Isolate frame implementation details

Makefile

@@ -11,6 +11,7 @@ $(OBJDIR)/%.o: src/../%.c
 APP := lzsa
 
 OBJS := $(OBJDIR)/src/main.o
+OBJS += $(OBJDIR)/src/frame.o
 OBJS += $(OBJDIR)/src/shrink.o
 OBJS += $(OBJDIR)/src/expand.o
 OBJS += $(OBJDIR)/src/libdivsufsort/lib/divsufsort.o

VS2017/lzsa.vcxproj

@@ -179,6 +179,7 @@
   <ItemGroup>
     <ClInclude Include="..\src\expand.h" />
     <ClInclude Include="..\src\format.h" />
+    <ClInclude Include="..\src\frame.h" />
     <ClInclude Include="..\src\libdivsufsort\include\config.h" />
     <ClInclude Include="..\src\libdivsufsort\include\divsufsort.h" />
     <ClInclude Include="..\src\libdivsufsort\include\divsufsort_private.h" />
@@ -187,6 +188,7 @@
   </ItemGroup>
   <ItemGroup>
     <ClCompile Include="..\src\expand.c" />
+    <ClCompile Include="..\src\frame.c" />
     <ClCompile Include="..\src\libdivsufsort\lib\divsufsort.c" />
     <ClCompile Include="..\src\libdivsufsort\lib\sssort.c" />
     <ClCompile Include="..\src\libdivsufsort\lib\trsort.c" />

VS2017/lzsa.vcxproj.filters

@@ -45,6 +45,9 @@
     <ClInclude Include="..\src\libdivsufsort\include\divsufsort_private.h">
       <Filter>Fichiers sources\libdivsufsort\include</Filter>
     </ClInclude>
+    <ClInclude Include="..\src\frame.h">
+      <Filter>Fichiers sources</Filter>
+    </ClInclude>
   </ItemGroup>
   <ItemGroup>
     <ClCompile Include="..\src\main.c">
@@ -68,5 +71,8 @@
     <ClCompile Include="..\src\libdivsufsort\lib\trsort.c">
       <Filter>Fichiers sources\libdivsufsort\lib</Filter>
     </ClCompile>
+    <ClCompile Include="..\src\frame.c">
+      <Filter>Fichiers sources</Filter>
+    </ClCompile>
   </ItemGroup>
 </Project>

