Well. Here is the progress of JGMOD library.


December 1997
-------------
24  Got 2 MOD documentation from oulu. Neither give a very good description
    about the MOD format but with both combined, just enough.

25  Merry Christmas !! Decided to make a mod library and I called it JGMOD.
    You might be wondering what JG means. My actual name is Guan Foo Wah.
    Guan is my surname and people call me Jeffery, or Jeff. So JG stands for
    Jeffery Guan. I could call it GFWMOD but it really sounds terrible to me.

    Started making the library. The birth of this library !!! Able to load
    some of the MOD 31 instruments header (song name and all the 31
    instruments info). 

26  Revamped in the loading header function. It was untidy. Now it uses
    Allegro's pack_stuff routines. Able to load the MOD pattern for 4 channels
    mod.

27  Finally figured out how to load X channel mod. It is really easy. Some
    other damm documentation confuses me with it.  Able to load all the 31
    instruments sample. Having trouble to play those sample using Allegro's
    play_sample() function. Sounds too terrible.

28  Tried minus 127 with all of those 31 instruments samples and hell it
    works !! play_sample() have no problem playing those samples now. That
    was a lucky try. Made an interrupt to play those mod files assuming the
    music is at tempo 6 and bmp 125. Since it uses play_sample(), a new
    sample played will not cancel any sample on the same channel. Echos and
    feedback will be heard. Yuck !

29  Started making this history to whovever is interested.
    Now it uses Allegro's lower level digital sample routines. No more
    feedback and echos. YEAAA... Plays samples in loop mode if required.

30  Now plays the sample at the correct volume. Pattern Break command (13)
    is now supported. Able to play music at different tempo and bmp(15).

31  Supports Fine Volume Slide Up and Fine Volume Slide Down (14-10
    and 14-11).


January 1998
------------
3   Supports set sample offset(9).

5   Fixed Fine Volume Slide Up and Fine Volume Slide down bug

7   Changed the way to handle bpm by changing the interrupt rate
    when needed.

8   Fixed a pattern break bug. Supports MOD 15 instruments now. 

9   Volume slide (10) supported

10  Read the S3M format. Could not understand much. Change to NTSC clock rate
    cause S3M c2spd is in that rate. Implemented the finetune effect (14-5).

11  Implemented the period slide up (pormento up, 1) and slide
    down(promento down, 2) effect. Slide to note(3) is also supported but
    buggy.

12  Fixed slide to note bug, but a bug still exists.

26  Fixed another slide to note bug. I think it is free of bug now. Did a
    makefile for this library.

28  Supports Retrigger Sample (14-9) and Cut Sample (14-12).

29  Delay Sample (14-13) and Pormento to note+volume slide suported (5).
    Vibrato+volume slide is also supported but without the vibrato.
    Make prev_pattern() and next_pattern() to skip pattern.

30  Now the song can play in a loop mode. Fixed a bug in the skip pattern
    functions.  


Febuary 1998
------------
13  Opps. Haven't update this library for a long time. Did stop_mod function
    to stop the music. Updated the mod loader to support up to 32 channels
    but not player.

14  Did a loader for XM partially.

16  The player now supports up to 32 channels and the number of channels
    reserved for mod can be specified instead of the default 24 channels.

19  Changed how the mod interrupt works. Global commands will not be skiped
    if the no of channels reserved for the player is less than no of
    channels used by the song. Did a readme.txt. I am going to upload this to
    my web page tomorrow.

20  Library uploaded. 

23  Did the vibrato effect (4). Yea... !!!!

24  Added some preprocessor commands to jgmod.h. Did a cpp program to test
    with this library. It compiled fine. No errors at all. But when I ran the
    program, there is no sound at all. I have no clue at all why this
    is happening. I don't know much about cpp.

25  Did set pan effect (8) and 16 pos panning effect (14-8). This is a
    non-standard protracker effect. So, could not test my codes. Not sure if
    it really works. Also, did is_mod_playing() function to test the mod
    is playing anything.

27  Did pause_mod(), resume_mod() and is_mod_paused() function. Thought
    I found a bug in mod_interrupt() function in player.c. The player exits
    with SIGSEGV when a music playing in unloop mode ends without windows
    running. It has no problem when windows is running. After 2 hours of
    wondering, found the ass code. Jgmod.c is the problem. Jgmod.c is
    accessing 'of' pointer even it is pointing to NULL. Fixed the problem. If
    you are reading this, my advice. If you are messing with pointers, don't
    test your program under windows. It will usually appear to be NORMAL.
    Shitty win95.


March 98
--------
1   Change next_pattern() and prev_pattern() to next_mod_track() and
    prev_mod_track() respectively. Fixed a bug in is_mod_playing() and
    resume_mod(). Added goto_mod_track() and finished destroy_mod().
    Change readme.txt and my jgmod homepage contents.

2   Library uploaded. Fixed a pattern break bug. I was not reading the
    documentation properly. Corrected the set pan effect.

3   Redo the s3m loader because I forgot most of the s3m loader codes I did
    2 months ago. 

4   Continue the s3m loader. Able to load 8 bit and 16 bit samples. Only
    tested with 8bit unsigned samples. Fixed a sample offset bug. Tested the
    library with cpp program again. This time it works. This is puzzling.

5   Fixed the 16bit samples loader. Instrument lenght given in S3M is in
    samples. I thought it is in bytes. Change the way notes are stored.
    Previously notes are stored in a single huge array. Now, store notes
    in arrays according to their pattern. I change it because it would be
    easier this way when dealing with XM.

6   Continue the s3m loader. Able to detect the no of channels accurately
    but slow.

9   Busying downloading gcc280, gpp280 and lgp280. Fixed some compile
    problems with gcc280.

10  Added restart position support (only in loop mode). Did a faster channels
    detection for s3m but less accurate. One of partitions compressed with
    drivespace 3 in my hard disk had gone bad. Fixed the problem with
    Scandisk. I store most of my mod musics in that partition. I had
    100-110Mb of mod. After fixing, I have 13Mb left !!!! Now I have very
    less mod for testing. I have do redownload those musics again. *sigh*
    Never ever use drivespace even if you are short of space. 
    
11  Added the option to disable or enable protracker with 15 instruments.   
    Continue the s3m loader. Able to set panning.

13  Just got myself a modem. Busying downloading some of the musics which
    lost earlier. 

15  Able to load all of the S3M patterns. Just need to convert it to a common
    format (format used by protracker).

16  Converted the S3M frequency to hertz. Now JGMOD can play S3M
    basically. Yahoo !!! S3M command A (set tempo), B (pattern jump),
    and C (pattern break) supported.

17  S3M volume channel done. Command O (set sample offset), T (set bpm).
    Command D (volume slide) partially supported.
    
18  Continue coding command D. Fixed a restart position bug. S8 (set 16
    position panning) command supported.

19  SD (delay sample), X (set pan), G (slide to note), note off supported.
    Finished coding the command D.

20  S2 (finetune), SC (cut sample) supported. Delay sample had been wrongly
    converted. Fixed. Library uploaded.

21  Fixed a problem with s3m fine volume slide up. V (global volume)
    supported.

    Found the cause of the problem when playing the last track of 'Existing
    by Rick Rippon'. The header reported that there are total of 41 patterns
    but actually 42 patterns. Stupid S3M. Fixed. 

    Now this is the stupidest mistake I ever had. I left some of the SAMPLE
    structure pointing to a null sample data in the s3m loader. When Allegro
    plays this sample with volume greater than 0, it will exit will SIGSEV
    error. This cause Catch that goblin!" to exit with SIGSEV error at track
    18-03 (or pattern 14). This song requested to play instrument 17 which is
    empty. It took me to 4 hours just to find this bug. 
    
22  Added set_mod_volume() and get_mod_volume(). Pattern delay(14-14)
    supported. Command E(portamento down) and F(portamento up) implemented
    except the extra fine slide part. I don't know how. My sliding is
    somewhat different from what I heard from other players. Don't know
    why. Library uploaded.

26  Downloaded wip (25 March). A stereo variable is added to the sample
    structure which causes only MOD to play incorrectly. Fixed.

