Loadless Timer and Death Counter #37

eddynya · 2023-11-11T03:36:40Z

Death Counter

Keeps track of your deaths in story mode

Loadless Timers

Includes both fullgame and segment timers, and each can be toggled to not show at all, show at all times, show between worlds, and only show at the end of the run
If the segment timer is enabled in any way, at the end of the run, you can see all your split times + segment times for each world
Toggle-able warning that shows up on the name entry screen if no timers are enabled

Unified draw_timer() Code

The challenge mode segment timer, IW timer, RTA/Pause timer, and story timer all use the updated function

Known Issues

everything is explained here https://docs.google.com/document/d/1b2wigdLeCfmix8JAe3KXI8Z4aBpLXtcXYd08oKKzuYw/edit?usp=sharing

-timer no longer increments on mkb::STAGE_SELECTED due to (load?) variance -timer no longer increments between the stage select a press and the start of the 10 ball screen spin in bc of variance between emu and console (will test more though to be sure)

-the timer now increments for exactly 49 frames during fade out (which is the usual fade out time when stage selecting) after pressing stage select. this fade out time should not vary between sd/usb/dolphin -fixed the death counter bug where it would increment to 1 on the first stage automatically

ComplexPlane

This should be plenty to get you started!

ComplexPlane · 2023-11-11T13:35:44Z

scripts/gen_compile_commands.py

You can undo these permissions changes with git checkout scripts inside WSL. I think this is caused by having the repo cloned outside WSL, or accessing the repo inside WSL with a tool outside WSL... I forget what.

Point is, IMO the best experience is to git clone the repo from a WSL terminal inside your WSL home directory, and open it in vscode using code . inside the folder in a WSL terminal. I don't know what your setup looks like but nambo/rehtrop should definitely be able to sort you out if there's an issue.

Setting up my environment on my laptop and I ran into this too. I believe when vscode isn't connected to WSL, it attempts to Windows-ify the files? I'm also not sure if code . sets up the remote connection if you don't already have the WSL extension installed.

@eddynya If you don't have the WSL extension for vscode, you'll want to get it before trying Complex's solution. Once you have it, do ctrl+shift+p and type "connect to wsl" and select that option. Then try running git checkout scripts and the files should be reverted.

If this doesn't work and you want a simpler solution you can try git config core.filemode false. This will stop git from picking up these file changes.

I'm not sure about git config core.filemode false because we do want to respect executable bit of the script files and future scripts etc. You shouldn't run into this issue if you clone the repo inside your WSL home dir, and open code with code . (or Connect to WSL).

src/mods/storytimer.cppZone.Identifier

src/mods/storytimer.cpp

ComplexPlane · 2023-11-11T13:55:12Z

src/mods/deathcounter.cpp

+    // first framing should not increase the death counter, and retrying after breaking the tape should not increase it either
+    // to do: however, if you retry after breaking the tape on the very first frame (so the frame before goal init), it does count as a death when it should not
+    if (mkb::sub_mode == mkb::SMD_GAME_GOAL_INIT || mkb::g_storymode_stageselect_state == mkb::STAGE_SELECT_INTRO_SEQUENCE) {
+        s_can_die = false;
+    }


Why should retrying on first frame not count as a death? Wouldn't it be simpler to keep it consistent with the game's behavior?

src/mods/storytimer.cpp

ComplexPlane · 2023-11-11T15:49:22Z

src/mods/storytimer.cpp

+    if (mkb::g_storymode_stageselect_state == mkb::STAGE_SELECT_INTRO_SEQUENCE || mkb::g_storymode_stageselect_state == 3 || 
+        mkb::g_storymode_stageselect_state == mkb::STAGE_SELECT_IDLE || mkb::g_storymode_stageselect_state == 5 || mkb::g_storymode_stageselect_state == mkb::STAGE_SELECTED) {


There's a StoryModeStageSelectState enum you can make use of, but it looks like it's missing 5

src/mods/storytimer.cpp

ComplexPlane · 2023-11-11T16:36:33Z

src/utils/timerdisp.cpp

Ok, at this point I actually think timerdisp is trying to be way too many things for its own good. I did suggest generalizing draw_timer(), but I didn't imagine you'd want to do stuff like optionally draw two timers on the same line in a specific format.

