- Check for if a single local player isn't spectator or no-contested.
- Clean up code to return early if exclusionary conditions are met, rather than one long if conditional bundling them together
Instead of the absolute insanity that is V_DrawCroppedPatch, which makes you specify the number of pixels to crop off the top/left and then the number of pixels to show after the crop ... you just use V_SetClipRect to create a rectangle for any future draws to be cropped down to, and V_ClearClipRect afterwards to clear it.
Currently only supported by V_DrawStretchyFixedPatch. Ideally other drawing functions should also receive clipping rectangle support too.
Rather than setting udmf earlier in P_SetupLevel and continuing to awkwardly
check udmf everywhere, I'm adding a new global solely for compat purposes.
For now, have it mirror udmf to avoid supporting two binary formats.
(still keeping the 2.2 conversion code tho?)
- Drawn underneath absolutely everything else, because it's the least specific of all the guides minigen can provide
- "Common objects" includes:
- Rings/spheres
- Waypoints
- Item boxes/spots
- Overtime kiosk
- Rings
- Item Capsules
- The above were chosen because they're a good distinctor between sectors that are valid to drive on and sectors that would be valid were there no impassable lines or massive height differences preventing the player from getting there.
Related:
- K_TerrainHasAffect now has a "bad only" check mode.
- If true only report back for strong friction, any offroad, any damage, or stairjank.
The totally-not-a-secret reason I made this branch.
- doomednum 770 (associated with polyobject anchors 760/761 and skybox centerpoint 780)
- Place exactly two in a map to draw an implicit rectangle.
- Supports top-left/bottom-right AND bottom-left/top-right placements.
- I_Errors if you place too many (or only one).
- You don't *have* to have these, this is just a bonus if you're a map like Power Plant or CDSS1 negatively affected by your skybox.
- `P_InitMinimapInfo`
- Writes to `p_local.h` extern struct
- Handles everything previously distributed across multiple K_drawKartMinimapIcon calls (and most of AM_findMinMaxBoundaries)
- Reduces complexity of drawKartMinimapIcon significantly
- Last prerequisite before implementing user-placable minimap boundaries...
Supports values between 1 and 10, with 1 being the default.
If a multiplier greater than 1 is provided, the filename will take the form "MINIMAP-%d.png", where %d is the multiplier.
Return a `minigen_t` struct with explicit width and height instead of extending the UINT8 buffer by 2 to provide that information in a very datatype-limited way.
- All pre-existing colours are now actually what the code says they should be for the 2.2 palette
- RR-specific colour changes.
- Remove all noclimb-specific line colours
- Render areas that can't be stepped up/down onto as walls
- Add colours for
- Tripwire (cyan, but not 0xff cyan)
- Finish Line (grey)
- FOF info (blue)
- Use a low-intensity colour to signal possible offroad/hazard sector or stairjank step
- Rework to support drawing in multiple passes, so that information that is strictly more important (solid walls, finish line) will not be obscured at minimap resolution by nearby lines signalling offroad/stairjank
Related:
- Added K_TerrainHasAffect
- Returns true if terrain has any properties which would affect the player's gameplay, false if not.
- Generates "MINIMAP.png" in your srb2home
- Uses inherited automap code to render to a temporary buffer
- Because am_map.c is a mess of filescope static variables right now, this only works when the automap is disabled.
- Currently an equal alternate method to SLADE's map image export, but because we're in control, additional features can be added later...
- TODO: Off vertically by one pixel on GHZ. Otherwise effectively identical in shape
- TODO: the colours are rancid, I wonder if they were even updated for the 2.2 palette
Related:
- Use identical linear-time mechanisms for detecting borders of map geometry between automap and minimap
- Automap was previously using iteration over all vertices
- Minimap was previously pointlessly writing min/max values twice
This used to be used for `P_RelinkPointers()`, but was superseded by the far saner mobjnum. HOWEVER, its lingering effect was to leave `mobj->info` in an invalid state until a later `P_FinishMobjs()` was called. This is memory unsafety :D
After removing the last remnants of this ancient hack, it is now once again possible to connect to a server without crashing immediately. However, I did get a crash after a few seconds, so there is definitely still *something* nasty going on under the hood.
Foundational assistive work.