29  Lock the appropriate function and variables and JGMOD structure. Fixed a
    bug in destroy_mod. Calling play_mod after calling destroy_mod will not
    play the song correctly.

30  Implemented extra fine slide for E(portamento down) and F(portamento up).

31  Improved the vibrato effect and command H(vibrato) is supported. 
    Fixed finetune for s3m. SA (stereo control) and Set Vibrato
    waveform (mod and s3m) supported.


April 98
--------
1   XM and IT supported !!!!                                                        You have been fooled... haha.

3   Pattern loop is now supported for MOD and S3M. 

4   Command I(Tremor), K(vibrato+volume slide) and L(porta to note+volume
    slide) are now supported. Now I found why my pitch sliding is different
    from other players. When converting herz to period, I should have take
    c2spd into account but I didn't. Fixed by changing quite a number of
    lines except set finetune. I will do it another day because it is
    3.50AM now. I hope no new bug occured from this modification.

5   Finished finetune. Implemented command U (fine vibrato).
    Library uploaded

6   Did codes for cubic player note dots like for JGMOD.EXE.

9   Some minor changes to player.c and Arpeggio supported (for MOD and S3M).

10  Fixed a bug in portamento to note.

11  Player.c which is 32kb takes a long time to compile (about 8 sec). So I
    started splitting player.c to two files. Player.c and player2.c. 

15  Changed the priority of all insturments sample from a default of 255 to
    192 because of the Allegro wip April 12 version. Added JGMOD_DATE just
    like allegro. Done get_jgmod_sample() to get mod instruments. Done
    ex1.c, ex2.c and ex3.c, which is examples on how to program with JGMOD.

17  Fixed a bug in s3m loader. Now it can handle buggy note values.
    Implemented s3m apreggio (command J).

21  Tremolo and set tremolo waveform supported for MOD and S3M.

22  Restarted the XM loader. Now able to load all the patterns but I still
    haven't convert it into something JGMOD understand.

25  Added remove_mod() function.

26  Added the ability to speed up or slow down the music by using
    set_mod_speed(). toggle_pause_mode() done.

28  Library uploaded

30  XM header might report the wrong number of patterns. Now my XM loader
    can deal with this problem.


MAY 1998
--------
1   Continue the XM loader. Able to load instrument header

2   Continue the XM loader. Able to detect number of samples, load
    all the sample headers and load all the delta samples (8bit and 16bit).

3   Now, my library can play XM basically. There is still a lot to do before
    it sounds acceptable. Implemented command B(position jump),
    D(pattern break), 9(sample offset), EE(pattern delay), F(set tempo/BPM),
    C (set volume), G(set global volume), E6(loop pattern), E9(retrig note),
    ED(note delay), EC(note cut)    .

4   XM might request to play nonexistant instrument or sample which will
    cause a SIGSEV error. I also converted the sample number for notes in
    all instruments wrongly. This causes JGMOD to play incorrect sample.
    All fixed.

5   Fine volume slide down and up supported. Plays incorrect sample again.
    Fixed.

6   Fixed some bugs in the loader. Now it can load "Dreams 2" without
    SIGSEV error but the loader still could not load it correctly.

7   Command 8(Set Pan), 7(tremolo), E7(set tremolo waveform),
    1(pitch slide up), 2(pitch slide down) supported. 3(Slide to note)
    is also done but somewhat different from what I heard from Cubic player.

23  Fixed portamento to note bug for XM. Command 5(slide to note + volume
    slide) and 6 (vibrato + volume slide) done.

24  Command 4 (vibrato), E4(vibrato waveform), E1(fine porta up),
    E2 (fine porta down), H(global volume slide) implemented.
    
    Found out that I did not implement E1(fine porta up) and E2(fine porta
    down) for MOD. How the hell did I miss those ? Anyway, both has been
    implemented.

26  Changed all malloc() calls in loaders to calloc().

28  Implemented command K(key off) and note off as volume ramp for 200
    milisec temporary. This make more music sound much better.

39  Set vibrato speed and do vibrato for volume column done.

30  Player.c reaches 32Kb again while player2.c is at 19Kb. Transfered some
    codes to player2.c. Did panning slide left and right for volume column.

