State of Angband master branch

[d_m]

As of now (commit 87a261ebd99090fb3bd0c2a6342faa3b5fcda107 in the master branch of the angband/angband repo on github [1]) the following changes have been made:

1. Extensive code refactoring by elly and takkaria (in particular edit file parsing and the menu API)
2. new spell browsing from takkaria
3. vault changes from fizzix
4. fractional blows from magnate
5. new WELL RNG from d_m (non on github)
6. misc bugfixes

So, at this point there is a lot of good new stuff in there for people to play with, and more coming down the pipe! There are a few known rough edges but things should be playable at this point (and the rough edges are being smoothed as I write).

So, go ahead and grab a Windows nightly [2] (Mac nightlies were broken but should resume shortly). Or you can build from source. The procedure for cloning from git should be as simple as "git clone git://github.com/angband/angband.git". The build instructions are platform-dependent and should be available elsewhere on oook.

Also, if you're someone who likes submitting patches or fiddling with the code, I encourage you to get an account on github and fork angband. (EDIT: and if your github username is different, make sure we know who you are.)

Finally, if you have development questions, stop by #angband-dev on the IRC network freenode.

[1] https://github.com/angband/angband
[2] http://rephial.org/nightlies/

[fizzix]

feedback on vaults would be greatly appreciated.

[ewert]

My current end game warrior has had real fun with vaults. Maybe 3 greater ones, am around dlvl 85 I think but ready for Morgoth really. Much less diagonal vaults I think though, dunno why, probably just RNG. Oh and the zoos/graveyards etc. definitely had items in them, sure squelch took care of many, but it felt nice. Cleared especially zoos whenever found them, except one stupid one with like 4 greater Q's ...

Anyways, I think I will try to get the master branch changes in ivanilla, if master is working even on birth screen.

[Timo Pietilä]

Question: does that break savefile compability? If not I import my current char to it and continue with it (warrior, so fractional blows affect a lot).

[fizzix]

Originally Posted by ewert

Much less diagonal vaults I think though, dunno why, probably just RNG.

Diagonal vaults were moved to medium vaults mainly because I needed more. They were also beefed up a bit. Diagonal vaults were really the anomaly in lesser vaults. They were as good, if not better than about 3-4 of the greater vaults.

I still want to futz around with lesser and medium vault generation chances, I'm not entirely satisfied. Pits are too common because they appear any time a vault was attempted but didn't fit, this seems a bit buggy.

Originally Posted by ewert

Oh and the zoos/graveyards etc. definitely had items in them, sure squelch took care of many, but it felt nice. Cleared especially zoos whenever found them, except one stupid one with like 4 greater Q's ...

I actually didn't think there were enough items there...Although my endgame weapon was from the one zoo I cleared.

@Timo, I don't think it will break savefiles, since I've tested most of these changes already. I started with a post-death file though, not a mid-game one.

[Timo Pietilä]

Originally Posted by fizzix

I still want to futz around with lesser and medium vault generation chances, I'm not entirely satisfied. Pits are too common because they appear any time a vault was attempted but didn't fit, this seems a bit buggy.
.

How about trying smaller vault first? If vaults then become too common you can reduce the chance of trying vaults in first place.

[takkaria]

Originally Posted by Timo Pietilä

Question: does that break savefile compability? If not I import my current char to it and continue with it (warrior, so fractional blows affect a lot).

Nope, savefiles should work just fine.

[takkaria]

Originally Posted by ewert

My current end game warrior has had real fun with vaults. Maybe 3 greater ones, am around dlvl 85 I think but ready for Morgoth really. Much less diagonal vaults I think though, dunno why, probably just RNG. Oh and the zoos/graveyards etc. definitely had items in them, sure squelch took care of many, but it felt nice. Cleared especially zoos whenever found them, except one stupid one with like 4 greater Q's ...

Anyways, I think I will try to get the master branch changes in ivanilla, if master is working even on birth screen.

You might want to wait for 3.2-- I'm planning on pulling in your no_selling option too before release. (3.2 is coming RSN)

[fizzix]

Originally Posted by Timo Pietilä

How about trying smaller vault first? If vaults then become too common you can reduce the chance of trying vaults in first place.

This is a simple and perhaps ideal solution that would require a major rewrite of how the dungeon generation code works.

Briefly: the dungeon is broken up into 18x6 blocks of 11x11 squares. A block is chosen at random, and a room is attempted. If the room is a special room a section of blocks is reserved, and if other stuff exists in this section the room is rejected.

Both lesser vaults and medium vaults require 33x22 or (3x2 in block space). Pits and nests and other special rooms require 33x11 and are thus nearly twice as likely to fit.

A solution, that does not require a rewrite of the blocking mechanism, would be to change the rejection criteria by reading out the room dimensions first, I've been working on this on and off, but got distracted. This still doesn't change that a pit will be attempted if the vault fails, although it will reduce the chances of vault failure.

GVs are pretty much a mess because they're so big, they need to be the first room selected to exist or overwrite the rooms below. I chose to make them the first rooms chosen since that was much easier to implement and it freed up room for medium vaults (an idea that I got from you, perhaps accidentally).

[d_m]

I agree with fizzix--the current GV patch is not a perfect solution but does seem to give a real improvement, so I figured we should include it.

I am hoping to do a more drastic rewrite of generate.c at some point (but I've been working on it slowly for a year, so no promises...) I expect that if I don't get to it someone else on github will.