[DONE] [CR] Root Cellar - food preserving option. #24083

nexusmrsep · 2018-06-24T18:27:45Z

What does it do?

This PR introduces:

1. New constructable terrain: root cellar
2. Mechanics for root cellar to store food in temperature equal to underground temperature. See: PR #24073
3. Side implementation: Two constant temperatures used in calculation were moved to game constants.

Racionale:

Part 1: General idea

Discussion under PR #23986 has led to new ideas, one of which is need for more options to preserve food. It led to an obvious conclusion, that most common food preservation method, nowadays and in the past, is temperature control. Having no electric fridges however, people in the past ages used various methods to keep food cool, of which most common one was using actual ground's temperature in form of dugout shelters, root cellars, etc. Nowadays this form of food preservation still live in form of household basement food storages and wine cellars.

This PR introduces root cellars, and allow them to store food in reduced temperatures (equal to underground temperature = 43F / 6C, compare #24073). Combined with #23986 it offsets the introduction of wide-range of perishable comestibles, alowing long-term storage at reduced decay rates.

Part 2: Sources
Here is a very good Wikipedia article on its purpose.
Compare: here Page 101-102.
and here. Other good source: here
Also: here, and one more.

See intro's and discussion of #23986 and #24073.

Part 3: Why did i use those values?
Construction requires:

skill 4 survivalist (to know how does it work) and skill 4 fabrication (to know how to build it without collapsing it on your head)
digging tool and a previously dug deep pit - most important principle of the root cellar is to dig it into the ground, where it can't exchange temperature with earth
rocks or bricks - they provide rudimentary support preventing dirt walls from collapsing and are a base/support for the insulating roof,
hammering tool, 2x4's or heavy sticks (roof material, trapdoor), withered plants of piles of straw (insulation) - another principle is to insulate the top of the dugout from outside temperature, by making the roof; roof is then covered with dirt and earth (& rocks, sod), which provides additional insulation. Since dirt is not represented in game, a assume that what was dug out from the pit now lays on the roof, also some of withered plants and rocks may be considered as used for this purpose.
Think about the final product as a miniature hobbit house, just not as fancy.

[source: https://www.ruralintelligence.com/images/food/root-cellar-prairie440.jpg]

Implementation:

1. Added new 'root cellar' entry to terrain.json
2. Added new construction recipe 'Build Root Cellar' to construction.json
3. This allowed me to modify decay calculators to check if food is placed in root cellar (terrain check), and if so - to calculate rot using underground temperature, as stated above.

[WARNING and HELP NEEDED] This however does NOT work, although it should be plain and simple. This is why I set [WIP] and [CR] on this PR and humbly ask for help. However I would write if ( g->m.tername( location ) == "root cellar" ) in weather.cpp in get_rot_since() function, it would not return proper value. I compiled few versions and added debug messages (not present in pushed code) that would 'print' string for g->m.name(location), g->m.tername(location), and few variants avalaible in map.h/map.cpp, but they returned text: "nothing". Not "" (null), without errors, just plain string "nothing". Did the same to print tripoint x.y.z values of argument location pushed to get_rot_since() function, and it seems to give some coordinates. Spent too many hours trying to crack it down. It must be something trivial that i cant see, or something convoluted, like coordinantes conversion that happen somewhere in between. Other parts of the code seem to use the same method to extract terrain name at a tripoint location, so I'm puzzled. Please help.

EDIT: Thanks @Night-Pryanik and @Coolthulhu for setting me on the right path to answer for this.

4 As a side-job I moved base annual temperature to game_constants.h, and did the same with starting temperature value. This may be an overkill but it seems rational and better then leaving naked number inside the code itslef.

Leland · 2018-06-24T23:14:50Z

data/json/construction.json

+    },{
+        "type" : "construction",
+        "description" : "Build Root Cellar",
+        "category" : "CONSTRUCT",


Recommend using the pre_note to describe the purpose - as a player I'd be mystified.

Night-Pryanik · 2018-06-25T05:23:24Z

I think you should use something like this:

ter_id tid = g->m.ter( location );
if( tid ) = t_rootcellar { 
    ... ;
}

Coolthulhu · 2018-06-25T16:25:48Z

src/weather.cpp

    time_duration ret = 0;
+    // root cellars are considered as underground storage, ignores weather
+    if ( g->m.tername( location ) == "root cellar" ) {


This is called often - at least once per item. This check may be expensive since it affects all such calls.

