* Clamp hpPer as a sanity check
* Refactor update mana routines in control.cpp to use PlayerStruct::UpdateManaPercentage
* Remove unnecessary calls to UpdateHP/ManaPercentage
* Unify logic that draws the flasks into a single internal function
* Fix style violations
* Remove unnecessary save/loads of hp/manaPer
The way this variable is used expects a value in this range and it gets clamped again anyway, but this makes the code more similar to the manaPer logic for future refactoring
Identical logic in use, the checks to see if the values are negative can be accomplished using clamp and the same divide by zero guard. I'll combine the two attributes in a near future commit.
control_update_life_mana is called during the game loop before any of the drawing functions so hpPer/manaPer are calculated based on the latest current/max of their respective attributes every frame.
There's no need to do the calculation again in the draw routines. (There's also no need to calculate this every frame but that's going to take a bit to untangle.)
Added overload for DrawString taking a Point to avoid creating a rect for callers which only use position. This also documents the way DrawString operates when passed a clipping rectangle with a dimension of 0.
As part of this overload removed the logic for 0 width regions from DrawString. This does change the behaviour of the Rectangle version if called with a rect with width 0, all callers using that behaviour have been updated in this commit.
Using Rectangle/Size allowed simplifying the logic for certain calls where they could use DrawText alignment flags, previously this was manually aligning by calculating dimensions and offsetting the position. This also fixes#2169
Also includes a few instances where a temporary buffer was used to set the text to be drawn with unbounded sprintf calls, replaced those with snprintf as is recommended in modern C applications. Moving to C++ strings would be good in a future refactor.
Disabled by default because of these known issues:
1. When clicking on inventory item, it briefly appears a bit shifted (in the wrong coordinates).
This issue can happen with software cursor as well, but is a lot more
obvious with the hardware cursor.
2. Cursor is scaled with nearest-neighbour scaling, which may look a bit different from
how the rest of the graphics are scaled.
See also previous attempt: https://github.com/diasurgical/devilutionX/pull/955 by @viciious
Co-authored-by: Victor Luchits <vluchits@gmail.com>
This mostly change player to be by reference instead of by index.
But additionally it does stript checks for gold in the belt, move some
value types to the initialization and short circute a few functiongs.
Now that we no longer have a buffer border, we can render directly to
the video memory if the output surface is 8-bit and double buffering is
enabled.
Also adds a flag to force this even when double buffering is disabled,
because some systems emulate 8-bit mode with something else, so the
output is always buffered.
This required a tweak to the cursor to make sure that we clear it after
`SDL_Flip`.
obligaron
RISHABH BUDHIRAJA
Anders Jenbo
Andrew James
Gleb Mazovetskiy
Juliano Leal Goncalves
qndel
Kalebe Alves