Angband 4.2.5

[PowerWyrm]

The logic for reducing chance of successfully applying a skill (disarming, opening...) is not consistent everywhere. Sometimes a factor of 1/10 is applied for TMD_BLIND/no light/TMD_CONFUSED/TMD_IMAGE and sometimes a 1/10 factor is applied for TMD_BLIND/no light and then another 1/10 reduction is applied for TMD_CONFUSED/TMD_IMAGE.

10% factor: do_cmd_lock_door(), do_cmd_disarm_aux()
10% + 10% factors: do_cmd_open_chest(), do_cmd_disarm_chest(), calc_unlocking_chance()

The best would be to rename calc_unlocking_chance() and use it everywhere a skill power is calculated.

[mrfy]

Using 4.2.5-185-g02b5e5f72 on macosx, I'm noticing that the first action taken doesn't make a sound but the second does.

For example,
Picking up multiple arrows/shots/bolts, sometimes the first item picked up is silent but 2nd and following ones make the quiver pick up sound.

Or switch to tunneling, and the first wall removed doesn't make a sound but following ones do.

Submitted to github as #5962

[Nick]

Using 4.2.5-185-g02b5e5f72 on macosx, I'm noticing that the first action taken doesn't make a sound but the second does.

For example,
Picking up multiple arrows/shots/bolts, sometimes the first item picked up is silent but 2nd and following ones make the quiver pick up sound.

Or switch to tunneling, and the first wall removed doesn't make a sound but following ones do.

Submitted to github as #5962

Thanks for reporting.

[Ingwe Ingweron]

Has there been a change to "Afraid" status? Playing the nightly, I've noticed that @ can still melee while "Afraid" even though @ does not have pFear.

[backwardsEric]

Has there been a change to "Afraid" status? Playing the nightly, I've noticed that @ can still melee while "Afraid" even though @ does not have pFear.

Not intentionally, but a post 4.2.5 change caused two timed effects (AFRAID and TERROR) to not apply the side effects of fear to the player. A proposed fix for that is at https://github.com/angband/angband/pull/5965 . Thanks for reporting the problem.

[Nick]

Not intentionally, but a post 4.2.5 change caused two timed effects (AFRAID and TERROR) to not apply the side effects of fear to the player. A proposed fix for that is at https://github.com/angband/angband/pull/5965 . Thanks for reporting the problem.

That fix is in place now.

[gglibertine]

Very minor bug -- the potion is called "Resist Heat", but when you take it, your status bar says "Resist Fire". This is inconsistent with the behaviour of "Resist Cold", which matches in both places. (I observed this in the Android app but remember it being the same in the Windows build.)

[PowerWyrm]

Someone noticed me of a bug in my variant based directly off 4.2.5: there was a pile integrity failure about some copper coins and the log showed that just before that a player was digging some quartz vein and found some copper, followed by some message about discovering a hidden mimic. Do you thing it's possible that instead of copper then game generated hidden creeping coins instead? I don't see how this is possible in the 4.2.5 code and I don't remember changing anything about the mimic code in my variant...

[backwardsEric]

Like you, I do not see how place_gold() would generate a monster instead. Does the log make it clear that the mimic was revealed when the player tried to move onto the copper from the vein? Or could it have happened some time later at a different location?

There's one outstanding pull request related to mimics, https://github.com/angband/angband/pull/5979 . However, that would only be relevant for gauntlet, hard-centre, or lair levels. Even then, the pull request fixes up the link from the mimicked object to the monster and that link only comes into play when revealing a mimic in project_o().

[PowerWyrm]

The problem here is that the pile integrity check compares two identical objects (some copper), but the first one (square(cave, obj->grid)->obj) is address 02E06B70 and the other one (mon->mimicked_obj) is address 02FF5F90, so there must have been an object copy somewhere instead of just equaling the pointers.

[PowerWyrm]

I think I found the reason... and it's an interesting one. When a mimic is generated, we call place_monster() to place the monster then mon_create_mimicked_object() to create the corresponding mimicked object. Inside mon_create_mimicked_object(), the money object is generated via make_gold() and directly assigned to the mimic (mon->mimicked_obj = obj). However, the object is placed inside the cave via floor_carry(). Guess what happens if the mimic (copper) was placed on a square... that already contained some copper! Yeah, the mimicked object is absorbed by the object in the square and the amount is just increased... and mon->mimicked_obj now points do a deleted object.

[backwardsEric]

https://github.com/angband/angband/pull/6033 will prevent a mimicked object from stacking with another object. Monster placement uses empty grids (this is explicit in pick_and_place_distant_monster() and vault_monsters(); for vaults it is implicit by the mapping of vault symbols to stuff placed in the dungeon) as far as I can tell, so I don't expect that pull request to help with the pile integrity failure. It will keep dropped objects from merging with a mimicked object already on the floor.

[Nick]

I haven't looked, but what about a mimic moving onto a money pile?

[backwardsEric]

While it's mimicking, it does not move on its own (mon-move.c:1946). If moving on its own, the mimicked object has been removed from the dungeon. If another monster forces it to swap positions, that could do it, though it would have to happen out of sight of the player.

[Nick]

Yeah, I remember long ago (before the split between reality and what the player sees) there used to be a thing where you would detect money in a vault, then see it moving around as the mimic was pushed by other monsters.