Doesn't work anyway in my test compilation, nor for @Night-Pryanik 's idea above.
Currently testing it to its limits to find what's going on.
I might have a lead...

The front of my test function now looks like this:

ter_id terrain = g->m.ter( location ); add_msg( m_info, "Debug TP X: %s", location.x); add_msg( m_info, "Debug TP Y: %s", location.y); add_msg( m_info, "Debug TP Z: %s", location.z); add_msg( m_info, "Debug MS: %s", g->m.getmapsize() ); if ( terrain == t_rootcellar ) { add_msg( "Debug 1: PING" ); add_msg( m_info, "Debug 1: T: %s", terrain.id().c_str() ); add_msg( m_info, "Debug 1: RC: %s", t_rootcellar.id().c_str() ); for( time_point i = start; i < end; i += 1_hours ) { ret += std::min( 1_hours, end - i ) / 1_hours * get_hourly_rotpoints_at_temp( AVERAGE_ANNUAL_TEMPERATURE ) * 1_turns; } return ret; } else { add_msg( "Debug 1: PONG" ); add_msg( m_info, "Debug 2: T: %s", terrain.id().c_str() ); add_msg( m_info, "Debug 2: RC: %s", t_rootcellar.id().c_str() ); } if ( terrain == t_pit ) { add_msg( "Debug 2: PIT PIT PIT" ); }

RESULT:
X = 678, Y = 1983, Z = 0
MS=11 (?)
Debug 2: T: t_null
Debug 2: RC: t_rootcellar

t-null --> this suggests that map::ter() returns t_null, and that might be becouse XYZ are !inbound(), but why? maybe the ~~strange getmapsize() result~~? [edit: not at all strange, its a game constant...]

The huge coords mean that you're getting either absolute coords or coords in overmap. But map::ter expects local coords, which are not always available here (rot processing can happen when local map doesn't even exist).
This means that the approach with root cellar detection at rot calculation level can't work reliably.

I found the answer: 'tripoint location' argument pushed to the function should be first converted by map:getlocal() to local coordinates. I knew it, I knew it must have been coordinates conversion

This is called often - at least once per item. This check may be expensive since it affects all such calls.

How to avoid it then? I think it has to check where the item is located. Also for the sake of future PRs, if I follow Kevin's idea on storing food in proper furniture, I would use this method to discern where is it located, and if such location is valid.

You'd need to pass it from above, where item location in local coords is still available.
If you do it this way, make sure all paths go through the same functions, to avoid hard to find bugs where rot is different depending on how is it processed.

Coolthulhu · 2018-06-25T16:27:01Z

The special case of root cellar is different from items just stored underground. Instead, the root cellar always keeps the items at exact temperature, regardless of all environment.

nexusmrsep · 2018-06-25T17:33:14Z

Ok, after few hours I managed to narrow down and fix the problem. Map coordinates conversion is a pain! Thank you very much @Night-Pryanik and @Coolthulhu for reviewing and pointing me to the right direction on this.

Leland · 2018-06-25T17:49:09Z

data/json/construction.json

@@ -2390,7 +2390,8 @@
            [ [ "2x4", 6 ], [ "stick", 6 ] ],
            [ [ "withered", 12 ], [ "straw_pile", 12 ] ]
        ],
-        "pre_terrain" : "t_pit",
-        "post_terrain" : "t_rootcellar"
+        "pre_note":  "You need a deep pit to construct a root cellar.",


This is helpful but not exactly what I was referring to.

A "root cellar's" purpose isn't evident in its name. A user would have to construct the terrain and then examine it (and read the description) in order to determine it's purpose.

The pre_note would be a good place to make that evident from the beginning.

For instance, You first need a deep pit to construct a root cellar, and enable the storage of food in a cold environment.

Wordy, but you get the gist :)

@Leland
No, it would'nt. The sole purpose of a pre-note is to give some info on extra costruction/deconstruction requrements.
What you ask is actualy true for all constructions. Solution to this would be expanding construction menu with the description of a constructed item, so you will get the idea of what actualy you are constructing.