31  Implemented command P(panning slide)


JUNE 1998
---------
4   get_jgmod_sample_no() returns some invalid sample number which caused
    jgmod to access invalid memory. The funniest thing is that sometimes
    JGMOD does not exit immediately with SIGSEV when accesing the invalid
    memory. This causes note2hz() calculate frequency with wrong value of
    sample transpose and then exit with Floating Point Exception. I hope this
    will fix all SIGSEV problems I face some of the time with XM.
    
5   I have decided to redo some of the codes for the mod interrupt. Whatever
    changes to the volume, etc... are updated immediately. Now I wish to
    do whatever changes only at the end of the interrupt. This will make
    me easier for me to code and maintain.

6   Got the player to play basically. I am going to concentrate on MOD
    commands first. S3M and XM will be later.

    Command 1, 2, 4, 6, 10, 11, 12, 13, 15, E1, E2, E6, E10, E11, E14 done.

7   Fixed a bug with extended commands which are global commands. Command
    E4, E5, E7, E8, E9, ED, 0, 3, 5, 7, 8, 9 done. Phew ! All MOD effects are
    done. Now I will concentrate on S3M.
    
    S3M command A, B, C, D, E, F, G, H, I, J, K, L, O, R, T, U done.

8   Fixed a bug in MOD arpeggio and slide to note. Continue doing those
    S3M effects. Done all except Multi Retrig.

9   Done Multi Retrig for S3M. Now, JGMOD supports all standard S3M effects.
    I will concentrate on XM. Supports XM basically. XM effect
    3-6, 8, A-D, EA, EB, F, G done.

10  Fixed a bug with porta to note and sample offset. On June 4, I fixed a
    bug which may cause SIGSEV or Floating exception. I missed a section of
    code which could also returns an invalid sample number. Fixed.

13  Implemented more effects. All volume commands are done.

14  Implemented command X1 (extra fine porta up), X2 (extra fine porta down),
    R (Multi retrig note) and T (tremor). All XM commands are supported now
    except K (Key off) and L (set envelope position).

    set_mod_speed() no longer increse the pitch but only the speed.
    Done set_mod_pitch() to increase the pitch

    Uploaded JGMOD to my homepage.

17  XM official document stated that envelops are 48 bytes long when actually
    it is 24 words long. This leads me to think there are 24 envelope
    positions when there are actually 12 positions only. Changed.

28  Found out that XM (amiga frequency) cannot play instruments with negative
    finetune in B note (any octave) properly. This is because that finetune
    is not converted properly when loading XM which I think it causes JGMOD
    to access illegal memory.

    There is another illegal memory access in parse_new_note(). Both fixed


JULY 1998
---------
7   Added some codes to simulate key off (or note off)

11  Added some codes for initializing envelopes. Converting XM linear period
    to frequency no longer uses pow() and no longer cause some programs which
    uses FPU to crash with floating point exception (I think).

15  More initializing envelopes today. Normal and sustain volume envelope
    supported but there is a bug in sustain volume envelope.

16  Fixed the yesterday sustain volume bug and done loop volume envelope.
    Key off is properly supported (I think)

22  Improve the documentation just like what Linus Frost suggested. I hope
    the improvement is good enough. Library Uploaded

26  Converted all 16bit samples in XM to 8bit samples upon loading. This is
    to fix those a SIGSEV error that I encountered a long time ago. This
    conversion has no effect on the sound output quality at all since allegro
    will automatically convert 16bit to 8bit samples before mixing.

27  Added the option to force to 8bit samples in XM modules by using define
    in load_xm.c.

31  Fix XM portamento to note + volume slide bug. 
    JGMOD will restart the song if prev_mod_track() or goto_mod_track() to
    track 0.


AUGUST 1998
-----------
6   Fixed a bug that exits in volume envelope when sustain point is at tick 0

7   Panning Envelope supported. JGMOD has problem playing envelope with only
    one point specified. Instruments that should play volume envelope but
    no point specified in the envelope should not be played at all. Fixed

    Began coding on the conversion utility, JGM. Able to save JGM basic
    JGM header

8   Begain coding on the JGM loader. Able to load basic JGM header. Also the
    conversion utility able to save instrument header.