Here's the relevant concerns as I understand them:

Multiple modules want to draw timers left-aligned at certain parts of the screen, but they should "stack" to avoid overlapping

Modules need to format their timers differently (prefixes, more than one timer, etc.), but without duplicating excessive formatting code

Modules need to negotiate where their lines of text are rendered, but aside from that, they don't need to agree on how their timers are drawn.

So here's my proposal:

Make a separate module that is dedicated to dishing out "slots" ((x, y) coordinate pairs) for other modules to render arbitrary lines of text to. Modules that call this function earlier get "slots" higher on the screen. The function might look like:

enum class Region { TopLeft, TopRight }; Vec2d get_next_slot(Region pos);

Although, since your code is the only thing using the top left region, you might not need to implement this at all / just keep the current logic for how stuff in the top-right region is drawn.

Let individual modules render their own arbitrary line of text. Modules could use a shared function to format a single frame counter variable as a time as a helper, and e.g. call it multiple times if they need to render more than one timer on the same line.

Btw, this is how you build a formatted string by itself, so you can use it later:

char str[32] = {}; // Must be big enough to hold the longest possible string, plus one null character mkb::sprintf(buf, "int: %02d, float: %.02f, string: %s", 10, 1.5, "substring");

Using a shared timer formatting function would look something like:

char timer_str[32] = {}; timer_utils::format_timer(timer_str, my_framecount /* More options here */); draw::debug_text("SEG: %s", timer_str);

Note that you can't returna pointer to the buffer you sprintf into because it is a stack variable, so you have to pass it "down" instead of "up".

All the timer code here looks like it's gotten overly complex, so this sounds like a great idea! @eddynya I can help you get started with this, just dm me if you have questions!

ComplexPlane · 2023-11-11T16:45:29Z

src/mods/storytimer.cpp

+        if (s_is_on_spin_in == true) {
+            // increment the timer every frame during spin in
+            s_spin_in_timer +=1;
+        }
+        if (s_is_on_gameplay == true) {
+            //increment the timer every frame during gameplay
+            s_gameplay_timer +=1;
+        }
+        if (s_is_postgoal == true && s_stage_fade_out_timer <= stage_fade_out_time) {
+            // increment the timer every frame after game goal init happens; once you press stage select, a separate 49 frame timer is started (fade out from stage select
+            // to the first completely white frame takes 49 frames). once the timer hits 49 frames, stop incrementing the timer 
+            // until the 10 ball screen starts spinning in
+            s_postgoal_timer +=1;
+        }
+        if (mkb::g_storymode_stageselect_state == mkb::STAGE_SELECT_INTRO_SEQUENCE || mkb::g_storymode_stageselect_state == 3 || 
+        mkb::g_storymode_stageselect_state == mkb::STAGE_SELECT_IDLE) {
+            //increment the timer every frame on the story mode select screen until the a press input; we do not include the transition time after pressing a afterwards 
+            // even ignoring completely white frames, the time spent on mkb::g_storymode_stageselect_state == mkb::STAGE_SELECTED can be
+            // highly variable (up to over 40 frames sometimes!), so for the purpose of a loadless timer, it makes sense to cut this out from the timer
+            s_stage_select_timer +=1;
+        }
+        if (s_is_on_exit_game_screen == true) {
+            // increment the timer every frame on the exit game screen
+            s_exit_game_timer +=1;
+        }
+        if (s_is_on_fallout_screen == true) {
+            // increment the timer every frame during the fallout sequence and y/n screen
+            s_fallout_timer += 1;
+        }
+        if (s_is_timeover == true) {
+            // increment the timer every frame during the timeover sequence
+            s_timeover_timer += 1;
+        }


