Angband 4.2.0

[gglibertine]

So, that pesky and elusive run-past bug still exists. My new @ just ran past some copper coins and had to back up to retrieve them.

I am *only* noticing this happening with copper coins; other coins get picked up. Maybe an anomaly, maybe a clue?

[Pete Mack]

Was it coins or a monster?

[Ingwe Ingweron]

Was it coins or a monster?

In my case coins.

[gglibertine]

In my case coins.

Same -- copper coins. Some dropped by monsters, others were just lying there.

[gglibertine]

I am *only* noticing this happening with copper coins; other coins get picked up. Maybe an anomaly, maybe a clue?

Disregard this. It seemed like there was a pattern, but further experimentation doesn't bear it out.

[Pete Mack]

Is it repeatable? That is, if you try running by the coins in the same direction a second time, does it happen again? If so, a save file is possible.

[gglibertine]

Is it repeatable? That is, if you try running by the coins in the same direction a second time, does it happen again? If so, a save file is possible.

Nope, the second time @ successfully picks up the treasure. Quite a frustrating bug to nail down, this one. I'm keeping an eye out for any repeat behaviour.

[wobbly]

I can produce a version of this. Stand next to gold or an object that would normally be auto-picked up. Wizard mode in a monster to interrupt run. Walking onto the object will pick it up. Running onto it wont.

[PowerWyrm]

Another bug in effect_handler_EARTHQUAKE() spotted while porting the struct loc refactoring (commit 7be84e3):

old code:

y = py + ddy_ddd[i];
x = px + ddx_ddd[i];
if (map[16 + y - centre.y][16 + x - centre.x]) continue;

new code:

struct loc grid = loc_sum(pgrid, ddgrid_ddd[i]);
if (map[16 + y - centre.y][16 + x - centre.x]) continue;

This leaves x and y uninitialized (in fact they should be equal to 32), breaking earthquakes.

Fix: replace with grid.y and grid.x.

[spara]

More random feedback. I like the way the new books have been named so that the school is obvious. It's also nice that the description of the book says if @ can read the book. I did not notice that at first though, since the the first necro book had nothing in the description and I assumed the others did not have anything either in their description. The most awesome thing would be, if all books had some kind of flavor text and would mention which classes can read the book.

I understand that coming up with good flavor texts is much more work than one would assume, so at least adding notification about the classes that can read the book at some point would make the game a bit more approachable.

[Oraticus]

Another annoying bug... it used to be that if you were casting spells, you could hit a direction and the spell would travel that direction, allowing you to blind-fire down a corridor if you knew enemies were on their way down. I tried it today with lightning strike, and when I hit "right", it just blew up at my feet. I was using the num pad, so I tried it with num lock both on and off, and I even tried it with the arrow keys. Just burst at my feet each time. Maybe other spells work (Raise earth works, but that's because it beams anyways)... maybe just ball spells? Anyone able to test, or seeing this same behavior?

[fph]

I think that the idea is that lightning strike comes "from the sky" (or the dungeon ceiling?) like an air strike, instead of travelling in a straight line from your hands. (I didn't check, but I'd be curious to see if you can target a tile you can't see behind a corner).

[Nick]

I think that the idea is that lightning strike comes "from the sky" (or the dungeon ceiling?) like an air strike, instead of travelling in a straight line from your hands. (I didn't check, but I'd be curious to see if you can target a tile you can't see behind a corner).

That's correct. Lightning strike can target any grid the player can see, but you need to pick an actual grid to hit. I'll see if I can make that a bit clearer.

[Nick]

Another bug in effect_handler_EARTHQUAKE() spotted while porting the struct loc refactoring (commit 7be84e3):

old code:

y = py + ddy_ddd[i];
x = px + ddx_ddd[i];
if (map[16 + y - centre.y][16 + x - centre.x]) continue;

new code:

struct loc grid = loc_sum(pgrid, ddgrid_ddd[i]);
if (map[16 + y - centre.y][16 + x - centre.x]) continue;

This leaves x and y uninitialized (in fact they should be equal to 32), breaking earthquakes.

Fix: replace with grid.y and grid.x.

Thanks, as usual

For everyone else reporting stuff, I am reading this thread and will get to it. Eventually.

[Ingwe Ingweron]

...will get to it. Eventually.

Beleriand?