9   Able to load instrument header.

10  Able to save and load sample header.

11  Some adjustment to the sample header.

17  Playing an instrument that doesn't use volume envelope, with key off
    (or note off) specified is equilavent to note cut. Fixed.

20  Fixed 2 bugs in portamento to note. Library uploaded.


September 1998
--------------
9   Fixed a bug in the XM loader.

16  Correct the tempo effect. A tempo 0 command will be ignored.

30  XM arpeggio and XM set envelope position supported. The S3M loader will
    automatically convert all 16bit samples to 8bit samples to avoid SIGSEV
    error when playing the S3M as well.


October 1998
------------
3   I found out that Unreal uses S3M and IT some time ago, thanks to
    www.gamasutra.com. I bought the game today and found out it is really
    true except that the file has an extra header for Unreal in the beginning
    (about 145 bytes) of the file and with an extension of umx.

    Detecting the umx file is very simple but finding the size of the of
    the unreal header file is not very direct since the size of the header
    varies for all umx. There are 2 variables in the file which are very
    close to the size of the file but it doesn't give me any clue of
    detecting the size of the header. I found out a way of detecting the size
    of the header and thus able to find the first byte of the S3M header. Now
    loading the S3M from UMX is not a problem for JGMOD.

    Library uploaded

9   Corrected some errors in unreal.txt about the channels used for some
    musics.

29  JGMOD are able to handle musics more than 32 channels. Also loading XM
    and MOD more than 32 channels have no problem with JGMOD.
    JGMOD.EXE can scroll up or down if more than 33 channels are used, that
    is if the screen is not big enough to display all those 33 channels.

30  The s3m loader are able to load s3m up to 32 channels. Previously, only
    up to 16 channels.


November 1998
-------------
6   Instrument without note specified should be ignored.

7   Decided how the patterns should be saved. Coded the JGM converter and
    loader. Fix a lot of bug in the converted. HEck... Converter is harder
    to code than the loader. Fixed quite a lot of bugs in the converter.
    Keep me busy for about 4 hours debugging it. I think it is stable now.
    
    Now JGMOD can load JGM can Allegro datafile. Library Uploaded

8   Fix another bug in JGM. Now it can store patterns greater than 255 rows.
    Library uploaded again.

    Found another bug... damn... JGM has trouble storing extended command
    some extended command. Fixed. Shouldn't have upload it so fast.

27  Ported my library to Winallegro using VC5. WinJGMOD is here !!!

28  More testing for WinJGMOD. Winallegro can't play volume correctly. BAD.
    JGMOD uploaded.


December 1998
-------------
10  Started doing Allegro grabber plugins. Able to grab MODs.

11  Added function register_datafile_jgm() to load jgm.

14  Finished doing Allegro grabber plugins.

22  When instrument/sample and a portamento to note is specified, the volume
    should be initialize to the default volume specified within the
    instrument/sample.

    Added a function in JGMOD.EXE to redraw the screen by pressing the
    tilde key. This is only important if the screen is messed up when
    switching to windows and back to dos.

23  Fix a bug on yesterday's bug fix which only happens on XM. Added
    compress option thru makefile.

    Library uploaded.

29  Didn't close a file correctly in the detect_jgm() function. Fixed.


January 1999
------------
24  Upgraded to DJGPP 2.02. JGMOD have trouble opening certain MOD files.
    Don't really know why.

26. Found a bug in the jgmod_fseek() function. Now, JGMOD should not have
    troubles opening the MOD files. The grabber plugin and my Jtris game now
    works fine. Phew.

    Library uploaded.

31  Started doing my own file routines, for portability reasons.


February 1999
-------------
24  Added a variable, enable_lasttrk_loop to ignore the position jump in the
    last track of a song to avoid infinity loop.


April 1999
----------
20  Gosh. Haven't been updating JGMOD for a long long time. Seems like an eon.
    Been spending too much time playing games. Just finished the jgmod_error.
    Problems encountered by JGMOD will be reported in jgmod_error.

