From: Rudolf Polzer <divVerent@gmail.com>
Date: Fri, 13 Mar 2020 23:49:52 +0000 (-0400)
Subject: Expand the division by sensitivity into the math itself.
X-Git-Url: https://git.rm.cloudns.org/?a=commitdiff_plain;h=7b7193287a61cbe0d2eea65c6cb3f87921d21522;p=xonotic%2Fdarkplaces.git

Expand the division by sensitivity into the math itself.

Also, add some TODO comments for things to still figure out maybe.
Maybe it's better though to stick to QL's definition of the values.
---

diff --git a/cl_input.c b/cl_input.c
index bed0be60..b3bf7b38 100644
--- a/cl_input.c
+++ b/cl_input.c
@@ -592,18 +592,21 @@ void CL_Input (void)
 			// sensitivity, so we apply it but then dividie by
 			// sensitivity.value so that the later multiplication
 			// restores it again.
-			float accelsens = sensitivity.value;
+			float accelsens = 1.0f;
 			float adjusted_speed_pxms = (speed * 0.001f - m_accelerate_power_offset.value) * m_accelerate_power_strength.value;
+			float inv_sensitivity = 1.0f / sensitivity.value;
 			if (adjusted_speed_pxms > 0)
 			{
 				if (m_accelerate_power.value > 1.0f)
 				{
-					accelsens += expf((m_accelerate_power.value - 1.0f) * logf(adjusted_speed_pxms));
+					// TODO: How does this interact with sensitivity changes? Is this intended?
+					// Currently: more sensitivity = less acceleration at same pixel speed.
+					accelsens += expf((m_accelerate_power.value - 1.0f) * logf(adjusted_speed_pxms)) * inv_sensitivity;
 				}
 				else
 				{
 					// The limit of the then-branch for m_accel_power -> 1.
-					accelsens += 1.0f;
+					accelsens += inv_sensitivity;
 					// Note: QL had just accelsens = 1.0f.
 					// This is mathematically wrong though.
 				}
@@ -613,12 +616,12 @@ void CL_Input (void)
 				// The limit of the then-branch for adjusted_speed -> 0.
 				// accelsens += 0.0f;
 			}
-			if (m_accelerate_power_senscap.value > 0.0f && accelsens > m_accelerate_power_senscap.value)
+			if (m_accelerate_power_senscap.value > 0.0f && accelsens > m_accelerate_power_senscap.value * inv_sensitivity)
 			{
-			        accelsens = m_accelerate_power_senscap.value;
+				// TODO: How does this interact with sensitivity changes? Is this intended?
+				// Currently: senscap is in absolute sensitivity units, so if senscap < sensitivity, it overrides.
+			        accelsens = m_accelerate_power_senscap.value * inv_sensitivity;
 			}
-			// Divide by the standard linear sensitivity again.
-			accelsens /= sensitivity.value;
 
 			in_mouse_x *= accelsens;
 			in_mouse_y *= accelsens;