src/frame.c

@@ -0,0 +1,177 @@
+/*
+ * frame.c - frame implementation
+ *
+ * Copyright (C) 2019 Emmanuel Marty
+ *
+ * This software is provided 'as-is', without any express or implied
+ * warranty.  In no event will the authors be held liable for any damages
+ * arising from the use of this software.
+ *
+ * Permission is granted to anyone to use this software for any purpose,
+ * including commercial applications, and to alter it and redistribute it
+ * freely, subject to the following restrictions:
+ *
+ * 1. The origin of this software must not be misrepresented; you must not
+ *    claim that you wrote the original software. If you use this software
+ *    in a product, an acknowledgment in the product documentation would be
+ *    appreciated but is not required.
+ * 2. Altered source versions must be plainly marked as such, and must not be
+ *    misrepresented as being the original software.
+ * 3. This notice may not be removed or altered from any source distribution.
+ */
+
+#include <stdlib.h>
+#include "frame.h"
+#include "shrink.h"
+
+#define LZSA_ID_0   0x7b
+#define LZSA_ID_1   0x9e
+
+/**
+ * Get compressed file header size
+ *
+ * @return file header size
+ */
+int lzsa_get_header_size(void) {
+   return 3;
+}
+
+/**
+ * Get compressed frame header size
+ *
+ * @return frame header size
+ */
+int lzsa_get_frame_size(void) {
+   return 3;
+}
+
+/**
+ * Encode file header
+ *
+ * @param pFrameData encoding buffer
+ * @param nMaxFrameDataSize max encoding buffer size, in bytes
+ *
+ * @return number of encoded bytes, or -1 for failure
+ */
+int lzsa_encode_header(unsigned char *pFrameData, const int nMaxFrameDataSize) {
+   if (nMaxFrameDataSize >= 3) {
+      pFrameData[0] = LZSA_ID_0;                         /* Magic number */
+      pFrameData[1] = LZSA_ID_1;
+      pFrameData[2] = 0;                                 /* Format version 1 */
+
+      return 3;
+   }
+   else {
+      return -1;
+   }
+}
+
+/**
+ * Encode compressed block frame header
+ *
+ * @param pFrameData encoding buffer
+ * @param nMaxFrameDataSize max encoding buffer size, in bytes
+ * @param nBlockDataSize compressed block's data size, in bytes
+ *
+ * @return number of encoded bytes, or -1 for failure
+ */
+int lzsa_encode_compressed_block_frame(unsigned char *pFrameData, const int nMaxFrameDataSize, const int nBlockDataSize) {
+   if (nMaxFrameDataSize >= 3 && nBlockDataSize <= 0x7fffff) {
+      pFrameData[0] = nBlockDataSize & 0xff;
+      pFrameData[1] = (nBlockDataSize >> 8) & 0xff;
+      pFrameData[2] = (nBlockDataSize >> 16) & 0x7f;
+
+      return 3;
+   }
+   else {
+      return -1;
+   }
+}
+
+/**
+ * Encode uncompressed block frame header
+ *
+ * @param pFrameData encoding buffer
+ * @param nMaxFrameDataSize max encoding buffer size, in bytes
+ * @param nBlockDataSize uncompressed block's data size, in bytes
+ *
+ * @return number of encoded bytes, or -1 for failure
+ */
+int lzsa_encode_uncompressed_block_frame(unsigned char *pFrameData, const int nMaxFrameDataSize, const int nBlockDataSize) {
+   if (nMaxFrameDataSize >= 3 && nBlockDataSize <= 0x7fffff) {
+      pFrameData[0] = nBlockDataSize & 0xff;
+      pFrameData[1] = (nBlockDataSize >> 8) & 0xff;
+      pFrameData[2] = ((nBlockDataSize >> 16) & 0x7f) | 0x80;   /* Uncompressed block */
+
+      return 3;
+   }
+   else {
+      return -1;
+   }
+}
+
+/**
+ * Encode terminal frame header
+ *
+ * @param pFrameData encoding buffer
+ * @param nMaxFrameDataSize max encoding buffer size, in bytes
+ *
+ * @return number of encoded bytes, or -1 for failure
+ */
+int lzsa_encode_footer_frame(unsigned char *pFrameData, const int nMaxFrameDataSize) {
+   if (nMaxFrameDataSize >= 3) {
+      pFrameData[0] = 0x00;         /* EOD frame */
+      pFrameData[1] = 0x00;
+      pFrameData[2] = 0x00;
+
+      return 3;
+   }
+   else {
+      return -1;
+   }
+}
+
+/**
+ * Decode file header
+ *
+ * @param pFrameData data bytes
+ * @param nFrameDataSize number of bytes to decode
+ *
+ * @return 0 for success, or -1 for failure
+ */
+int lzsa_decode_header(const unsigned char *pFrameData, const int nFrameDataSize) {
+   if (nFrameDataSize != 3 ||
+      pFrameData[0] != LZSA_ID_0 ||
+      pFrameData[1] != LZSA_ID_1 ||
+      pFrameData[2] != 0) {
+      return -1;
+   }
+   else {
+      return 0;
+   }
+}
+
+/**
+ * Decode frame header
+ *
+ * @param pFrameData data bytes
+ * @param nFrameDataSize number of bytes to decode
+ * @param nBlockSize pointer to block size, updated if this function succeeds (set to 0 if this is the terminal frame)
+ * @param nIsUncompressed pointer to compressed block flag, updated if this function succeeds
+ *
+ * @return 0 for success, or -1 for failure
+ */
+int lzsa_decode_frame(const unsigned char *pFrameData, const int nFrameDataSize, unsigned int *nBlockSize, int *nIsUncompressed) {
+   if (nFrameDataSize == 3) {
+      *nBlockSize = ((unsigned int)pFrameData[0]) |
+         (((unsigned int)pFrameData[1]) << 8) |
+         (((unsigned int)pFrameData[2]) << 16);
+
+      *nIsUncompressed = ((*nBlockSize & 0x800000) != 0) ? 1 : 0;
+      *nBlockSize &= 0x7fffff;
+      return 0;
+   }
+   else {
+      return -1;
+   }
+}