This deserves a separate PR and i will try to approach it, as it bothers me too.

nexusmrsep · 2018-06-30T09:55:11Z

@Coolthulhu

You'd need to pass it from above, where item location in local coords is still available.
If you do it this way, make sure all paths go through the same functions, to avoid hard to find bugs where rot is different depending on how is it processed.

Following your statement I reworked functions above calc_rot() to push local coordinates (instead of global abs) down to it, and in consequence down to get_rot_since(). Then get_rot_since() either uses those coordinates or converts them to global by getabs() where needed.

This I believe solves the problem you mentioned.

Coolthulhu · 2018-06-30T13:52:21Z

You forgot at least one use case of calc_rot. Double check everything - changing coord type is one of the least safe operations, because the compiler won't warn that you're supplying the wrong type.

Coolthulhu · 2018-06-30T13:54:16Z

src/weather.cpp

        }
        return ret;
    }
    // if on- or above-ground it uses progressive weather-determined temperatures at location
    const auto &wgen = g->get_cur_weather_gen();
    for( time_point i = start; i < end; i += 1_hours ) {
-        w_point w = wgen.get_weather( location, i, g->get_seed() );
+        w_point w = wgen.get_weather( g->m.getabs( location ), i, g->get_seed() );


Extract the getabs to a variable to avoid calculating it in a loop if the compiler fails to notice that it doesn't change.

Coolthulhu · 2018-06-30T13:56:00Z

src/weather.cpp

+    // root cellars are considered as underground storage, ignores weather, constant temperature
+    if ( g->m.ter( location ) == t_rootcellar ) {
+        for( time_point i = start; i < end; i += 1_hours ) {
+        ret += std::min( 1_hours, end - i ) / 1_hours * get_hourly_rotpoints_at_temp( AVERAGE_ANNUAL_TEMPERATURE ) * 1_turns;


Scrap the loop. Just calculate it all in one go.
Also, indentation.

nexusmrsep · 2018-06-30T15:40:52Z

Extract the getabs to a variable to avoid calculating it in a loop if the compiler fails to notice that it doesn't change.
Scrap the loop. Just calculate it all in one go.

Done & done. Also I merged underground rot calculation into one sub-routine as this is more elegant way to do it, when the previous results would be the same anyway.

You forgot at least one use case of calc_rot. Double check everything - changing coord type is one of the least safe operations, because the compiler won't warn that you're supplying the wrong type.

I've double-checked and all calc_rot calls seems covered:

player::eat
player::process_food
map::has_rotten_away (with recursive use)
item::process_food

Where is the missing one?

Coolthulhu · 2018-06-30T16:19:31Z

src/weather.cpp

+    // root cellars are considered as underground storage, ignores weather, constant temperature
+    if( g->m.ter( location ) == t_rootcellar || location.z < 0 ) {
+        ret = ( end - start ) / 1_hours * get_hourly_rotpoints_at_temp( AVERAGE_ANNUAL_TEMPERATURE ) *
+              1_turns;


Don't assign ret if you return it on the next line. Move the ret declaration below the block, in this block return the value you'd assign to ret without assigning it.

Also, check for location.z before checking for terrain type. If one check in if is orders of magnitude faster and more common than the other, it should be moved to front so that it is eliminated faster.

good point, thanks

nexusmrsep · 2018-06-30T16:22:04Z

I have also done some rudimentary testing of the compiled version: in summer season one pizza was dropped in the root cellar, another was dropped on adjacent tile. As expected one in the cellar was decaying in a slower pace that the one outside. No apparent errors occured.

This reverts commit fdc1a99.

nexusmrsep · 2018-07-17T16:56:13Z

Can I ask for the update on this PR? [edit: on the current state. Is there more I need to do on my side?]

ZhilkinSerg · 2018-07-19T20:10:32Z

I will try to test it on weekend.

nexusmrsep · 2018-07-30T19:04:50Z

Side note: if #24555 works, then I think @Coolthulhu 's advice for pushing coordinates "from above" to temperature calculations would become obsolete and I could trim this PR by half, using instead the new_game trick (root cellar check would be done only if new_game == false).

