Updated changelog, todo, specifications and readme

CHANGELOG.md

@@ -1,8 +1,20 @@
 # Changelog
 
+### v2.3.1-alpha
+
+* Now if a CD-ROM image contains any wrong sector MSF (like PSX libcript protection), then MSF cleanup is disabled because the process will not be lossles if not.
+* Fixed a bug in lzma compression which makes the program to create bigger files. Was much more notizable in images with a lot of merged CDDA and data sectors (Medievil ECM file was 2gb+, which is much more than the original image).
+* If there was an error processing the file, then the output file will be removed (added an option to keep it).
+
+### v2.3.0-alpha
+
+* Added FLAC compression using flaczlib class which helps to make it in a similar way to zlib
+* Added FLAC repository as submodule and updated Makefile
+* Fixed some memory leaks
+
 ### v2.2.0-alpha
 
-* Added lz4 HC compression using a lzlib4 class which helps to make it similar to zlib
+* Added lz4 HC compression using a lzlib4 class which helps to make it in a similar way to zlib
 * Added lz4 repository as submódule and updated Makefile
 * Refactorized analisys, encodong and decoding processes
 * Fixed a bug in header write/read

README.md

@@ -20,9 +20,9 @@ This tests were made using the original ECM tool made by Neill Colett, this ECM
 
 |            | Original Size |    Original ECM   | ECM Tool Reloaded |        7zip       | Original ECM + 7z | ECM Tool Reloaded (Compressed) |
 |:----------:|:-------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:------------------------------:|
-| FFVIII CD3 |  721.508.928  | 635.001.493 (12%) | 574.742.104 (20%) | 438.608.753 (39%) | 365.071.928 (49%) |        367.058.165 (49%)       |
-| Firebugs   |  608.401.248  |  597.995.580 (2%) |  560.062.104 (8%) | 498.312.033 (18%) | 493.397.731 (19%) |        368.291.992 (39%)       |
-| Tekken 3   |  687.922.368  |  645.620.932 (6%) | 606.047.556 (12%) | 457.049.639 (34%) | 426.481.504 (38%) |        421.560.711 (39%)       |
+| FFVIII CD3 |  721.508.928  | 635.001.493 (12%) | 574.742.104 (20%) | 438.608.753 (39%) | 365.071.928 (49%) |        362.903.256 (50%)       |
+| Firebugs   |  608.401.248  |  597.995.580 (2%) |  560.062.104 (8%) | 498.312.033 (18%) | 493.397.731 (19%) |        363.280.352 (40%)       |
+| Tekken 3   |  687.922.368  |  645.620.932 (6%) | 606.047.556 (12%) | 457.049.639 (34%) | 426.481.504 (38%) |        417.325.163 (39%)       |
 
 
 As you can see, the ECM Tool Reloaded uncompressed size is improved a bit by removing more data and dummy sectors from the streams. The CDDA tracks in Firebugs have caused a poor reduction ratio in original ECM tool because it keeps all the sectors. Reloaded conversely removes the GAP data, so  reduction ratio is improved a bit.
@@ -68,4 +68,30 @@ Optional options:
 
 # Changelog
 
-Changelog can be viewed in [CHANGELOG.md](CHANGELOG.md) file
+Changelog can be viewed in [CHANGELOG.md](CHANGELOG.md) file
+
+# FAQ
+
+## How this program works
+
+CDROM data is splitted in sectors, and those sectors are of 2352 bytes size. Depending of the sector type, it may contains Error Detection Code and Error Correction Code data, which will help to the CDROM drive to detect and recover a sector if it is damaged (for example by dust or even little scratches). This data is generated during the burn process and can be easily generated using an algorithm, and like it is not usefull to keep the data outside a CDROM, can be removed. On the decoding process the program will generate that data again and then the image file will be exactly like the original (checked using MD5, SHA1...).
+
+You can find more info in the [CDROM Rainbow books](https://en.wikipedia.org/wiki/Rainbow_Books) 
+
+## Why I have created this "Reloaded" version
+
+The original program created by Neill Corlett is a very good tool... I was using it for a long time to store my CDROM images collection. The problem is that original ecm tool process the file without a clear block separation, storing the sectors info scattered in file. That makes the file unseekable or at least very hard to be seeked, and that is why my program was created.
+
+The ECM2 format keeps the sectors data in header, so just reading that data you can generate an index to seek in file very fast. Then you just will have to read the sector data depending of which type is, and regenerate the removed data easily. I think that this will be usefull for emulators and I am planning to do some tests in the future.
+
+ECM2 format also removes more data than original ECM format (depending of sector type, of course), like MSF if not libcrypt protection exists in game, sync, redundant flags, and zeroed sectors. This will help to save up to 8% of the image size.
+
+Finally, ECM2 format includes compression support which not only will allow to do everything in the same program, also will allow to compress every stream type with a different method, getting higher compression ratios in images that contains CDDA tracks thanks to the FLAC algorithm. The best example is the above Firebugs comparison in compression tests.
+
+## Can be used with Audio CD images
+
+Of course!... but the resulting file will be just the compressed stream if compression is used, and the whole data if not. That data will be packed in a ECM2 file format, which make it unplayable using an Audio Player program (even if FLAC compression is used). The best for that kind of images is to use an Audio CD ripper which allows you to create FLAC or another format files, which can be easily played on an Audio Player.
+
+## Why it still in Alpha
+
+I still working in this tool, so the ECM2 format can suffer changes. Also the tool is not extensive tested and there are bugs that must be fixed before remove the Alpha status. Just now I am perfoming several tests, and on those tests I have located two bugs (one of them was critical).

