$Id: debug.txt,v 1.2 2003/04/12 16:16:05 andrew_belov Exp $ARJ DEBUGGER'S GUIDEThis file is a reference for everyone who has to deal with ARJ technicalsupport and debugging.DEBUG CONFIGURATION        Debugging begins with compiling   ARJ for debug mode.   Supply a        DEBUG=1 clause to the make program, as in the following example:        nmake MODE=OS232 COMPILER=CSET2 DEBUG=1 prepare all         This will yield   an unoptimized executable with some   internal        assertions and, possibly,  further debugging functionality which        is platform-specific (e.g. debug  malloc() in IBM C Set/2 static        libraries).        The debug flavor also   supports a simple malloc() tracing  with        "-hdm", probably useful for working  out memory shortages in the        16-bit world.TOP 5 TEST CASES        Whenever a large volume of changes is committed, or ARJ is being        tested under  a new  platform,  we  suggest  the following  test        scenario:        1. Ensure  that all  modules  compile   normally, and the  build        process finishes without errors  when "make package" is  issued.        This warrants basic functionality to be there.        2. Examine the usual operation of ARJ module  with "arj a -jt1",        submitting   a  combination  of  options while   archiving  some        "payload" (the volume to be archived is of little importance, 10        files totaling 1 megabyte would be enough for testing):	arj a test -+ -v100K -jt1 -hk5 -i5 -jm -hg -garble        The "-jt1"  will let you see  if  there are  any major omissions        which would result  in  broken  archives. Re-check the   created        archive with:        arj q test.arj  [should report no damage]        arj v test.arj        arj t -v -garble test.arj        arj x -v -garble test.arj        3. Check the ARJSFX branches:        ARJSFXV -> arj a test -+ -va -je        ARJSFX -> arj a test -+ -je        ARJSFXJR -> arj a test -+ -je1 [may need "-2x", see the output]        Each of   the SFX levels  shares  a  portion of  ARJ  code while        bringing in some distinct routines,   so  we put a priority   on        testing the SFX.        4. Check REARJ:        rearj test.zip        (need another archiver and a REARJ.CFG for that). REARJ utilizes        the platform-specific process creation routines.        5. Verify the compatibility with the  mainstream versions of ARJ        available at the project   download page, or from  ARJ Software,        Inc.   Extract    the created  archive by   using    a "branded"        distribution of ARJ, then do  quite the contrary by packing  the        archive in one version, then extracting it with another one.        Pay special attention to the encryption issue:        XOR encryption: arj a test -hg -g...        GOST 40-bit cipher: arj a test -hg! -g...        GOST 128-bit modular cipher: arj a test -hg -g...        Supply your  password of choice for   "-g..." when archiving and        extracting/testing.APPENDIX: DEBUG OPTIONSA list of debug options (-hd<opt>, followed by an optional argument):  .     (ASR) Displays each file being added to a filelist.  ?     With ARJ v 2.75.03+, allows use of Extended license keys in standard        ARJ. The filelist capacity shall remain the same.  a     Rigorous archive handling. Enables detailed information in index        files, and forces recovery of ARJ-PROTECTed archives.  b<u>  Sets VBUF size. VBUF is used for buffering of archive input stream.  c     Skips CRC checking for extracted files, thereby causing CRC errors.  d     Sets the dictionary size on compression, thus making ARJ compress        better. The archives can still be extracted... with ARJZ only, though.  f     Notifies about len_cnt fixes in encoding routine.  i<ld> Defines the flist_main capacity in ARJ. Extended license is not        required in this case.  k     Zeroes compsize after the file has been unpacked.  l     Displays the encryption version.  m     (ASR, DEBUG builds only) Displays malloc()/farmalloc() arguments.  n     Inhibits all kinds of file generation activities.  o<s>  Sets archive modification time to the desired timestamp.  p<u>  Sets TBUF size. TBUF is used for buffering of archive output stream.  q     Reverts to the standard filelist type for flist_main.  s     Enables unlimited ranges for numeric switch values.  t     Enables the poor man's profiler (displays execution ticks).  v     Dumps the command-line tokens as they are processed, shows memory        statistics and may have other effects as a "verbose" command.  x     Inhibits any sort of hollow file processing.  y     Uses a case-preserving way of filename/path storage.  z<ld> Sets the physical filelist capacity limit.Common examples:* -hdaflmv is a usual bug reporter's test case.* -hdt may be used for performance measurements.