This reverts commit f0abb03.

This reverts commit c8d88e9.

nexusmrsep · 2018-08-01T15:41:31Z

Word of comment here: the "revert revert revert" commits are actually the above-mentioned trim after #24555. Since the new walk-around method worked there, I discarded all changes that pushed location data from above to get_rot_since from calc_rot, while applying the new method in conflict resolving branch merging commit.

That is good news overall, as this PR is now not as risky, and in fact pretty straightforward.

ZhilkinSerg · 2018-08-01T17:02:39Z

Two nachos (one in root cellar and one in player inventory) rot at the same speed.

nexusmrsep · 2018-08-01T17:04:35Z

Two nachos (one in root cellar and one in player inventory) rot at the same speed.

I'll take a look in a minute.

nexusmrsep · 2018-08-01T17:19:59Z

Fixed. It seems in all this pursuit after current upstream I didn't noticed that one of tripoints were not in the right scale.

ZhilkinSerg · 2018-08-01T17:28:22Z

Looks good now.

nexusmrsep · 2018-08-01T17:30:17Z

Looks good now.

Glad to hear that. Hallelujah.

sfsworms · 2018-08-03T20:34:15Z

Mmmmh, there seems to be an issue with NPC: they don't recognize the root cellar as a furniture and when asked to clean the camp they remove the food from there.

ZhilkinSerg · 2018-08-03T20:46:54Z

That is because root cellar is terrain, not a furniture.

sfsworms · 2018-08-04T11:11:40Z

Oh, make sense. Slightly annoying though. Should I create an issue for it?

nexusmrsep added 4 commits June 24, 2018 12:33

root cellar object and construction recipe

290a2ca

Merge remote-tracking branch 'upstream/master' into rootcellar

9181760

resolved conflicts

050cd42

different approach, moving const temp values to game_constatnts

bf7ba1c

Leland reviewed Jun 24, 2018

View reviewed changes

Coolthulhu reviewed Jun 25, 2018

View reviewed changes

fixed function, pre_note, added headers

8f40f89

corrected number of " ) "

80dba0d

Leland reviewed Jun 25, 2018

View reviewed changes

nexusmrsep mentioned this pull request Jun 29, 2018

[Rdy] [CR] Immortal food overhaul - long lasting but rotting food. #23986

Merged

global/local coordinates rework for rot calculation

1a1e59c

nexusmrsep changed the title ~~[WiP] [CR] Root Cellar - food preserving option.~~ [Rdy] [CR] Root Cellar - food preserving option. Jun 30, 2018

astyle fix

98a7f4e

Coolthulhu reviewed Jun 30, 2018

View reviewed changes

nexusmrsep added 2 commits June 30, 2018 17:27

function compacting for optimization

44f2acf

astyle

be4313d

Coolthulhu reviewed Jun 30, 2018

View reviewed changes

more optimizations

7216bef

ZhilkinSerg self-assigned this Jul 1, 2018

nexusmrsep mentioned this pull request Jul 2, 2018

Future of food preservation. #24145

Closed

ZhilkinSerg removed their assignment Jul 8, 2018

Merge branch 'master' into rootcellar

6c22ada

Revert "revert of coordinates conversion for calc_rotget"

f0abb03

This reverts commit fdc1a99.

Merge branch 'master' into rootcellar

f8b7fa8

nexusmrsep changed the title ~~[Rdy] [CR] Root Cellar - food preserving option.~~ [DONE] [CR] Root Cellar - food preserving option. Jul 29, 2018

ZhilkinSerg added the (P2 - High) High priority (for ex. important bugfixes) label Jul 29, 2018

nexusmrsep mentioned this pull request Jul 30, 2018

Fix for temperature related: chargen crash (in labs) and ice lab food decay bug #24555

Merged

nexusmrsep and others added 3 commits August 1, 2018 17:20

Merge branch 'master' into rootcellar

d406882

Revert "Revert "revert of coordinates conversion for calc_rotget""

9dfac50

This reverts commit f0abb03.

