From 4affb3811c4a57cb4e033ddd40f5e99236b59250 Mon Sep 17 00:00:00 2001
From: Dirk Engling <erdgeist@erdgeist.org>
Date: Sat, 8 Feb 2014 23:39:56 +0100
Subject: Speed up decompress script by avoiding to wastefully copying a
 megabyte of data each time we scan for a new decompression offset

---
 src/decompress.c | 55 +++++++++++++++++++++++++++++--------------------------
 1 file changed, 29 insertions(+), 26 deletions(-)

diff --git a/src/decompress.c b/src/decompress.c
index 6074db6..fbf33ec 100644
--- a/src/decompress.c
+++ b/src/decompress.c
@@ -1,52 +1,55 @@
-#include <zlib.h>
+#include <stdlib.h>
 #include <stdio.h>
 #include <unistd.h>
 #include <fcntl.h>
 #include <string.h>
+#include <zlib.h>
+#include "mystdlib.h"
 
 #define XORLEN (29)
 #define HUGEBLOCK (1024*1024)
 
 int main(int argc, char **argv) {
-//  int infile = open("/Volumes/DasTelefonbuch/atb/phonebook.db", O_RDONLY);
-
-  int infile = open( argv[1], O_RDONLY);
+  MAP in = map_file( argv[1], 1 );
 
   unsigned const char xorkey [XORLEN] = "Just for Fun. Linus Torvalds.";
-  unsigned char input  [HUGEBLOCK];
+  unsigned char input  [XORLEN];
   unsigned char output [HUGEBLOCK];
   char respath[32]; /* file_XXXXX\0 */
-  int i, offs = 0, zres = 0, filenum = 0, resfile, avail=1;
-  z_stream z; memset( &z, 0, sizeof(z));
+  int i, zres = 0, filenum = 0, resfile;
+  size_t offs = 0;
 
-  while( avail ) {
-    do {
-      lseek( infile, offs, SEEK_SET );
-      avail = read( infile, input, HUGEBLOCK );
-      if( !avail) break;
-      for( i=0; i<XORLEN; ++i ) input[i] ^= xorkey[i];
+  z_stream z; memset( &z, 0, sizeof(z));
 
-      z.next_in  = input;  z.avail_in  = avail;
-      z.next_out = output; z.avail_out = HUGEBLOCK;
-      inflateInit( &z ); zres = inflate( &z, Z_NO_FLUSH );
+  while( offs < in->size - XORLEN ) {
+    for( i=0; i<XORLEN; ++i ) input[i] = in->addr[offs+i] ^ xorkey[i];
+    z.next_in  = input;  z.avail_in = XORLEN;
+    z.next_out = output; z.avail_out = HUGEBLOCK;
+    inflateInit( &z ); zres = inflate( &z, Z_NO_FLUSH );
+    if( (zres != Z_OK) && (zres != Z_STREAM_END) )
+      goto error_continue;
 
-      if( (zres != Z_OK) && (zres != Z_STREAM_END) ) {
-        inflateEnd(&z); memset( &z, 0, sizeof(z)); 
-        offs++;
-      }
-    } while((zres != Z_OK) && (zres != Z_STREAM_END));
+    z.next_in  = in->addr + offs + XORLEN; z.avail_in = in->size - offs - XORLEN;
+    while( zres == Z_OK ) zres = inflate( &z, Z_NO_FLUSH );
 
-    if( !avail ) break;
-    if( zres == Z_OK) while( inflate( &z, Z_NO_FLUSH ) == Z_OK );
+    if( zres != Z_STREAM_END ) {
+error_continue:
+      inflateEnd(&z); memset( &z, 0, sizeof(z));
+      offs++;
+      continue;
+    }
 
     sprintf( respath, "file_%05X", filenum++ );
     resfile = open( respath, O_RDWR | O_CREAT, 0644 );
-    /* I know, I know, error checking */
+    if( resfile < 0 ) {
+      fprintf( stderr, "Could not open output file %s\n", respath );
+      exit(1);
+    }
     write( resfile, output, z.total_out );
     close( resfile );
     offs += z.total_in;
 
-    inflateEnd(&z); memset( &z, 0, sizeof(z)); 
+    inflateEnd(&z); memset( &z, 0, sizeof(z));
   }
-  close( infile );
+  unmap_file(&in);
 }
-- 
cgit v1.2.3