[Crawl] Patch Bug Report: Explore and Magic Mapping

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

Yet another (minor) patch bug:
When you read or cast Magic Mapping on a fully explored level, then
Explore (Ctrl+O) acts like on an unexplored level. Fully explored means
that Magic Mapping didn't reveal any new terrain. Should be easy to
confirm...

Rubinstein

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

Rubinstein <picommander@t-online.de> wrote in news:csmedf$hks$02$1@news.t-
online.com:

> Yet another (minor) patch bug:
> When you read or cast Magic Mapping on a fully explored level, then
> Explore (Ctrl+O) acts like on an unexplored level. Fully explored means
> that Magic Mapping didn't reveal any new terrain. Should be easy to
> confirm...

I think what's happening is that Magic Mapping removes items from the map
memory, and replaces them with unexplored tiles. So when you Ctrl-O, it
bounces around to every space that had an item, checking to see what's
there.

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

Chipacabra wrote:

> Rubinstein <picommander@t-online.de> wrote in news:csmedf$hks$02$1@news.t-
> online.com:
>
> > Yet another (minor) patch bug:
> > When you read or cast Magic Mapping on a fully explored level, then
> > Explore (Ctrl+O) acts like on an unexplored level. Fully explored means
> > that Magic Mapping didn't reveal any new terrain. Should be easy to
> > confirm...
>
> I think what's happening is that Magic Mapping removes items from the map
> memory, and replaces them with unexplored tiles. So when you Ctrl-O, it
> bounces around to every space that had an item, checking to see what's
> there.

As much as I hate to write a me-too: I agree with Chip. If you look at the map
after magic mapping, you'll see single-dots ("magic mapped but not explored")
in some? all? squares in explored territory where previously known items are
laying. So... bug? Yes. Patch bug? My impression is that no, this was going on
before the patch, but I have no proof for that impression.

Erik

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

Erik Piper wrote:
> Chipacabra wrote:
>
>> Rubinstein <picommander@t-online.de> wrote in
>> news:csmedf$hks$02$1@news.t- online.com:
>>
>> > Yet another (minor) patch bug:
>> > [MM on already explored area, then 'Explore']
>>
>> I think what's happening is that Magic Mapping removes items from the
>> map memory, and replaces them with unexplored tiles.

If this is true, it's a bug in Crawl itself, isn't it?

> If you look at the map after magic mapping, you'll see single-dots
> ("magic mapped but not explored") in some? all? squares in explored
> territory where previously known items are laying. So... bug? Yes.
> Patch bug? My impression is that no, this was going on before the
> patch, but I have no proof for that impression.

I didn't watch this before after MM in a partial explored area, but I
probably missed it. But I'm sure it's not the case in a completely
explored area: everything is in place (after casting or reading a scroll
of MM), nothing changed, but 'Explore' acts like everything has been
forgotten. Maybe you're right and the bug doesn't origin from the patch
but from the main code, I don't know...

Rubinstein

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

Rubinstein <picommander@t-online.de> wrote:
> Erik Piper wrote:

>> If you look at the map after magic mapping, you'll see
>> single-dots ("magic mapped but not explored") in some? all?
>> squares in explored territory where previously known items are
>> laying.

All magic-mapped squares with items on them. Rubinstein will
probably not see the single-dot effect because Crawl uses straight
ASCII on Linux.

> I didn't watch this before after MM in a partial explored area,
> but I probably missed it. But I'm sure it's not the case in a
> completely explored area: everything is in place (after casting or
> reading a scroll of MM),

Do you see all items in your vicinity, but outside your LOS when you
hit X for the level-map?

> nothing changed, but 'Explore' acts like everything has been
> forgotten. Maybe you're right and the bug doesn't origin from the
> patch but from the main code, I don't know...

It's a feature. The original behaviour of explore was to consider
magic-mapped squares as fully explored. That was considered
unsatisfactory for various reasons, so the new behaviour of treating
magic-mapped squares as unexplored was introduced. As you've
observed, it's less than perfect when magic-mapping already-explored
territory, but it's the best that's possible without getting intimate
with the magic-mapping code.