src/frame.h

@@ -0,0 +1,104 @@
+/*
+ * frame.h - frame definitions
+ *
+ * Copyright (C) 2019 Emmanuel Marty
+ *
+ * This software is provided 'as-is', without any express or implied
+ * warranty.  In no event will the authors be held liable for any damages
+ * arising from the use of this software.
+ *
+ * Permission is granted to anyone to use this software for any purpose,
+ * including commercial applications, and to alter it and redistribute it
+ * freely, subject to the following restrictions:
+ *
+ * 1. The origin of this software must not be misrepresented; you must not
+ *    claim that you wrote the original software. If you use this software
+ *    in a product, an acknowledgment in the product documentation would be
+ *    appreciated but is not required.
+ * 2. Altered source versions must be plainly marked as such, and must not be
+ *    misrepresented as being the original software.
+ * 3. This notice may not be removed or altered from any source distribution.
+ */
+
+#ifndef _FRAME_H
+#define _FRAME_H
+
+/**
+ * Get compressed file header size
+ *
+ * @return file header size
+ */
+int lzsa_get_header_size(void);
+
+/**
+ * Get compressed frame header size
+ *
+ * @return frame header size
+ */
+int lzsa_get_frame_size(void);
+
+/**
+ * Encode file header
+ *
+ * @param pFrameData encoding buffer
+ * @param nMaxFrameDataSize max encoding buffer size, in bytes
+ *
+ * @return number of encoded bytes, or -1 for failure
+ */
+int lzsa_encode_header(unsigned char *pFrameData, const int nMaxFrameDataSize);
+
+/**
+ * Encode compressed block frame header
+ *
+ * @param pFrameData encoding buffer
+ * @param nMaxFrameDataSize max encoding buffer size, in bytes
+ * @param nBlockDataSize compressed block's data size, in bytes
+ *
+ * @return number of encoded bytes, or -1 for failure
+ */
+int lzsa_encode_compressed_block_frame(unsigned char *pFrameData, const int nMaxFrameDataSize, const int nBlockDataSize);
+
+/**
+ * Encode uncompressed block frame header
+ *
+ * @param pFrameData encoding buffer
+ * @param nMaxFrameDataSize max encoding buffer size, in bytes
+ * @param nBlockDataSize uncompressed block's data size, in bytes
+ *
+ * @return number of encoded bytes, or -1 for failure
+ */
+int lzsa_encode_uncompressed_block_frame(unsigned char *pFrameData, const int nMaxFrameDataSize, const int nBlockDataSize);
+
+/**
+ * Encode terminal frame header
+ *
+ * @param pFrameData encoding buffer
+ * @param nMaxFrameDataSize max encoding buffer size, in bytes
+ *
+ * @return number of encoded bytes, or -1 for failure
+ */
+int lzsa_encode_footer_frame(unsigned char *pFrameData, const int nMaxFrameDataSize);
+
+/**
+ * Decode file header
+ *
+ * @param pFrameData data bytes
+ * @param nFrameDataSize number of bytes to decode
+ *
+ * @return 0 for success, or -1 for failure
+ */
+int lzsa_decode_header(const unsigned char *pFrameData, const int nFrameDataSize);
+
+/**
+ * Decode frame header
+ *
+ * @param pFrameData data bytes
+ * @param nFrameDataSize number of bytes to decode
+ * @param nBlockSize pointer to block size, updated if this function succeeds (set to 0 if this is the terminal frame)
+ * @param nIsUncompressed pointer to compressed block flag, updated if this function succeeds
+ *
+ * @return 0 for success, or -1 for failure
+ */
+int lzsa_decode_frame(const unsigned char *pFrameData, const int nFrameDataSize, unsigned int *nBlockSize, int *nIsUncompressed);
+
+#endif /* _FRAME_H */

src/main.c

@@ -30,6 +30,7 @@
 #include <sys/time.h>
 #endif
 #include "format.h"
+#include "frame.h"
 #include "shrink.h"
 #include "expand.h"
 
