Just for the record; IMO this was a terrible idea. An off-line game should have off-line help.

I strongly agree with this. Not everyone has unlimited connectivity or data.

For 4.2.2, the .zip file for Windows and the .dmg file for macOS include the HTML versions of the help. In the .zip file, the help is in the docs directory; in the .dmg file, it is in the Docs directory.

Wand of Clone Monster bug
 

The wand only works when the mob is awake. If that is a new feature, I'm behind the times. But, it used to work regardless of mob state in 3.x.

In Project-mob.c -

/* Attempt to clone. */
if (multiply_monster(cave, context->mon))
context->hurt_msg = MON_MSG_SPAWN;

--- does not seem to work because, of bool sleep in mon-make.c's code:

* If `sleep` is true, the monster is placed with its default sleep value,
* which is given in monster.txt.
*
... */
bool place_new_monster(struct chunk *c, struct loc grid,
struct monster_race *race, bool sleep, bool group_ok,
struct monster_group_info group_info, byte origin)

In this case, if the mob is already asleep, the wand of clone monster successfully speeds up the mob, but does not clone it because the mob is asleep and generate.c cannot place a "wake", cloned mob asleep in a grid square. So, I believe the line in mon-move.c -

bool multiply_monster(struct chunk *c, const struct monster *mon)

simply returns FALSE without dropping a new mob or a dump file to debug. Good code, possibly faulty logic.

How to make Late Game Mages better.
 

I've made my suggestion numerous times, but I'll reiterate it once more.

In order to improve End Game interest / fun AND to improve Mage and spell-casters all around, I have put it to the Development Crew to have classes acquire spells in a spell-scroll system, adding them to binders.

This way, spell-casters do not crank up power in huge swaths as they currently do with the Book-at-once system. Given that the powerful books are found when the spell-casters are typically at or beyond the class levels necessary to learn them, they tend to learn the Book-at-Once. Hence, finding Raal's in 3.x or Arcane in 4.x for Mage classes equates to a phenomenal increase in power for the player ... just for picking up a copy.

Moving to a spell-scroll system, spell casters will only increase power piece-meal as scrolls are found, learned, and bound to binders. Otherwise, they're simply finding single-use scrolls, e.g., like a Scroll of Protection From Evil, read-once, consumed, gone forever. The UI knows the difference if the player "G"ains the scroll (and the system then prompts the player for a Binder to attach it to) or "r"eads it for 1-time use.

Spell-casters will be able to learn ~80% of the total spells, but the total number of spells out there should increase from the current ~40-some per class to 60-80. More utility spells. More attack spells. More defense spells per class. Say Max level 50 Mage can learn a total of 50-52 spells. Higher level spell scrolls will be rarer to find (but, should be easier to come across than the current rarity of High Level Books). Nearly all spells should be find-able by dungeon level 80.

You are not penalizing Mages. You are not bolstering other classes (warriors, thieves, etc.). You are giving the spell-casters (and abilities for thief/warrior classes, if one wishes to pursue those classes as well) more flexibility, more choice, more strategy in the game. You increase the level of interest into the Late-Game without penalizing a soul, without giving mobs yet another indefensible attack (Time, Gravity, Water, Force, Plasma, Mana, and so on) or another 10000 of health and +15 speed and 40 mob summons. You're also not giving warriors 14 extra swings per round or +3 extra Movement steps in attempt to make warriors somehow as good as mages supposedly are.

I can't reproduce that with the current development version: using a clone monster wand on an asleep monster works just as well as on an awake one. That agrees with what I see in the code: multiply_monster() doesn't check whether the monster being cloned is asleep or awake and place_new_monster() doesn't either - it's only told that the new monster will be awake when it is placed.

What can cause the cloning part of the activation to fail is not finding an empty square that's within one grid of the monster and within line of sight. So, using the wand on a monster in a dead end of a corridor when you're right next to it is guaranteed not to generate copy. With an empty space between you and the monster, it's virtually certain (1 - (8/9)^18000) that you'll get a copy.

Moreover, the raw docs are quite readable and sitting in the docs directory, and backwardsEric has included instructions explaining how to compile the docs locally yourself. Of course, the instructions are in the docs...

And I'll reiterate my answer, which is in short "Intriguing idea, I'm not totally revamping the game when the paint's not even dry on the last revamp".

Wasn't an issue with finding a grid square. Try this---the mob was not an original mob on the level, but was generated later on into the level, asleep.

Nexus Effect Altered?
 

@Wobbly prompted. I answered.


Next up, when / why was NEXUS altered to be Up/Down level only?
No more stat-scramble or just being sent away ~50-60 squares away?