Darshan

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

Darshan Shaligram wrote:
> Rubinstein <picommander@t-online.de> wrote:
>> Erik Piper wrote:
>
>>> If you look at the map after magic mapping, you'll see single-dots
>>> ("magic mapped but not explored") in some? all? squares in explored
>>> territory where previously known items are laying.
>
> All magic-mapped squares with items on them. Rubinstein will probably
> not see the single-dot effect because Crawl uses straight ASCII on
> Linux.

Reminds me of my early Nethack days, where I was playing the ASCII
version in X-terminal sometimes. I just tried this with Crawl (I'm
usually playing in Linux terminal), but there's no difference here.

>> I didn't watch this before after MM in a partial explored area,
>> but I probably missed it. But I'm sure it's not the case in a
>> completely explored area: everything is in place (after casting or
>> reading a scroll of MM),
>
> Do you see all items in your vicinity, but outside your LOS when you
> hit X for the level-map?

I see *all* (formerly seen) items, not only those in my LOS.

>> nothing changed, but 'Explore' acts like everything has been
>> forgotten. Maybe you're right and the bug doesn't origin from the
>> patch but from the main code, I don't know...
>
> It's a feature. The original behaviour of explore was to consider
> magic-mapped squares as fully explored. That was considered
> unsatisfactory for various reasons, so the new behaviour of treating
> magic-mapped squares as unexplored was introduced. As you've observed,
> it's less than perfect when magic-mapping already-explored territory,
> but it's the best that's possible without getting intimate with the
> magic-mapping code.

Ok, so shall it be (I really didn't want to wake sleeping dogs). Btw, I
wonder why nobody asked why I'm casting MM on an already explored level.
Easy answer: before casting MM you can't be quite sure whether it's
completely explored (other reasons like rising the divination skill also
comes to mind).
I'm still somewhat curious about the "various reasons" you mentioned
above. Right now I even can't imagine a single one...

Rubinstein

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

bork bork bork Rubinstein bork 1:34:51 PM bork 1/20/2005 bork bork:

[...]

> > It's a feature. The original behaviour of explore was to consider
> > magic-mapped squares as fully explored. That was considered
> > unsatisfactory for various reasons, [...]

> I'm still somewhat curious about the "various reasons" you mentioned
> above. Right now I even can't imagine a single one...

Just one? That's simple enough. Magic Mapping doesn't show if a square has
items on it; it just shows what terrain it is. Obviously, you want to know if
there are indeed items in the magic-mapped terrain.

Can't think of any others, though.

> Rubinstein

I'm surprised that you're not running into the same "tiles that are mapped
and holding items becomes merely magic-mapped if magic mapping is applied to
them" bug that I am. Maybe some Windows/Unix difference again (yeah, blame it
on that! Blame it on that!)

Erik

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

Erik Piper wrote:
> bork bork bork Rubinstein bork 1:34:51 PM bork 1/20/2005 bork bork:
>
> [...]
>
>> > It's a feature. The original behaviour of explore was to consider
>> > magic-mapped squares as fully explored. That was considered
>> > unsatisfactory for various reasons, [...]
>
>> I'm still somewhat curious about the "various reasons" you mentioned
>> above. Right now I even can't imagine a single one...
>
> Just one? That's simple enough. Magic Mapping doesn't show if a square
> has items on it; it just shows what terrain it is.

True, but it also doesn't remove already known items from the X-map,
which is what 'Explore' is acting like...

> Obviously, you want to know if there are indeed items in the
> magic-mapped terrain.

I still don't get it: the items you can see in the X-map (after Magic
Mapping) are *only* the already known items. Why should 'Explore' behave
weird because of that?
The problem IMO is that 'Explore' isn't able to distinguish between
Magic Mapped and already known area. It seems to treat *all* squares
(including the already known ones) covered by MM as newly explored
area.

> I'm surprised that you're not running into the same "tiles that are
> mapped and holding items becomes merely magic-mapped if magic mapping
> is applied to them" bug that I am. Maybe some Windows/Unix difference
> again (yeah, blame it on that! Blame it on that!)

Brillant idea! Let us start an OS war, preferably in the DoomRL
section... <evil grin>

Rubinstein

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

bork bork bork Rubinstein bork 3:17:49 PM bork 1/20/2005 bork bork:

[...]