@@ -67,6 +68,7 @@ static int lzsa_compress(const char *pszInFilename, const char *pszOutFilename,
    long long nOriginalSize = 0LL, nCompressedSize = 0LL;
    int nFlags;
    int nResult;
+   unsigned char cFrameData[16];
    bool bError = false;
 
    f_in = fopen(pszInFilename, "rb");
@@ -163,14 +165,13 @@ static int lzsa_compress(const char *pszInFilename, const char *pszOutFilename,
    }
 
    if ((nOptions & OPT_RAW) == 0) {
-      unsigned char cHeader[3];
+      int nHeaderSize = lzsa_encode_header(cFrameData, 16);
-
+      if (nHeaderSize < 0)
-      cHeader[0] = 0x7b;                              /* Magic number: 0x9e7b */
+         bError = true;
-      cHeader[1] = 0x9e;
+      else {
-      cHeader[2] = 0;                                 /* Format version 1 */
+         bError = fwrite(cFrameData, 1, nHeaderSize, f_out) != nHeaderSize;
-
+         nCompressedSize += (long long)nHeaderSize;
-      bError = fwrite(cHeader, 1, 3, f_out) != 3;
+      }
-      nCompressedSize += 3LL;
    }
 
    if (nOptions & OPT_VERBOSE) {
@@ -205,15 +206,14 @@ static int lzsa_compress(const char *pszInFilename, const char *pszOutFilename,
             /* Write compressed block */
 
             if ((nOptions & OPT_RAW) == 0) {
-               unsigned char cBlockSize[3];
+               int nBlockheaderSize = lzsa_encode_compressed_block_frame(cFrameData, 16, nOutDataSize);
-
+               if (nBlockheaderSize < 0)
-               cBlockSize[0] = nOutDataSize & 0xff;
-               cBlockSize[1] = (nOutDataSize >> 8) & 0xff;
-               cBlockSize[2] = (nOutDataSize >> 16) & 0xff;
-               nCompressedSize += 3LL;
-
-               if (fwrite(cBlockSize, 1, 3, f_out) != (size_t)3) {
                   bError = true;
+               else {
+                  nCompressedSize += (long long)nBlockheaderSize;
+                  if (fwrite(cFrameData, 1, nBlockheaderSize, f_out) != (size_t)nBlockheaderSize) {
+                     bError = true;
+                  }
                }
             }
 
@@ -236,13 +236,11 @@ static int lzsa_compress(const char *pszInFilename, const char *pszOutFilename,
                break;
             }
 
-            unsigned char cBlockSize[3];
+            int nBlockheaderSize = lzsa_encode_uncompressed_block_frame(cFrameData, 16, nInDataSize);
-
+            if (nBlockheaderSize < 0)
-            cBlockSize[0] = nInDataSize & 0xff;
+               bError = true;
-            cBlockSize[1] = (nInDataSize >> 8) & 0xff;
+            else {
-            cBlockSize[2] = ((nInDataSize >> 16) & 0x7f) | 0x80;   /* Uncompressed block */
+               if (fwrite(cFrameData, 1, nBlockheaderSize, f_out) != (size_t)nBlockheaderSize) {
-
-            if (fwrite(cBlockSize, 1, 3, f_out) != (size_t)3) {
                   bError = true;
                }
                else {
@@ -251,7 +249,8 @@ static int lzsa_compress(const char *pszInFilename, const char *pszOutFilename,
                   }
                   else {
                      nOriginalSize += (long long)nInDataSize;
-                  nCompressedSize += 3LL + (long long)nInDataSize;
+                     nCompressedSize += (long long)nBlockheaderSize + (long long)nInDataSize;
+                  }
                }
             }
          }
@@ -265,21 +264,19 @@ static int lzsa_compress(const char *pszInFilename, const char *pszOutFilename,
       }
    }
 