You probably want to use else for a lot of these...

src/mods/storytimer.cpp

ComplexPlane · 2023-11-11T17:17:09Z

Oh just noticed the conflicts... yeah you will need to pull and rebase/merge master to reconcile w/ rehtrop's latest changes

-now using 1 enum class for both timers instead of 2

ComplexPlane · 2023-11-12T18:22:59Z

src/mods/storytimer.cpp

+static bool s_in_story;
+static bool s_is_on_world[11];
+static bool s_is_between_worlds;
+static bool s_is_run_complete;
+static bool s_is_on_spin_in;
+static bool s_is_on_gameplay;
+static bool s_is_postgoal;
+static bool s_is_on_stage_select_screen;
+static bool s_is_on_exit_game_screen;
+static bool s_is_on_fallout_screen;
+static bool s_is_timeover;
+static bool s_can_increment_stage_counter;
+static bool s_lower_stage_counter;
+static bool s_start_STAGE_FADE_OUT_TIMEr;


Vast majority of these don't need to be static variables (or perhaps not even variables at all)

I made most of these local variables, if you think most don't need to be variables at all I can do that. My reasoning for keeping them as variables is that (I hope) writing it this way makes it easier to read than having like 20 submodes in an if()

-rewrote completed stages code (will probably try and simplify still) -small changes to death counter

Resolving merge conflicts with rehtrop's changes -Changed the RTA/Pause timer to use the new function I made as a bandaid fix until I address Complex's comment -Changed one of the timer widgets to say "Segment & Loadless Timers" until I decide how I want to reorganize this part of the menu

Pulled updated ghidra headers

(not finished)

removed some static u32 variables that were being kept track of

rehtropsmb

Some more stuff to take a look at. I still haven't tested in-game, this was just from doing an overview of the code. I'll do some gameplay testing eventually :)

Also make sure to remove the script changes from your PR! And pull in the most up-to-date code so you can fix your merge conflicts!

src/systems/pref.cpp

src/systems/menu_defn.cpp

src/mods/deathcounter.cpp

rehtropsmb · 2023-11-20T08:19:21Z

src/mods/deathcounter.cpp

+    if (s_can_die == true && (mkb::sub_mode == mkb::SMD_GAME_READY_INIT || mkb::sub_mode == mkb::SMD_GAME_RINGOUT_INIT 
+    || mkb::sub_mode == mkb::SMD_GAME_TIMEOVER_INIT || mkb::sub_mode == mkb:: SMD_GAME_SCENARIO_RETURN)){
+        // you can die either by retrying after dropping in, falling out, timing over, or stage selecting after dropping in (but before breaking the tape)
+        s_death_count += 1;


s_can_die would be set to false so it seems like this would only be able to be ran once anyways