> > [...]
> >
> >> > It's a feature. The original behaviour of explore was to consider
> >> > magic-mapped squares as fully explored. That was considered
> >> > unsatisfactory for various reasons, [...]
> >
> >> I'm still somewhat curious about the "various reasons" [Darshan mentions]
> >> above. Right now I even can't imagine a single one...
> >
> > Just one? That's simple enough. Magic Mapping doesn't show if a square
> > has items on it; it just shows what terrain it is.
>
> True, but it also doesn't remove already known items from the X-map,
> which is what 'Explore' is acting like...

It shouldn't -- but that's not what I was talking about above anyway. But I
thought you were saying that it was always behaving properly (not replacing
known-item squares on the X-map with merely-magic-mapped squares) for you
anyway?!?

> > Obviously, you want to know if there are indeed items in the
> > magic-mapped terrain.
>
> I still don't get it: the items you can see in the X-map (after Magic
> Mapping) are only the already known items. Why should 'Explore' behave
> weird because of that?

Let me put my point another way.

Case 1: Explore treats merely-magic-mapped squares the same as truly-mapped
squares. Some of those squares will have items on them. Sooner or later, if
you trust Explore to explore for you, you'll miss some squares with items
because of this.

Case 2: Explore treats merely-magic-mapped squares the same as unexplored
territory. The above stuff... uh, doesn't happen.

NOW do you see why...

....err, you DO know that "explore" means "Ctrl-O", not "the X-map", right?

> The problem IMO is that 'Explore' isn't able to distinguish between
> Magic Mapped and already known area. It seems to treat all squares
> (including the already known ones) covered by MM as newly explored
> area.

[...]

Erik

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

"Erik Piper" <efrniokr@sdky.cz> wrote in
news:35c2ccF4j8r7kU1@individual.net:

> Let me put my point another way.
>
> Case 1: Explore treats merely-magic-mapped squares the same as
> truly-mapped squares. Some of those squares will have items on them.
> Sooner or later, if you trust Explore to explore for you, you'll miss
> some squares with items because of this.
>
> Case 2: Explore treats merely-magic-mapped squares the same as
> unexplored territory. The above stuff... uh, doesn't happen.
>
> NOW do you see why...
>
> ...err, you DO know that "explore" means "Ctrl-O", not "the X-map",
> right?
>
>> The problem IMO is that 'Explore' isn't able to distinguish between
>> Magic Mapped and already known area. It seems to treat all squares
>> (including the already known ones) covered by MM as newly explored
>> area.
>
> [...]
>
> Erik
>

Okay, here's what I think is happening: When Magic Mapping goes off, it
leaves already explored areas alone and adds unexplored areas to the X-
map as known-but-unexplored. However, any area that contains an item is
marked as known-but-unexplored, even if it was fully explored before. So,
if you fully explore a level, and then cast MM, every single space on the
map with an item flips over to known-but-unexplored. Looking at the X-
map, they will show as known-but-unexplored. Ctrl-O will go check them
out. I can reproduce this behavior in my game 100% of the time, with
every version of Darshan's patch that I've used. I haven't tried it with
the stock game yet. Even if the unpatched game behaves the same, most
people wouldn't notice because it barely affects anything except Ctrl-O
exploring.

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

Chipacabra wrote:
> "Erik Piper" <efrniokr@sdky.cz> wrote in
> news:35c2ccF4j8r7kU1@individual.net:
>
>> Let me put my point another way.
>> [...]
>>
>>> The problem IMO is that 'Explore' isn't able to distinguish between
>>> Magic Mapped and already known area. It seems to treat all squares
>>> (including the already known ones) covered by MM as newly explored
>>> area.
>>
>> [...]
>
> Okay, here's what I think is happening: When Magic Mapping goes off,
> it leaves already explored areas alone and adds unexplored areas to
> the X- map as known-but-unexplored. However, any area that contains an
> item is marked as known-but-unexplored, even if it was fully explored
> before. So, if you fully explore a level, and then cast MM, every
> single space on the map with an item flips over to
> known-but-unexplored.

Sounds reasonable. If true, it also would mean it's not repairable
unless you go into the main code (not a problem of the patch itself).

> Looking at the X- map, they will show as known-but-unexplored.
^^^^
You really mean *show*? You actually see the differences on the X-map?
MM on already explored squares doesn't show any visible differences
here. Let me guess: you're also playing the windows version? Then it
probably has something to do with the odd behavior Erik is talking about
and which is not present in the Linux version...