Also guarantees a consistent cup name length in memory, as some places read/wrote 15 bytes, and some places read/wrote 16 bytes. We settle on the latter (including null terminator) for the broadest backwards compatibility.
G: partial merge, missing lots of cup stuff
- Defaults to 1, AKA Sonic 1/2 monitor
- Set to 2 for Sonic 3k monitor
- Supports a range of 1-9 and A-Z
- Permits fine-grained rapid-prototype stretch goal Chaotix monitor stuff later in development
R_AllocTextureColumnTables wasn't checking if the tables were never allocated, making the renderer later attempt to render midtextures for drawsegs that don't contain actual midtextures.
* Instead of doing constant G_MapNumbers when finding the relationship between maps and cups...
* Add a cache of level IDs to cups, to go with the strings.
* Add a cache of the cup pointer to maps, so we don't have to search through all cups to find our map. (done in P_InitMapData)
* Pre-emptive work: G_IsSpecialStage and P_GetNextEmerald now reference cup data instead of a hardcoded ID set.
* Remove a bunch of old stuff from mapheaderinfo_t/associated, and reorder what stays
* Countdowntimer? 💥
* Startrings? 💥
* sstimer/ssspheres? 💥
* forcecharacter? 💥 (distinct from forceskin)
* interscreen? 💥
* sstage_start/end and smpstage_start/end? 💥💥💥💥
* You've been blocked
* G_MapNumber now returns a special NEXTMAP_INVALID if not found, for more consistent reference.
* Incorporate a good chunk of the `edit-headers` branch. Can't clear maps individually because of the new restrictions on sequential mapheaders, but we can add a "disable in vote screen, not even for map hell/archive" flag to a map at some future juncture for equivalent functionality...
- mapvisited and recorddata_t (previously mainrecords)
- Changed how gamedata is saved and loaded
- Change the versioncheck (funny hex provided by chengi) AND call it `developringdata.dat` in develop builds
- Fix a bunch of off-by-ones in condition and emblem data
ALSO, for Time Attack:
- Fix menu not showing off your times
- Now save times even when gamedata modified, since the menu didn't care (come back to it?)
- Don't save times or do intermission screen if the Capsule Attack ended because you lost all your bumpers
- Fix a crash adding ghosts in Capsule Attack
- Refactor significantly (now has its own func, `G_GetNextMap`)
- If gametype uses cups, iterate through cups to find the current level, then grab the next valid level
- If not, get the next valid mapheader for your gametype
- SOC `nextmap`/`marathonnext` is not just deprecated but REMOVED
- Hide the NEXTMAP_ constants again, but leave support dummied out for if we have them publically accessible again
- Also get rid of a bunch of OTHER mapheader stuff we're never gonna use
- NiGHTS Grades? NOPE
- Vanilla titlecard patches? NOPE
- Boss music fadeout/replacement? NOPE
- Select Heading? NOPE
- You've been blocked.
- Don't show maps without lumps on the level select list
- this is me being petty, but making it NOTIMEATTACK in SOC instead of TIMEATTACK so we can reconsider the maps with/without them.
- G_MapNumber now handles special NEXTMAP_ nextmapspecial_t constants that exist at the end of the available type.
- Cleanup of G_DoCompleted
- Add bounds checking to the various SOC maincfg map starts (spstage_start, etc)
- Add lump checking to titlemap behaviour
- Now a struct with zone-allocated buffer member
- Instead of bunging it on the end of the buffer, the time for G_SometimesGetDifferentGametype is its own struct member
Also:
- improved error prints for SOC condition definitions
- improved bounds checking to use `nummapheaders` for iterating over mapheaderinfo
There are still situations that use NUMMAPS like mapvisited, randmapbuffer, etc, which need to be addressed before merger.
* Required an adjustment of everywhere using G_MapNumber to return the raw header number, instead of off-by-one ala gamemap.
* Fixing gamemap is a viable improvement for a future commit, but this commit is already pretty big.
* Remove SCANTHINGS, since it used G_MapNumber and didn't work with long map names OR virtres anyways.
* Support freeing new information in CLEAR LEVELS maincfg event, since I tried to use that to test cups.
* Make Patch_Free's usability match Z_Free -- passing NULL is permitted and a no-op.
