Main | My View | View Issues | Change Log | Wiki | Tavern | News
  

Viewing Issue Simple Details [ Jump to Notes ] [ Wiki ] [ View Advanced ] [ Issue History ] [ Print ]
ID Category Severity Reproducibility Date Submitted Last Update
0004101 [DCSS] Bug Report major N/A 2011-06-05 08:41 2012-01-10 04:49
Reporter evilmike View Status public  
Assigned To dpeg
Priority high Resolution done  
Status closed   Product Branch 0.9 ancient branch
Summary 0004101: D:27 swamped with draconians
Description I've made number of vaults that contain the entrance to zot, since it was mentioned in ##crawl-dev that it would be good to have some. These have been designed as branch entries (for zot), but also as branch endings (for the main dungeon). Thus, they are larger than typical branch entry vaults, and are a bit more challenging (some harder than others).

The monsters are draconians and dragons. For the draconians, they are mostly the unclassed (base draconian) type, with most vaults only having a few classed ones. Thus, the difficulty should be appropriate for the bottom of the dungeon, but still easier than zot. None of these vaults contain any real loot (one of them contains a small amount of gold, but that's all).

Note that in order for these vaults to work as intended, crawl needs to generate just one entrance to zot, rather than 4+ (as it does currently). I have already submitted a patch for this, which changes 3 lines in dungeon.cc (it's a pretty simple change).


There is a weird issue with placing these vaults, currently. In order to guarantee that one gets placed, I am using "PLACE: D:27" (this line guarantees one gets picked right at the start of level generation), and have avoided tagging them with zot_entry. The reason is that if I just use the zot_entry tag, the game will only pick a vault _after_ generating most of the level, and thus won't be able to find a spot to put the vault (it will try several times and fail... try this in debug mode to see what I mean). On the other hand, if I use "PLACE: D:27" and the zot_entry tag at the same time, it will pick one vault for the PLACE line, and another vault for the zot_entry tag (and likely fail at this part). Thus, I have opted for just using PLACE, as a workaround.

Unfortunately, if there are no vaults tagged with zot_entry for it to choose, the game will randomly place a portal on the level. This happens in addition to picking one of the D:27 vaults (in other words, two portals will be generated: one inside a vault, and one outside). I've worked around this by using an empty dummy vault tagged with zot_entry. I've explained this in a lengthy comment in the vault file itself.

I'd like to have come up with a better solution than this hack, but I don't know the crawl source well enough to do that (and I'm not much of a C++ programmer anyway). Luckily, this method works for now, even if it's rather inelegant. I suspect that the best solution in the long term would be to make vaults tagged with zot_entry be placed at the beginning of level generation.
Additional Information
Tags No tags attached.
Attached Files txt file icon zot-entry.txt [^] (10,941 bytes) 2011-06-05 08:41 [Show Content]
txt file icon zot-entry-updated.txt [^] (9,535 bytes) 2011-06-07 22:06 [Show Content]
txt file icon zot-entry-updated2.txt [^] (7,685 bytes) 2011-06-08 10:59 [Show Content]
? file icon morgue-Mathis-20110710-111651.map [^] (5,025 bytes) 2011-07-10 18:56
txt file icon morgue-Mathis-20110710-111651.txt [^] (46,337 bytes) 2011-07-10 18:56 [Show Content]

- Relationships

-  Notes
(0013395)
evilmike (developer)
2011-06-05 08:43

 Here is the patch I mentioned: https://crawl.develz.org/mantis/view.php?id=4100 [^]

All it does is stop stairs and hatches from being put on D:27 (which would get converted to portals).
(0013415)
KiloByte (manager)
2011-06-07 21:26

 I'm afraid it's a bad idea to place branch entrances as PLACE:D:27. Even if you removed all other primary vaults for that depth, the dungeon builder doesn't guarantee it will be placed -- to the contrary, it gives up placing a large vault pretty fast. Branch entrances need to be guaranteed, especially Zot.
(0013416)
evilmike (developer)
2011-06-07 22:05

 Oh, I didn't realize even PLACE: D:27 made them non-guaranteed. I've gone ahead and reuploaded this, so everything is tagged as a normal branch entry (and no PLACE) lines. That will at least allow them to be usable if the issue with failing to place them is resolved. I don't know how easy that would be, though.
(0013419)
Zaba (developer)
2011-06-08 09:30

 Yup, you shouldn't use PLACE like this, and just tagging zot_entry them is enough (provided the patch that removes the special case is applied).

The updated version looks fine to me.
(0013420)
evilmike (developer)
2011-06-08 11:05

 Alright, I've reuploaded them yet again, after some advice from ##crawl-dev. I removed the ORIENT line from all of them and redesigned them a bit so they take up less space. That pretty much solves all of the previous problems regarding placement. It still struggles with the castle one sometimes, but that's because it's a lot bigger than the rest, and it can't be made any smaller. Anyway, thanks for the help, everything actually works as intended now.
(0013690)
dpeg (administrator)
2011-07-05 21:59

 Wonderful! All in. Perhaps the Castle could have some silly wand at times...
(0013702)
jeffqyzt (reporter)
2011-07-06 14:41

 RE: Perhaps the Castle could have some silly wand at times...

Should have a wand of random effects (a.k.a. wishful thinking) ;-)
(0013738)
dolphin (reporter)
2011-07-10 18:20

 I think I may have just played one of these (like a gnoll castle, but with draconians). Is it supposed to have 200-300 draconians? Or is there an error with the monster generation?

(0.9-a1-953-ga0a87fa)
(0013739)
dpeg (administrator)
2011-07-10 18:33

 evilmike_zot_entry_castle is not blame. Can you give more pointers, please?
(0013740)
dolphin (reporter)
2011-07-10 18:57

 This is what I know that I know.
- I killed about 237 draconians (235 regular, 1 annihilator, 2 knights) before making a stupid mistake. I'm blaming repitition
-I was on D:27
-There was a castle that resembled the first gnoll castle, with a three tile dirt entrance gate, and a few windows on the outskirts.
-The Zot entrance was somewhere inside that area, since everywhere else was explored.

Here's the morgue and map file.
(0013743)
dpeg (administrator)
2011-07-10 22:33

 This bug report came as feedback, sorry for the weird format of this item.
I don't have time to investigate, apologies for that. I did have another look into the new Zot maps and they don't do anything obviously wrong.
(0013744)
evilmike (developer)
2011-07-10 23:54

 From looking at that map file, the vault there isn't one of the ones I made (the castle I made is surrounded by a moat and has larger "towers" than the one in that .map file).

I'm not fully sure what vault it is (it looks like lemuel_castle though).
(0013749)
mumra (developer)
2011-07-11 12:57

 I had this as well; literally hundreds of Dracs, they just didn't stop coming. Managed to wipe them all out in the end but it was ridiculously tough.

I had a look at large_lemuel_castle.des and I see this (for depth>=25):

            e.subst("w:w1")

So if I'm reading it correctly there's a 50% chance that the *entire moat will get swapped for Draconians* (which explains exactly what I saw, pretty sure dolphin got the same). And this could happen anywhere from D:25 onwards.
(0013760)
KiloByte (manager)
2011-07-12 01:34

 Thanks for investigation, mumra. I guess it was meant to be "l" (lava).

I've changed it in trunk: evilmike, dpeg, if it's not what you want, please say so. And if it is, you may close.

- Issue History
Date Modified Username Field Change
2011-06-05 08:41 evilmike New Issue
2011-06-05 08:41 evilmike File Added: zot-entry.txt
2011-06-05 08:43 evilmike Note Added: 0013395
2011-06-07 21:26 KiloByte Note Added: 0013415
2011-06-07 22:05 evilmike Note Added: 0013416
2011-06-07 22:06 evilmike File Added: zot-entry-updated.txt
2011-06-08 09:30 Zaba Note Added: 0013419
2011-06-08 10:59 evilmike File Added: zot-entry-updated2.txt
2011-06-08 11:05 evilmike Note Added: 0013420
2011-07-05 20:50 dpeg Assigned To => dpeg
2011-07-05 21:59 dpeg Note Added: 0013690
2011-07-05 21:59 dpeg Status new => resolved
2011-07-05 21:59 dpeg Fixed in Branch => 0.9 development branch
2011-07-05 21:59 dpeg Resolution open => done
2011-07-06 14:41 jeffqyzt Note Added: 0013702
2011-07-10 18:20 dolphin Note Added: 0013738
2011-07-10 18:33 dpeg Note Added: 0013739
2011-07-10 18:56 dolphin File Added: morgue-Mathis-20110710-111651.map
2011-07-10 18:56 dolphin File Added: morgue-Mathis-20110710-111651.txt
2011-07-10 18:57 dolphin Note Added: 0013740
2011-07-10 22:33 dpeg Note Added: 0013743
2011-07-10 22:33 dpeg Priority normal => high
2011-07-10 22:33 dpeg Severity feature => major
2011-07-10 22:33 dpeg Status resolved => feedback
2011-07-10 22:33 dpeg Resolution done => reopened
2011-07-10 22:33 dpeg Category Upload: Maps and Vaults => Bug Report
2011-07-10 22:33 dpeg Summary Zot entry vaults (7) => D:27 swamped with draconians
2011-07-10 23:54 evilmike Note Added: 0013744
2011-07-11 12:57 mumra Note Added: 0013749
2011-07-12 01:34 KiloByte Note Added: 0013760
2011-08-18 11:08 KiloByte Status feedback => resolved
2011-08-18 11:08 KiloByte Resolution reopened => done
2012-01-10 04:49 evilmike Status resolved => closed

Mantis 1.1.8[^]
Copyright © 2000 - 2009 Mantis Group
Powered by Mantis Bugtracker