src/mods/deathcounter.cpp

src/mods/storytimer.cpp

rehtropsmb · 2023-11-20T21:15:47Z

src/mods/storytimer.cpp

+    // if the segment timer is enabled in any capacity, show all 10 split times + iw times after the
+    // tape is broken on the last stage
+    if (TimerOptions(pref::get(pref::U8Pref::SegmentTimerOptions)) != TimerOptions::DontShow &&
+        is_run_complete) {
+        // haven't attempted the mkb::sprintf suggestion yet
+        timerdisp::draw_timer(IW_TIME_LOCATION_X, segment_timer_location_y, IW_TIME_TEXT_OFFSET,
+                              "W1:", s_timer_group[0].split, s_timer_group[0].segment, true, false,
+                              draw::WHITE);
+        timerdisp::draw_timer(IW_TIME_LOCATION_X, segment_timer_location_y + 1, IW_TIME_TEXT_OFFSET,
+                              "W2:", s_timer_group[1].split, s_timer_group[1].segment, true, false,
+                              draw::WHITE);
+        timerdisp::draw_timer(IW_TIME_LOCATION_X, segment_timer_location_y + 2, IW_TIME_TEXT_OFFSET,
+                              "W3:", s_timer_group[2].split, s_timer_group[2].segment, true, false,
+                              draw::WHITE);
+        timerdisp::draw_timer(IW_TIME_LOCATION_X, segment_timer_location_y + 3, IW_TIME_TEXT_OFFSET,
+                              "W4:", s_timer_group[3].split, s_timer_group[3].segment, true, false,
+                              draw::WHITE);
+        timerdisp::draw_timer(IW_TIME_LOCATION_X, segment_timer_location_y + 4, IW_TIME_TEXT_OFFSET,
+                              "W5:", s_timer_group[4].split, s_timer_group[4].segment, true, false,
+                              draw::WHITE);
+        timerdisp::draw_timer(IW_TIME_LOCATION_X, segment_timer_location_y + 5, IW_TIME_TEXT_OFFSET,
+                              "W6:", s_timer_group[5].split, s_timer_group[5].segment, true, false,
+                              draw::WHITE);
+        timerdisp::draw_timer(IW_TIME_LOCATION_X, segment_timer_location_y + 6, IW_TIME_TEXT_OFFSET,
+                              "W7:", s_timer_group[6].split, s_timer_group[6].segment, true, false,
+                              draw::WHITE);
+        timerdisp::draw_timer(IW_TIME_LOCATION_X, segment_timer_location_y + 7, IW_TIME_TEXT_OFFSET,
+                              "W8:", s_timer_group[7].split, s_timer_group[7].segment, true, false,
+                              draw::WHITE);
+        timerdisp::draw_timer(IW_TIME_LOCATION_X, segment_timer_location_y + 8, IW_TIME_TEXT_OFFSET,
+                              "W9:", s_timer_group[8].split, s_timer_group[8].segment, true, false,
+                              draw::WHITE);
+        // use segment timer spacing for w10 since "W10" is 3 characters long, not 2
+        timerdisp::draw_timer(SEGMENT_TIMER_LOCATION_X, segment_timer_location_y + 9,
+                              SEGMENT_TIMER_TEXT_OFFSET, "W10:", s_timer_group[9].split,
+                              s_timer_group[9].segment, true, false, draw::WHITE);
+    }


Should this be it's own bool preference? Or just the EndOfRun case?

The way it works now is that if the segment timer is on at all (so either showing at all times, between worlds, or at the end of the run), when the run is finished, all 10 split and segment times are shown. I don't think this should get its own bool pref since this information is shown after the run, and so people who want to run pace blind aren't affected by it

retrying on the tape break frame no longer increases the death counter

nothing significant really

-timer and seg timer are consistently offset from each other currently when they shouldn't be

also fixed the update order bug causing the segment timer to be 1f behind what it should be

issue: the w1 split does not show up

the segment timer now stops on the goal tape break frame instead of using SMD_GOAL_INIT and subtracting 2 frames manually

-removed unused variables -fixed error where split 1 was off by 1f

eddynya added 13 commits October 19, 2023 02:59

loadless timer + segment timer stuff

b944f0b

fixing segment timer after resetting

03b8941

somewhat fixed some issues with the timer missing frames

60848bd

fixed the missing frames issue (kinda jank fix) for the loadless timer

835c8d8

fixed segments not showing up when run ends

bd51ae1

added a story death counter

f9cfe63

fixed timer/death counter spacing + timer not on warning

c78143f

some code cleanup

d7ceef5

unified timer code, fixed spacing on old timers

94d1fe8

make the loadless timer count during timeover

261494c

cut out parts of the timer

688050f

-timer no longer increments on mkb::STAGE_SELECTED due to (load?) variance -timer no longer increments between the stage select a press and the start of the 10 ball screen spin in bc of variance between emu and console (will test more though to be sure)