Revert "Revert "reverts part 2""

f05e491

This reverts commit c8d88e9.

fix for location depth conversion

9840e69

ZhilkinSerg merged commit 2298186 into CleverRaven:master Aug 1, 2018

nexusmrsep deleted the rootcellar branch August 1, 2018 17:30

DracoGriffin mentioned this pull request Aug 3, 2018

Use github's multiple issue template feature. #24615

Merged

int-ua mentioned this pull request Jun 15, 2020

[CR] Skip unnecessary weather calculations for root cellar #41324

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[DONE] [CR] Root Cellar - food preserving option. #24083

[DONE] [CR] Root Cellar - food preserving option. #24083

nexusmrsep commented Jun 24, 2018 •

edited

Loading

Leland Jun 24, 2018 •

edited

Loading

Night-Pryanik commented Jun 25, 2018

Coolthulhu Jun 25, 2018

nexusmrsep Jun 25, 2018 •

edited

Loading

Coolthulhu Jun 25, 2018

nexusmrsep Jun 25, 2018 •

edited

Loading

nexusmrsep Jun 25, 2018 •

edited

Loading

Coolthulhu Jun 25, 2018

Coolthulhu commented Jun 25, 2018

nexusmrsep commented Jun 25, 2018

Leland Jun 25, 2018

nexusmrsep Jun 25, 2018

nexusmrsep commented Jun 30, 2018 •

edited

Loading

Coolthulhu commented Jun 30, 2018

Coolthulhu Jun 30, 2018

Coolthulhu Jun 30, 2018

nexusmrsep commented Jun 30, 2018 •

edited

Loading

Coolthulhu Jun 30, 2018

nexusmrsep Jun 30, 2018

nexusmrsep commented Jun 30, 2018

nexusmrsep commented Jul 17, 2018 •

edited

Loading

ZhilkinSerg commented Jul 19, 2018

nexusmrsep commented Jul 30, 2018

nexusmrsep commented Aug 1, 2018

ZhilkinSerg commented Aug 1, 2018

nexusmrsep commented Aug 1, 2018

nexusmrsep commented Aug 1, 2018

ZhilkinSerg commented Aug 1, 2018

nexusmrsep commented Aug 1, 2018

sfsworms commented Aug 3, 2018

ZhilkinSerg commented Aug 3, 2018

sfsworms commented Aug 4, 2018

[DONE] [CR] Root Cellar - food preserving option. #24083

[DONE] [CR] Root Cellar - food preserving option. #24083

Conversation

nexusmrsep commented Jun 24, 2018 • edited Loading

What does it do?

Racionale:

Implementation:

Leland Jun 24, 2018 • edited Loading

Choose a reason for hiding this comment

Night-Pryanik commented Jun 25, 2018

Choose a reason for hiding this comment

nexusmrsep Jun 25, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nexusmrsep Jun 25, 2018 • edited Loading

Choose a reason for hiding this comment

nexusmrsep Jun 25, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Coolthulhu commented Jun 25, 2018

nexusmrsep commented Jun 25, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nexusmrsep commented Jun 30, 2018 • edited Loading

Coolthulhu commented Jun 30, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nexusmrsep commented Jun 30, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nexusmrsep commented Jun 30, 2018

nexusmrsep commented Jul 17, 2018 • edited Loading

ZhilkinSerg commented Jul 19, 2018

nexusmrsep commented Jul 30, 2018

nexusmrsep commented Aug 1, 2018

ZhilkinSerg commented Aug 1, 2018

nexusmrsep commented Aug 1, 2018

nexusmrsep commented Aug 1, 2018

ZhilkinSerg commented Aug 1, 2018

nexusmrsep commented Aug 1, 2018

sfsworms commented Aug 3, 2018

ZhilkinSerg commented Aug 3, 2018

sfsworms commented Aug 4, 2018

nexusmrsep commented Jun 24, 2018 •

edited

Loading

Leland Jun 24, 2018 •

edited

Loading

nexusmrsep Jun 25, 2018 •

edited

Loading

nexusmrsep Jun 25, 2018 •

edited

Loading

nexusmrsep Jun 25, 2018 •

edited

Loading

nexusmrsep commented Jun 30, 2018 •

edited

Loading

nexusmrsep commented Jun 30, 2018 •

edited

Loading

nexusmrsep commented Jul 17, 2018 •

edited

Loading