-   unsigned char cFooter[4];
    int nFooterSize;
 
    if ((nOptions & OPT_RAW) != 0) {
       nFooterSize = 0;
    }
    else {
-      cFooter[0] = 0x00;         /* EOD frame */
+      nFooterSize = lzsa_encode_footer_frame(cFrameData, 16);
-      cFooter[1] = 0x00;
+      if (nFooterSize < 0)
-      cFooter[2] = 0x00;
+         bError = true;
-      nFooterSize = 3;
    }
 
    if (!bError)
-      bError = fwrite(cFooter, 1, nFooterSize, f_out) != nFooterSize;
+      bError = fwrite(cFrameData, 1, nFooterSize, f_out) != nFooterSize;
    nCompressedSize += (long long)nFooterSize;
 
    if (!bError && (nOptions & OPT_VERBOSE)) {
@@ -322,6 +319,7 @@ static int lzsa_decompress(const char *pszInFilename, const char *pszOutFilename
    long long nStartTime = 0LL, nEndTime = 0LL;
    long long nOriginalSize = 0LL;
    unsigned int nFileSize = 0;
+   unsigned char cFrameData[16];
 
    FILE *pInFile = fopen(pszInFilename, "rb");
    if (!pInFile) {
@@ -330,20 +328,17 @@ static int lzsa_decompress(const char *pszInFilename, const char *pszOutFilename
    }
 
    if ((nOptions & OPT_RAW) == 0) {
-      unsigned char cHeader[3];
+      const int nHeaderSize = lzsa_get_header_size();
 
-      memset(cHeader, 0, 3);
+      memset(cFrameData, 0, 16);
-
+      if (fread(cFrameData, 1, nHeaderSize, pInFile) != nHeaderSize) {
-      if (fread(cHeader, 1, 3, pInFile) != 3) {
          fclose(pInFile);
          pInFile = NULL;
          fprintf(stderr, "error reading header in input file\n");
          return 100;
       }
 
-      if (cHeader[0] != 0x7b ||
+      if (lzsa_decode_header(cFrameData, nHeaderSize) < 0) {
-         cHeader[1] != 0x9e ||
-         cHeader[2] != 0) {
          fclose(pInFile);
          pInFile = NULL;
          fprintf(stderr, "invalid magic number or format version in input file\n");
@@ -440,19 +435,21 @@ static int lzsa_decompress(const char *pszInFilename, const char *pszOutFilename
 
    while (!feof(pInFile) && !nDecompressionError) {
       unsigned int nBlockSize = 0;
+      int nIsUncompressed = 0;
 
       if (nPrevDecompressedSize != 0) {
          memcpy(pOutData + BLOCK_SIZE - nPrevDecompressedSize, pOutData + BLOCK_SIZE, nPrevDecompressedSize);
       }
 
       if ((nOptions & OPT_RAW) == 0) {
-         unsigned char cBlockSize[3];
+         const int nFrameSize = lzsa_get_frame_size();
 
-         memset(cBlockSize, 0, 3);
+         memset(cFrameData, 0, 16);
-         if (fread(cBlockSize, 1, 3, pInFile) == 3) {
+         if (fread(cFrameData, 1, nFrameSize, pInFile) == nFrameSize) {
-            nBlockSize = ((unsigned int)cBlockSize[0]) |
+            if (lzsa_decode_frame(cFrameData, nFrameSize, &nBlockSize, &nIsUncompressed) < 0) {
-               (((unsigned int)cBlockSize[1]) << 8) |
+               nDecompressionError = 1;
-               (((unsigned int)cBlockSize[2]) << 16);
+               nBlockSize = 0;
+            }
          }
          else {
             nBlockSize = 0;
@@ -465,16 +462,14 @@ static int lzsa_decompress(const char *pszInFilename, const char *pszOutFilename
       }
 
       if (nBlockSize != 0) {
-         bool bIsUncompressed = (nBlockSize & 0x800000) != 0;
          int nDecompressedSize = 0;
 