some cleanup

e7f7c67

ComplexPlane requested changes Nov 11, 2023

View reviewed changes

ComplexPlane reviewed Nov 11, 2023

View reviewed changes

src/mods/storytimer.cpp Outdated Show resolved Hide resolved

eddynya added 2 commits November 11, 2023 18:30

small fixes before tackling bigger issues

e8623d9

-now using 1 enum class for both timers instead of 2

Delete storytimer.cppZone.Identifier

3286dfa

ComplexPlane reviewed Nov 12, 2023

View reviewed changes

ComplexPlane requested a review from rehtropsmb November 13, 2023 00:28

eddynya and others added 9 commits November 13, 2023 16:51

small changes

475cffe

-rewrote completed stages code (will probably try and simplify still) -small changes to death counter

Merge branch 'master' of github.com:ComplexPlane/SMB2PracticeMod

80636e4

Pulled updated ghidra headers

updated names to match new ghidra header

2f1d3c7

starting to rewrite segment timer code

4a69bcd

(not finished)

another approach

19d253c

fixed some issues with the rewrite

eca1c36

pre cleanup backup

b5eeeff

clean up arc 2 part 1

cb66e06

eddynya added 2 commits November 17, 2023 18:47

added count deaths on stage 1 toggle to deathcounter

9a38703

optimized memory used by timer

290410a

removed some static u32 variables that were being kept track of

rehtropsmb requested changes Nov 20, 2023

View reviewed changes

eddynya and others added 24 commits November 21, 2023 17:42

pre merge commit

066e8d0

Update pref.cpp

00df2a7

Merge branch 'master' of github.com:ComplexPlane/SMB2PracticeMod

01d4c8a

fixed deathcounter bug

e2f6eaa

retrying on the tape break frame no longer increases the death counter

started addressing storytimer comments

4cda08a

nothing significant really

small changes before bigger changes

9170ba3

backup before fixing bugs

a788754

-timer and seg timer are consistently offset from each other currently when they shouldn't be

removed unnecessary static variables in the TimerGroup struct

47d96c1

also fixed the update order bug causing the segment timer to be 1f behind what it should be

simplified code for the number of completed stages

e192874

removed the s_completed_stages_world array

bf4555a

Delete storytimer.cppZone.Identifier

91e4558

removed unused variables

7b0cb90

rewrote end of run segment timer code using mkb::sprintf

fd65cb3

fixed some end of run segment formatting

6493439

issue: the w1 split does not show up

fixed the w1 not showing issue

3f13d8b

improved code for stopping the segment timer

4e4ab77

the segment timer now stops on the goal tape break frame instead of using SMD_GOAL_INIT and subtracting 2 frames manually

clean up

edb2f4c

reverted draw_timer function to only draw one timer

7648a30

debugging stuff

058e037

removed debugging variables/cleanup

1f4a237

Merge branch 'ComplexPlane:master' into master

f694f0e

update scripts

9968535

minor things

941f805

-removed unused variables -fixed error where split 1 was off by 1f

Merge branch 'master' of https://github.com/eddynya/SMB2PracticeMod

67b8c3d

		if (mkb::g_storymode_stageselect_state == mkb::STAGE_SELECT_INTRO_SEQUENCE \|\| mkb::g_storymode_stageselect_state == 3 \|\|
		mkb::g_storymode_stageselect_state == mkb::STAGE_SELECT_IDLE \|\| mkb::g_storymode_stageselect_state == 5 \|\| mkb::g_storymode_stageselect_state == mkb::STAGE_SELECTED) {

Loadless Timer and Death Counter #37

Are you sure you want to change the base?

Loadless Timer and Death Counter #37

Uh oh!

Conversation

eddynya commented Nov 11, 2023

Uh oh!

ComplexPlane left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rehtropsmb Nov 13, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

ComplexPlane commented Nov 11, 2023

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rehtropsmb left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

rehtropsmb Nov 13, 2023 •

edited

Loading

rehtropsmb left a comment •

edited

Loading