22  Read a little of IT file format doc. Finished the detection for IT and
    started doing the IT loader.

24  Decided that the JGMOD examples should be part of JGMOD distribution.
    Library uploaded.


May 1999
--------
13  Added keys to reduce and increase range of notes displayed 
    (using F5 & F6), and to change the relative positions of the notes
    (using F7 & F8) in JGMOD.EXE.

16  Added get_mod_info() to get some information about MODs. Added EX7
    to examples to show how to use get_mod_info().


June 1999
---------
6   Fixed a pattern break and position jump bug. I noticed this bug about a
    year ago and refused to fix it because it could cause infinity loop. The
    pattern break bug will be handled correctly if the mod song is in loop
    mode. Otherwise, it will not. JGMOD could play backward.s3m properly now.

    I tought I found out that JGMOD plays instruments at tick 1. Should be at
    tick 0. Actually, JGMOD plays instruments at tick 0. Then, immediatelly,
    the tick is increased by one. This made it looks JGMOD plays instruments
    at tick 1. It really fooled me for a while, after not touching my core
    JGMOD codes for such a long time. JGMOD codes looks quite alien to me
    now. Changed the logic of the core. Increase, then play.

    set_mod_pitch() no longer allow 0 pitch 

21  Fixed the arpeggio bug, I think.
    Library Uploaded.


November 1999
-------------
26  Bought Unreal Tournament(UT). What an exciting game.

29  Just received a JGMOD version of Linux from George Foot. Hm.. Maybe I
    will call it LnxJGMOD? or LinuxJGMOD? Duh.. I have no idea.

    The detection for S3M in Unreal has been changed to detect S3M more
    reliably than before. JGMOD is now able to detect and load UT's S3M and
    XM (both in UMX). Also added codes to detect Unreal and UT Impulse
    Tracker(IT) files. JGMOD should not have any problems at all when
    detecting any UMX modules.


January 2000
------------
5   Changed the directory structure. Library and object files are no longer
    kept in the same directory but in it's own directory just like Allegro.

6   Finished porting to Linux generally.

7.  Forgotten to close some files in get_mod_info(). Fixed.
    get_mod_info() now able to detect and get information about Unreal
    Tournament XM and IT files.

    JGMOD plugin now supports Unreal and Unreal Tournament UMX (S3M and XM)
    files. More information is displayed about MOD in the plugin. 
    
9   Wrote plugins.lnx on now to install the plugins. The previous
    plugins.txt has been renamed to plugins.dj

    Library uploaded.

10  Re-porting JGMOD to MSVC. Got the jgmod.exe and jgm.exe compiled. jgm.exe
    crashes when converting MOD to JGM.

11  Bought and installed Linux Mandrake 6.1. Excellent KDE desktop.
    Continue porting JGMOD to MSVC. Found JGMOD have linking problems under
    Linux with Allegro 3.9.30. Fixed.

12  Finished porting to MSVC. All of the files compiled successfully with
    Allegro 3.9.30. Like the previous version of WinJGMOD, JGMOD still cannot
    play MODs correctly and sometimes, it quits playing MOD. There is
    something wrong with the timer and allegro doesn't support bidirectional
    and looping samples. 

13  Fixed some linking problem again under Linux.

16  Library Uploaded.

17  Fixed some problem with the timer halting in MSVC. Linux shared file
    library is supported.
    Library Uploaded.


September 2000
--------------
11  Having trouble running DJGPP utilities on Windows ME. DJGPP have trouble
    renaming files. ARgh.... Have to install Win95B on a seperate partition,
    with the help of BootMagic. HD space is very very limited now. ARGhh...

    Tested Henrik Stokseth MingW32 patch for JGMOD. Thanks man. Everything
    works fine.

12  Tested Robin Burrows wdsound.c which improves Win32 mixing routine for
    JGMOD on Mingw32. MSVC can't compile the wdsound.c. This is exiciting.
    All the Mod musics sound correctly on DirectX platform. YEAH.

    Library Uploaded.


April 2001
----------
18  Applied the diff file by Angelo Mottola to support BEOS 5.

    Library Uploaded.
