Bugs and complaints on current master

[PowerWyrm]

It might be my memory playing tricks with me, but didn't @ use to get some experience from disabling traps?

The xp you get when disarming traps is now something like depth / 5, so it's rounded down to zero until you reach 250ft.

[Ingwe Ingweron]

A new build is up on the nightlies page which (at least)...
Please report any problems.

At the outset, monster.txt ---> lore.txt is broken again.

Edit: Three crashes so far, one was while killing a troll, another was when a monster breathed down a hallway, killing some monsters in the process. It makes me suspect something to do with drops, but haven't been able to duplicate them, so no savefile yet.

[Ingwe Ingweron]

Three crashes so far, one was while killing a troll, another was when a monster breathed down a hallway, killing some monsters in the process. It makes me suspect something to do with drops, but haven't been able to duplicate them, so no savefile yet.

Revision - The crash is something to do with monster breath and spells. Several crashes from hound breaths, a druj spell, and Medusa breath.

[Nick]

Ah, good - off to a flying start

[Nick]

New build up which fixes the lore issue. For the others, I note that Medusa doesn't breathe, so I guess that was a bolt or ball spell. What type of hounds are you seeing this with?

[Ingwe Ingweron]

New build up which fixes the lore issue. For the others, I note that Medusa doesn't breathe, so I guess that was a bolt or ball spell. What type of hounds are you seeing this with?

If I recall correctly, the hounds breathed inertia. I'll work on getting a savefile for you.

[Ingwe Ingweron]

Here's a savefile with Curse weirdness and CRASH on Remove Curse. Belwether.zip

Curse Weirdness: @ has found Aiglos, which was identified as cursed when found, but on initially trying ?Remove Curse while in @'s pack, was informed "you have no curses to remove". ?Identify indicate "you have nothing to identify". @ must wield the weapon. Oddly enough, the curse can disappear when examining the weapon (not the "??", but the steelskin description, in this case. Go up or down stairs, or some other action that causes a refresh and the curse description would reappear. In addition, why are artifacts not normally cursed by their nature subject to these regular curses attaching to them? I think artifacts should be immune from such regular curses.

Remove Curse CRASH: Now, to the CRASH. When the curse is shown on Aiglos, @ reads ?Remove Curse, successfully removing the Steelskin curse, but, then on examination of the weapon a whole host of curse descriptions appears. Take one step to the right and CRASH. Or, <Esc> from the descriptions and CRASH. Basically, CRASH!

Hopefully the above description is clear enough to explain the curse/remove curse problems inherent in the attached savefile. Though, I'm very tired and think maybe I'm just rambling.

[PowerWyrm]

Here's a savefile with Curse weirdness and CRASH on Remove Curse.

I'm just looking at the commit about curses (a222ddd). When they were a linked list, there was no problem because a curse would exist or not depending on whether or not it was a "link". With arrays, a curse exists if its power is nonzero in the array. However, when a curse is appended, its power is calculated from the object level... and the result can be zero. An easy fix would be to ensure that the power of the curse is at least 1 when it is appended to the object.

[PowerWyrm]

Oh yes, arrays of curses are browsed using a "for" loop that sometimes starts with 0 and sometimes starts with 1. Since the Steelskin curse is the last in the txt file, it is the first one in the "curses" array. That could explain the crash.

[PowerWyrm]

Also inside those "for" loops, all curses with power = 0 should be excluded. For example (item_tester_uncursable):

for (i = 1; i < z_info->curse_max; i++) {
if (c[i].power < 100) return true;
}

should be:

for (i = 0; i < z_info->curse_max; i++) {
if (c[i].power == 0) continue;
if (c[i].power < 100) return true;
}

[PowerWyrm]

From the same commit, it seems like all randomization of curses has been lost: combat values (tohit, todam, toac) and modifiers. Since the "object" part of the curse has been removed, I'm not even sure curses do anything anymore (except the obvious ones).

[Nick]

Summary of my understanding of curse problems:

• Loops starting from 0 or 1 should (in general) both be fine, as curses[0] is a blank one
• Curse objects have not been removed as far as I'm aware; variability is certainly gone, and the whole of curse.txt needs to be overhauled
• Being able to generate curses with power 0 is indeed an oversight
• The crash is caused by freeing the object curse array but not setting the pointer to 0

I should get to fixing on the weekend.

[PowerWyrm]

Another change: the "Oops! It feels deathly cold!" message has been removed. To warn the player when wielding a cursed item, maybe keep the sound effect (using MSG_CURSED instead of MSG_WIELD if the item is cursed)?

[Nomad]

Another change: the "Oops! It feels deathly cold!" message has been removed. To warn the player when wielding a cursed item, maybe keep the sound effect (using MSG_CURSED instead of MSG_WIELD if the item is cursed)?

It seems like a bad idea to have information that's only available to the player as a sound cue. (I've always played with the sound turned off and I'm sure many other people do, plus it's inaccessible to deaf players.)

I feel like there should be a message on wielding a cursed object, just not the old "Oops! It feels deathly cold!" message, which is too strongly associated with the traditional sticky curse. Maybe "A dark aura surrounds you" or something like that?

[PowerWyrm]

Just a little code improvement (object_runes_known):

for (i = 0; i < z_info->curse_max; i++) {
		if (!curses_are_equal(obj, obj->known))
			return false;
	}

No need for the loop, curses_are_equal() already loops on the curses.