Rubinstein

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

Rubinstein <picommander@t-online.de> wrote:
[...]
> As I stated above, it seems much better if 'Explore' simply(?)
> would ignore *all* effects of magic mapping.

That, unfortunately, is not yet possible. Explore currently has to
work by the same map that you see when you hit X, and that is
affected by magic mapping.

What does surprise me is that magic mapping doesn't overwrite item
memory on the X map for you; as far as I can see, it does, even on
Linux (the magic-mapping code is the same, after all). You should
see items replaced by commas on the level-map when you magic-map an
already mapped level. If you aren't seeing any difference at all,
neither will Explore, so I believe you're mistaken.

Darshan

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

Rubinstein <picommander@t-online.de> wrote in news:cstpvr$spo$03$1@news.t-
online.com:

>> Looking at the X- map, they will show as known-but-unexplored.
> ^^^^
> You really mean *show*? You actually see the differences on the X-map?
> MM on already explored squares doesn't show any visible differences
> here. Let me guess: you're also playing the windows version? Then it
> probably has something to do with the odd behavior Erik is talking about
> and which is not present in the Linux version...
>
> Rubinstein
>

Yes, in the windows version the X-map displays differently for explored and
for MMed. After casting MM, I can look at my X-Map and see every space that
used to have an item in it.

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

Darshan Shaligram wrote:
> Rubinstein <picommander@t-online.de> wrote:
> [...]
>> As I stated above, it seems much better if 'Explore' simply(?)
>> would ignore *all* effects of magic mapping.
>
> That, unfortunately, is not yet possible. Explore currently has to
> work by the same map that you see when you hit X, and that is
> affected by magic mapping.

Not "yet"? You have some plans to work on it?

> What does surprise me is that magic mapping doesn't overwrite item
> memory on the X map for you
> [...]

Oh oh, what have I done again!
I tried to prove what I was so sure about: fired a screen session to get
hardcopies from all parts of the game (also from X-map) but all I proved
was: the fool was me. Don't know why I didn't see it at first, probably
because of using the limited version from a low level casting (rather
than from the scroll).
At least I've got my punishment: all those testing was at the cost of my
last 2 precious scrolls of MM (fools don't bother with wizard-mode).
Sorry, guys... :-(

Rubinstein

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

Chipacabra wrote:
> Rubinstein <picommander@t-online.de> wrote in news:cstpvr$spo$03$1@news.t-
> online.com:
>
>>> Looking at the X- map, they will show as known-but-unexplored.
>> ^^^^
>> You really mean *show*?
>> [...]
>
> Yes, in the windows version the X-map displays differently for
> explored and for MMed. After casting MM, I can look at my X-Map and
> see every space that used to have an item in it.

Did you read my last posting?
My general big sorry also was aimed to you and Erik. ;-)
I hope you both accept my apologize...

Rubinstein

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

Rubinstein wrote:

> Erik Piper wrote:
> > bork bork bork Rubinstein bork 3:17:49 PM bork 1/20/2005 bork bork:
> >
> > [...]
> >
> >> > [...]
> >> >
> >> >> > It's a feature. The original behaviour of explore was to
> >> >> > consider magic-mapped squares as fully explored.
>
> Dunno how the code deales with fully explored/magic mapped squares, but
> wouldn't it be better (and more logical, if technically possibel) to let
> 'Explore' completely ignore any process of magic mapping?

It is (for at least the one reason that I have been trying, perhaps too
complicatedly, to explain). It didn't ignore it before. Now it does, precisely
because it's more logical.

Erik

 

[Guest]

Archived from groups: rec.games.roguelike.misc (More info?)

Darshan Shaligram wrote:

> Rubinstein <picommander@t-online.de> wrote:
> [...]
> > As I stated above, it seems much better if 'Explore' simply(?)
> > would ignore all effects of magic mapping.
>
> That, unfortunately, is not yet possible. Explore currently has to
> work by the same map that you see when you hit X, and that is
> affected by magic mapping.

But in *effect* it is already true -- MMed squares are explored just as if you
had never seen them. Of course, programming-wise, I'm sure they are not ignored.

Erik