From 020c60feec8e3e2f199f92de0a8e75661c822761 Mon Sep 17 00:00:00 2001
From: helixhorned <helixhorned@1a8010ca-5511-0410-912e-c29ae57300e0>
Date: Sun, 18 Mar 2012 23:17:17 +0000
Subject: [PATCH] Port tvlineasm2 to C and enable transmaskvline2 code also for
 all-C builds.

With the same setup as before, a screen-filling translucent wall (with nothing
drawn behind it) renders at about 7 fps faster (from 60-something fps initially)

git-svn-id: https://svn.eduke32.com/eduke32@2498 1a8010ca-5511-0410-912e-c29ae57300e0
---
 polymer/eduke32/build/include/a.h  |  3 ++
 polymer/eduke32/build/src/a-c.c    | 66 +++++++++++++++++++++++++++++-
 polymer/eduke32/build/src/engine.c |  6 +--
 3 files changed, 71 insertions(+), 4 deletions(-)

diff --git a/polymer/eduke32/build/include/a.h b/polymer/eduke32/build/include/a.h
index 9f7411c59..a59de21d6 100644
--- a/polymer/eduke32/build/include/a.h
+++ b/polymer/eduke32/build/include/a.h
@@ -135,6 +135,9 @@ void mvlineasm4(int32_t cnt, char *p);
 void setuptvlineasm(int32_t neglogy);
 int32_t tvlineasm1(int32_t vinc, intptr_t paloffs, int32_t cnt, uint32_t vplc, intptr_t bufplc, intptr_t p);
 
+void setuptvlineasm2(int32_t neglogy, intptr_t paloffs1, intptr_t paloffs2);
+void tvlineasm2(uint32_t vplc2, int32_t vinc1, intptr_t bufplc1, intptr_t bufplc2, uint32_t vplc1, intptr_t p);
+
 void msethlineshift(int32_t logx, int32_t logy);
 void mhline(intptr_t bufplc, uint32_t bx, int32_t cntup16, int32_t junk, uint32_t by, intptr_t p);
 
diff --git a/polymer/eduke32/build/src/a-c.c b/polymer/eduke32/build/src/a-c.c
index 3c7af40e5..a55b014a1 100644
--- a/polymer/eduke32/build/src/a-c.c
+++ b/polymer/eduke32/build/src/a-c.c
@@ -22,6 +22,7 @@ extern void *reciptable;
 static int32_t bpl, transmode = 0;
 static int32_t glogx, glogy, gbxinc, gbyinc, gpinc;
 static char *gbuf, *gpal, *ghlinepal, *gtrans;
+static char *gpal2;
 
 //Global variable functions
 void setvlinebpl(int32_t dabpl) { bpl = dabpl; }
@@ -208,7 +209,7 @@ void mvlineasm4(int32_t cnt, char *p)
 
 void setuptvlineasm(int32_t neglogy) { glogy = neglogy; }
 // cnt+1 loop iterations!
-void tvlineasm1(int32_t vinc, intptr_t paloffs, int32_t cnt, uint32_t vplc, intptr_t bufplc, intptr_t p)
+int32_t tvlineasm1(int32_t vinc, intptr_t paloffs, int32_t cnt, uint32_t vplc, intptr_t bufplc, intptr_t p)
 {
     char ch;
 
@@ -242,8 +243,71 @@ void tvlineasm1(int32_t vinc, intptr_t paloffs, int32_t cnt, uint32_t vplc, intp
         }
         while (--cnt);
     }
+
+    return vplc;
 }
 
+void setuptvlineasm2(int32_t neglogy, intptr_t paloffs1, intptr_t paloffs2)
+{
+    glogy = neglogy;
+    gpal = (char *)paloffs1;
+    gpal2 = (char *)paloffs2;
+}
+// Pass: asm1=vinc2, asm2=pend
+// Return: asm1=vplc1, asm2=vplc2
+void tvlineasm2(uint32_t vplc2, int32_t vinc1, intptr_t bufplc1, intptr_t bufplc2, uint32_t vplc1, intptr_t p)
+{
+    char ch;
+
+    int32_t cnt = (asm2-p-1)/bpl;  // >= 1
+    const int32_t vinc2 = asm1;
+
+    const char *const buf1 = (char *)bufplc1;
+    const char *const buf2 = (char *)bufplc2;
+    const int32_t logy = glogy, ourbpl = bpl, transm = transmode;
+
+    char *pp = (char *)p;
+
+    cnt++;
+
+    if (transm)
+    {
+        do
+        {
+            ch = buf1[vplc1>>logy];
+            if (ch != 255) pp[0] = gtrans[pp[0]|(gpal[ch]<<8)];
+            vplc1 += vinc1;
+
+            ch = buf2[vplc2>>logy];
+            if (ch != 255) pp[1] = gtrans[pp[1]|(gpal2[ch]<<8)];
+            vplc2 += vinc2;
+
+            pp += ourbpl;
+        }
+        while (--cnt > 0);
+    }
+    else
+    {
+        do
+        {
+            ch = buf1[vplc1>>logy];
+            if (ch != 255) pp[0] = gtrans[(pp[0]<<8)|gpal[ch]];
+            vplc1 += vinc1;
+
+            ch = buf2[vplc2>>logy];
+            if (ch != 255) pp[1] = gtrans[(pp[1]<<8)|gpal2[ch]];
+            vplc2 += vinc2;
+
+            pp += ourbpl;
+        }
+        while (--cnt);
+    }
+
+    asm1 = vplc1;
+    asm2 = vplc2;
+}
+
+
 //Floor sprite horizontal line functions
 void msethlineshift(int32_t logx, int32_t logy) { glogx = logx; glogy = logy; }
 // cntup16>>16 + 1 iterations
diff --git a/polymer/eduke32/build/src/engine.c b/polymer/eduke32/build/src/engine.c
index fe8a84d9e..0daa8e5bc 100644
--- a/polymer/eduke32/build/src/engine.c
+++ b/polymer/eduke32/build/src/engine.c
@@ -3875,7 +3875,7 @@ static void transmaskvline(int32_t x)
 //
 // transmaskvline2 (internal)
 //
-#ifndef ENGINE_USING_A_C
+#if 1 //ndef ENGINE_USING_A_C
 static void transmaskvline2(int32_t x)
 {
     int32_t i, y1, y2, x2;
@@ -3950,7 +3950,7 @@ static void transmaskvline2(int32_t x)
 //
 // transmaskwallscan (internal)
 //
-static inline void transmaskwallscan(int32_t x1, int32_t x2)
+static void transmaskwallscan(int32_t x1, int32_t x2)
 {
     int32_t x;
 
@@ -3963,7 +3963,7 @@ static inline void transmaskwallscan(int32_t x1, int32_t x2)
 
     x = x1;
     while ((startumost[x+windowx1] > startdmost[x+windowx1]) && (x <= x2)) x++;
-#ifndef ENGINE_USING_A_C
+#if 1 //ndef ENGINE_USING_A_C
     if ((x <= x2) && (x&1)) transmaskvline(x), x++;
     while (x < x2) transmaskvline2(x), x += 2;
 #endif