TODO

@@ -6,6 +6,6 @@
 * ~~Try to use mmap instead manual buffers to see if works faster.~~ Using FILE buffers.
 * ~~Add options to compress the stream directly instead to have to use external programs~~ For now zlib.
 * ~~Convert the program into a library to make it usable by another programs (for now is a class).~~ Doesn't worth. A class is enougth to be used in another programs.
-* Try to ~~detect and~~ compress CDDA tracks using FLAC, APE, or similar.
+* ~~Try to detect and compress CDDA tracks using FLAC, APE, or similar.~~
 * Verify that all CD-ROM modes are working (For now only CDDA, Mode 2 and Mode 2 XA 1 were tested)
-* Add a sumary about the sectors types processed and sizes
+* ~~Add a sumary about the sectors types processed and sizes~~

ecmtool.cpp

@@ -498,7 +498,7 @@ static ecmtool_return_code ecmify(
 
     // End Of TOC reached, so file should have be processed completly
     if (ftello(in) != in_total_size) {
-        printf("\n\nThere was an error processing the input file...\n");
+        printf("\n\nThe input file was not fully readed...\n");
         printfileerror(in, infilename);
         return_code = ECMTOOL_FILE_READ_ERROR;
     }
@@ -535,15 +535,21 @@ static ecmtool_return_code ecmify(
         free(out_buffer);
     }
     // Delete the tocs data
-    delete[] sectors_toc;
-    delete[] streams_toc;
-    delete sTools;
+    if (sectors_toc) {
+        delete[] sectors_toc;
+    }
+    if (streams_toc) {
+        delete[] streams_toc;
+    }
+    if (sTools) {
+        delete sTools;
+    }
 
     if (!return_code) {
         printf("\n\nFinished!\n");
     }
     else {
-        printf("\n\nThere was an error processing the input file: %d", return_code);
+        printf("\n\nThere was an error processing the input file: %d\n\n", return_code);
         // We will remove the file if something went wrong
         if (!options->keep_output) {
             if (remove(outfilename)) {
@@ -726,16 +732,22 @@ static ecmtool_return_code unecmify(
     if (out_buffer) {
         free(out_buffer);
     }
-    delete[] sectors_toc;
-    delete[] streams_toc;
-    delete sTools;
+    if (sectors_toc) {
+        delete[] sectors_toc;
+    }
+    if (streams_toc) {
+        delete[] streams_toc;
+    }
+    if (sTools) {
+        delete sTools;
+    }
 
     // If something went wrong, inform the user and delete the output file
     if (!return_code) {
         printf("\n\nFinished!\n");
     }
     else {
-        printf("\n\nThere was an error processing the input file: %d", return_code);
+        printf("\n\nThere was an error processing the input file: %d\n\n", return_code);
         // We will remove the file if something went wrong
         if (!options->keep_output) {
             if (remove(outfilename)) {

ecmtool.h

@@ -2,7 +2,7 @@
 //
 #define TITLE "ecmtool - Encoder/decoder for Error Code Modeler format, with advanced features"
 #define COPYR "Copyright (C) 2021 Daniel Carrasco"
-#define VERSI "2.3.0-alpha"
+#define VERSI "2.3.1-alpha"
 //
 // This program is free software: you can redistribute it and/or modify
 // it under the terms of the GNU General Public License as published by