-         nBlockSize &= 0x7fffff;
          if ((int)nBlockSize > BLOCK_SIZE) {
             fprintf(stderr, "block size %d > max size %d\n", nBlockSize, BLOCK_SIZE);
             break;
          }
          if (fread(pInBlock, 1, nBlockSize, pInFile) == nBlockSize) {
-            if (bIsUncompressed) {
+            if (nIsUncompressed) {
                memcpy(pOutData + BLOCK_SIZE, pInBlock, nBlockSize);
                nDecompressedSize = nBlockSize;
             }
@@ -539,6 +534,7 @@ static int lzsa_compare(const char *pszInFilename, const char *pszOutFilename, c
    long long nOriginalSize = 0LL;
    long long nKnownGoodSize = 0LL;
    unsigned int nFileSize = 0;
+   unsigned char cFrameData[16];
 
    FILE *pInFile = fopen(pszInFilename, "rb");
    if (!pInFile) {
@@ -547,20 +543,17 @@ static int lzsa_compare(const char *pszInFilename, const char *pszOutFilename, c
    }
 
    if ((nOptions & OPT_RAW) == 0) {
-      unsigned char cHeader[3];
+      const int nHeaderSize = lzsa_get_header_size();
 
-      memset(cHeader, 0, 3);
+      memset(cFrameData, 0, 16);
-
+      if (fread(cFrameData, 1, nHeaderSize, pInFile) != nHeaderSize) {
-      if (fread(cHeader, 1, 3, pInFile) != 3) {
          fclose(pInFile);
          pInFile = NULL;
          fprintf(stderr, "error reading header in compressed input file\n");
          return 100;
       }
 
-      if (cHeader[0] != 0x7b ||
+      if (lzsa_decode_header(cFrameData, nHeaderSize) < 0) {
-         cHeader[1] != 0x9e ||
-         cHeader[2] != 0) {
          fclose(pInFile);
          pInFile = NULL;
          fprintf(stderr, "invalid magic number or format version in input file\n");
@@ -679,6 +672,7 @@ static int lzsa_compare(const char *pszInFilename, const char *pszOutFilename, c
 
    while (!feof(pInFile) && !nDecompressionError && !bComparisonError) {
       unsigned int nBlockSize = 0;
+      int nIsUncompressed = 0;
 
       if (nPrevDecompressedSize != 0) {
          memcpy(pOutData + BLOCK_SIZE - nPrevDecompressedSize, pOutData + BLOCK_SIZE, nPrevDecompressedSize);
@@ -687,13 +681,14 @@ static int lzsa_compare(const char *pszInFilename, const char *pszOutFilename, c
       int nBytesToCompare = (int)fread(pCompareData, 1, BLOCK_SIZE, pOutFile);
 
       if ((nOptions & OPT_RAW) == 0) {
-         unsigned char cBlockSize[3];
+         const int nFrameSize = lzsa_get_frame_size();
 
-         memset(cBlockSize, 0, 3);
+         memset(cFrameData, 0, 16);
-         if (fread(cBlockSize, 1, 3, pInFile) == 3) {
+         if (fread(cFrameData, 1, nFrameSize, pInFile) == nFrameSize) {
-            nBlockSize = ((unsigned int)cBlockSize[0]) |
+            if (lzsa_decode_frame(cFrameData, nFrameSize, &nBlockSize, &nIsUncompressed) < 0) {
-               (((unsigned int)cBlockSize[1]) << 8) |
+               nDecompressionError = 1;
-               (((unsigned int)cBlockSize[2]) << 16);
+               nBlockSize = 0;
+            }
          }
          else {
             nBlockSize = 0;
@@ -706,16 +701,14 @@ static int lzsa_compare(const char *pszInFilename, const char *pszOutFilename, c
       }
 
       if (nBlockSize != 0) {
-         bool bIsUncompressed = (nBlockSize & 0x800000) != 0;
          int nDecompressedSize = 0;
 
-         nBlockSize &= 0x7fffff;
          if ((int)nBlockSize > BLOCK_SIZE) {
             fprintf(stderr, "block size %d > max size %d\n", nBlockSize, BLOCK_SIZE);
             break;
          }
          if (fread(pInBlock, 1, nBlockSize, pInFile) == nBlockSize) {
-            if (bIsUncompressed) {
+            if (nIsUncompressed) {
                memcpy(pOutData + BLOCK_SIZE, pInBlock, nBlockSize);
                nDecompressedSize = nBlockSize;
             }