Refactor fields (step 6) #32481
Refactor fields (step 6) #32481
Conversation
ZhilkinSerg
added
[JSON]
ZhilkinSerg
force-pushed
the
refactor-fields-step-6
branch
from
ad8a9ff to
83e74e3
Compare
| bool has_fume = false; | ||
| for( const auto &field : fields ) { | ||
| if( field.first == fd_incendiary || field.first == fd_fire ) { | ||
| has_fire = true; | ||
| } |
There was a problem hiding this comment.
Maybe add fd_fire_vent and fd_flame_burst fields too?
There was a problem hiding this comment.
It was fixed in following linked PR.
| intensity_levels.emplace_back( intensity_level ); | ||
| } | ||
| optional( jo, was_loaded, "underwater_age_speedup", underwater_age_speedup, 0_turns ); | ||
| optional( jo, was_loaded, "decay_amount_factor", decay_amount_factor, 0 ); | ||
| optional( jo, was_loaded, "apply_slime_factor", apply_slime_factor, 0 ); | ||
| optional( jo, was_loaded, "dirty_transparency_cache", dirty_transparency_cache, false ); |
There was a problem hiding this comment.
It seems to me that this dirty_transparency_cache bool may be able to be determined automatically at runtime based on whether the field emits light or is translucent. Determining it automatically would imo be more robust than requiring people modifying/adding new fields to set it correctly.
There was a problem hiding this comment.
I agree. That is not final state of field type object and I plan to revisit it later. Right now it is just easier to keep flags for all field types in one place in explicit form even if they are somewhat redundant.
There was a problem hiding this comment.
Makes sense, this jsonization is a pretty massive undertaking already and I'm totally on board with keeping the code simpler for now.
| @@ -413,6 +413,8 @@ bool map::process_fields_in_submap( submap *const current_submap, | |||
| debugmsg( "Whoooooa intensity of %d", cur.get_field_intensity() ); | |||
| } | |||
|
|
|||
| dirty_transparency_cache = curtype.obj().dirty_transparency_cache; | |||
There was a problem hiding this comment.
The logic here is wrong, it needs to "latch" to dirty. Right now it will be un-set when it processes an invisible field.
| dirty_transparency_cache = curtype.obj().dirty_transparency_cache; | |
| dirty_transparency_cache |= curtype.obj().dirty_transparency_cache; |
I think this will do it, however when I test the smoke from a smoke grenade still isn't acting non-transparent.
There was a problem hiding this comment.
Re-tested with current master and it's already broken, so that's great :/
There was a problem hiding this comment.
Maybe be I broke it earlier? Need to test in 0.D...
There was a problem hiding this comment.
0.D is working, so I need to bisect.
There was a problem hiding this comment.
Smoke is non-transparent starting from intensity level 2.
Maybe it is not transparency, but field intensity issue (i.e. smoke from smoke bombs aren't achieving level 2 or 3 anymore)?
I've tested on savegame from 0.D with exploded smoke bomb and there is thick smoke in the center which is non-transparent, but newly exploded smoke bomb only gives intensity level 1.
There was a problem hiding this comment.
Should be fixed in #32551.
Summary
SUMMARY: Infrastructure "Refactor fields (step 6)"Purpose of change
Unhardcode data structures to allow field type customization/addition via json.
Describe the solution
Current pull request does the following: