes.

The fixed binary does not have the bug. In principle the game can be finished now. I'll give it a go and then upload the executable so the community can enjoy the game as intended.

The game is now completed. I'll post the fixed executables and map. A detailed solution will follow later.

Added to the site.

I've updated the game's source code with Exemptus's fixes:
https://gitlab.com/auraes/unaw-dos/-/tr ... type=heads