TinyPTC x11 ChangeLog + Bug(Fixes) list.

*** Note ***

English isn't my native language. So, some grammar/spelling errors might have
made their way into this document. You've been warned.

Alessandro Gatti <a.gatti@tiscali.it>

*** **** ***

20021208 0.7.3

Fixed a typo in the python source and updated the documentation.

20021207 0.7.2

The perl script for the assembler include generation was replaced with a
plain shell script in order to overcome the remote possibility of perl
not being installed on the system TinyPTC gets built on. Now TinyPTC can
be built as a python extension by using the excellent SWIG package, just
use the 'python' target of the makefile and look at python_test.py.

20020706 0.7.1

Kevin Nickels (knickels@users.sourceforge.net) sent a patch for the BGR888
converter that fixes the colour order on little endian platforms in the xlib
target. A minor error in the documentation has been corrected, and the MMX
assembly include file builder perl script has been slightly cleaned up.

20020517 0.7

Fred Howell wrote the XVideo target handling code and was kind enough to send
it in order to have that included in the major next release of TinyPTC/x11. It
has been tested both on NVidia-based cards with their binary-only drivers, on
ATI-based cards, and with Matrox-based ones, all of them under XFree 4.x. The
XVideo target features realtime scaling on the windowed output (fullscreen
support will be added on a later stage...), with a lightning fast MMX converter
from RGB to YV12. Make sure to check it out!

20011103 0.6.1

Version 0.6.0 was entirely done during one long coding night, so a few bugs
were present. Anyway, I made the C functions callable by C++ as well (tested
with g++), the example configuration file now has some enabled #defines that
let the example work flawlessly (otherwise it won't run if you haven't got
low-res modes defined in your X configuration. I've got them in, that's why I
didn't notice that...). Then, g++ complained on a few targets for some
duplicated variable definition. That's funny, gcc didn't complain. By the way,
I also tested the whole package with GCC 3.0.1, and it works fine. So, this is
mostly a "cleaned up" version.

20011102 0.6.0

After nine months, I found the time to work again on TinyPTC/X11. Big changes
this time... First of all, now TinyPTC/X11 is released under the LGPL license,
then I (finally) added DGA 2.0 support, although its heuristic stuff for best
videodepth choice isn't tested at all, since my hardware supports 32bpp
natively (if it doesn't work, please tell me, so I can attempt to fix the
problem). Then, I splitted all the various targets in a source/header pair of
their own, in order to improve manteniability and easy new targets addition.
Currently you compile and link all the various targets even if you use only
one of them but don't worry, because with some kind of #define magic and a fair
job done by strip, the executable size won't grow at all. The makefile has been
reworked in order to use automated dependencies and for generate automatically
the NASM includefile for mmx.s from mmx.h. Then, I had to change a variable
name in the supplied example due to clashes with a function with the same name.
Oh, I also added DBE support, if you need it. Of course the documentation has
been updated and the makefile now creates a static library called libtinyptc.a
that can be included in your own programs without having to recompile tinyptc
all the time. Have fun!.

20010206 0.5.1

The makefile now contains a different optimization flag, now the executable is
a bit smaller (at least on my box). I updated the documentation, too (I forgot
to add the __PTC_MEMCPY__ #define). Added a cleanup callback function, in order
to let it be a bit more programmer-friendly. Now you can use PTC_SUCCESS and
PTC_FAILURE in order to check the results of the TinyPTC functions you call.
Documentation now contains NetBSD compilation workaround.

20001027 0.5

The mmx converters were added and the makefile was cleaned up in order to be
a bit more flexible and to support the external assembler code. I left out a
debug fprintf somewhere in the DGA code, now it's just history. The
documentation was updated, too. Nothing more to say, over and out.

20001306 0.4

Bugs are calling... And we're supposed to fix them. There was a glitch in the
32bpp->15bpp conversion code, thanks to tolemaC/Elerium Core for the bug
report. Now the 24bpp conversion modules should work for big-endian machines as
well. Try them if you've got such a system and then tell us.

20000531 0.3.1

Sometimes, fixing a bug, brings another one in. Well, now the only thing that
worked was the 32bpp code. Strange, isn't it ? ;) Anyway, this is fixed now.

20000527 0.3

TinyPTC was developed again after a little pause. Great news this time... I
added color conversion and it seems to work quite well, too. It's rather slow
but I think that some compiler optimizations can speed it up a bit as well. If
it's still too slow for you, feel free to write your own converters and submit
them to Gaffer =) As an added bonus, the 24-32 bpp dilemma seems to be
disappeared... Maybe it's fixed, try and see. Tell me if there's something that
doesn't work.

20000508 0.2

TinyPTC was rewritten from scratch (literally), now the sources are easier to
read and I threw some comment in, too. The XShm support is finished and seems
to work. Tell me if I'm wrong about that. The DGA part wasn't working properly,
it blarfed on the fullscreen-emulation blit code, now it works again. I also
added a 'best-fit' videomode switch logic for fullscreen-emulation code.

20000505 0.1.1

Some kind of preliminary XShm support is now included. There wasn't an UID
check for DGA mode (it needs Root access in order to access video memory
directly). I forgot to include it when I joined the two x11 sourcetrees
together (windowed and DGA). The bitdepth is incorrectly reported as 24 even on
32bpp servers. Strangely enough, my server correctly reports 32bpp.
The pitch value calculation method is wrong. On some cards the assumption 
pitch=width*bpp isn't true (TNT cards for example but not Banshees). Damn, I'm 
so lame... =) I was calculating it with the logical screenwidth of the 
resulting viewport instead of calculating it with the physical screenwidth 
returned in the videomode info.	The makefile was modified in order to include a
check to "tinyptc.h" during the 'make' operation. 

20000503 0.1

Initial release. Supports X11 Windowed and X11 DGA access. No color 
conversions, just a bare bones emulation of the windows version.
