* When rupees are given as starting items, they still modify `gSaveContext.rupeeAccumulator`, which means that if you make a new file, start it, and reload without saving, or exit SoH before starting the new file, those rupees were lost. This adds a check for `gPlayState` being NULL, and if it is, adds those initial rupees to the rupee count directly.
* Update soh/src/code/z_parameter.c
Co-authored-by: briaguya <70942617+briaguya-ai@users.noreply.github.com>
---------
Co-authored-by: briaguya <70942617+briaguya-ai@users.noreply.github.com>
* Modified Affordable logic to select randomly from 10, 105, 205, and 505 depending on the selected wallet tier. Updated the tooltip to reflect.
* Clarified logic in comments and variable names.
* Streamlined affordable check for starter wallet.
One more function comment.
* More streamlining of affordable price generation, comment clarification.
* Changed OnReceiveItem hook to pass GetItemEntry, which required the following changes:
Reworked the references to it in `z_parameter` to call a single function with the item return and hook call in it for ease of editing.
Modified the pendingSale functionality to set and pass a modIndex value as well, as mod items (like randomizer) in shops still use the vanilla sale path.
* Missed some files for the pendingSale changes.
Also added Randomizer_Item_Give to the OnReceiveItem system.
* Ice traps now trigger OnReceiveItem.
* All ice traps truly do work now.
As a side effect, item autosave doesn't work for shop/scrub/merchant transactions, requires new OnSaleEnded hook that can also call the autosave.
* Removed unnecessary calls to ItemTable_RetrieveEntry where GetItemEntry properties were being used to call it. General code cleanup.
* Added OnSaleEnd hook for when rupees are finished deducting after a sale.
Migrated AutoSave to its own function, registered AutoSave function to OnReceiveItem and OnSaleEnd hooks to help with autsaving after buying items.
Some futureproofing for AutoSave function with parameters for skipping autosave, for when transition end is migrated to AutoSave function (whether through direct call or through a hook).
* Renamed hook paramaters, and registered hook function parameters, to a more descriptive alternative.
* Missed a couple, fixed a typo.
* One more missed paramater name refactor.
Refactored all references to OnReceiveItem to OnItemReceive to mirror upcoming full hook refactor for name ordering conventions.
Up-to-date with develop.
This leaves the default of 10.15, but allows a user to change it in the
cache. When using Homebrew libraries on newer macOS version, the linker
emits warnings:
```
ld: warning: dylib (/opt/homebrew/lib/libSDL2.dylib) was built
for newer macOS version (13.0) than being linked (11.0)
```
I use PyEnv and rebuilding after recent commits was giving me a cmake error about not being able to find the Python3 executable. Advising other users who might run into this problem on how to fix it.
* save build version to savefile
* adjust rando hash icons to use fade in/out
* add dialog message support on the file select screen and display rando version warning
* remove duplicated message functions and use stubbed play state instead for rando warning
* add major/minor/patch version saving to file and compare against
* use strncpy and memset for build version
* don't show rando warning one copy/erase screens
* review feedback
* Add german and french translations for rando warning
Co-authored-by: PurpleHato <linkvssangoku.jr@gmail.com>
---------
Co-authored-by: PurpleHato <linkvssangoku.jr@gmail.com>
* Restored vanilla pause buffering and input behavior broken by 'easy frame advance' cheat
Currently, holding A and repeatedly pausing and unpausing will cause that A input to be repeated, resulting in Link rolling without additional A inputs. Similarly, holding Z and unpausing in front of a candidate actor will cause Link to switch Z-targeting to that actor.
As a side effect, this also does not allow inputs to be buffered before the Subscreen has finished its closing animation, which is how vanilla also works. This is slightly unfortunate because the "freeze" before the game resumes is considerably shorter in Shipwright, but it'd be better to implement a fake delay there if we want to make it more forgiving.
* Fixed writing past end of the buffer when loading personal item tracker notes longer than 8 characters
* Revert "Fixed writing past end of the buffer when loading personal item tracker notes longer than 8 characters"
This reverts commit 721c18d488.
