Sat Jan  1 11:56:02 2000  Albrecht Kleine  <kleine@ak.sax.de>
	-tya*.*
	 some general clean up for release;
	 removed some reasons for compiler warning errors: 
	 added type casts, renamed FSCALE to F_SCALE (due FreeBSD)

	file #300 (final)
--------------------------------------------------------------------
Mon Dec 20 18:49:01 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c
	 * made PASS2 independent of `define COMBINEOP' (intro 
	   was in #253)
	 * clean up, removed some `#if 0' stuff
	-tyaexc.c
	 bugfixed wrong message: was value instead of index
	 in CompTriggerArrOutOfBoundsException2()
	-conf*,make*
	 added some changes for working with blackdown jdk1.1.8v1

	file #299
--------------------------------------------------------------------
Sat Dec 18 14:41:52 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c
	 repaired bug in self recursive calls (this bug came in #288, 
	 that has broken running Mo Dejong's Jacl-1.2.3 test suite
	 for 3 months)
	-tyaexc.c,tyarechelp.c,tyarecode.c
	 removed array optimizer (cmp #293) in pass2 due
	 index expected in wrong register if arrayoutofboundsexception
	 occured
	-tya.h,tyarecode.h,tyarechelp.c
	 using symbolic const for ADD,SUB in pass2() etc.

	file #298
--------------------------------------------------------------------
Wed Dec 15 20:57:54 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarechelp.c	
	 added comparisation with constants to Pass2()
	-tyarecode.c
	 added preparation for pass #2 for opcodes iconst_3,_4,_5

	file #297
--------------------------------------------------------------------
Sun Dec 12 21:00:24 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarechelp.c
	 added some is_dest-checks in pass2 where neccessary
	 in pass2()
	-tya.c
	 added hooks for CompilerRegisterNatives(),
	 CompilerUnregisterNatives()
	 
	file #296	
--------------------------------------------------------------------	
Sat Dec 11 09:20:40 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c
	 repaired bugfix in opcode combination optimizer DF_const2
	 (compare workaround in #269)

	file #295
---------------------------------------------------------------------
Wed Dec  8 21:52:58 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarechelp.c,tyarecode.c,tya.h
	 continued #293, 
	 defined constants for basic operation types in tya.h,
	 extended iptrans structure

	file #294
---------------------------------------------------------------------
Tue Dec  7 19:40:00 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarechelp.c
	 extended second pass for some other usual cases
	 (for details look diff output)

	file #293
---------------------------------------------------------------------
Mon Dec  6 20:53:21 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarechelp.c
	 * made Comp_POP_StoreLocalVar() more general for 
	   beeing every possible register as data source
	 * moved Pass2() and LoopFinder() to here

	file #292
---------------------------------------------------------------------
Sun Dec  5 19:24:20 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c,tya.c
	 optimisation of an opcode _triple_ for the 1st time,
	 using a 2nd pass
	-tyarechelp.c
	 made Comp_PUSH_LocalVarToReg() more general for 
	 using every possible register as destination

	file #291
---------------------------------------------------------------------
Sat Dec  4 17:47:31 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.h,tyarecode.c,tyarechelp.c
	 continued COMBINEOP optimization for opcode pairs like
	 *load/if* using NEXT_TESTS_WORD macro

	file #290
---------------------------------------------------------------------
Mon Nov 29 21:25:40 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c,tya.h,tyarecode.c
	 added some loop finder algorithm
	-tyarecode.c
	 backpatch struct: currcnt now stores ip value (old was ip+1)

	file #289
----------------------------------------------------------------------
Sun Oct 31 17:53:57 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya* 
	 cont. using #284 ,  minor modifications tyarecode.c
	 for faster invocation in certain cases

	file #287,#288
----------------------------------------------------------------------
Tue Oct 26 18:33:38 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya*
	 two experimental studies in invocation without setting any 
	 java_frame data, instead maintaining frame data while
	 in CompiledFramePrev_hook. Later completely removed,
	 not yet suitable, but has potential.

	file #285,#286 (not for general usage)
-----------------------------------------------------------------------
Fri Oct  1 18:29:44 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c,tyarecode.c,tyaconfig.h (affects for1.1 only)
	 continued #283 reducing number of useless tests
	 whether clinit has already been run or not by 
	 maintaining another flag (CLINITRUN)

	file #284
	***  build as TYA1.5v2, but no more public release  ***
----------------------------------------------------------------------
Thu Sep 30 18:11:18 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c,tyarecode.c,tyaconfig.h (affects for1.1 only)
	 added workaround code simulating class initializing
	 process just like TYA does on jdk12: this let us pass 
	 Artur's InitTest.java also on 1.1 platform :)
	 (but still suboptimal)

	file #283
-----------------------------------------------------------------------
Sun Sep 26 10:49:15 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaruntime.c
	 bugfix in RTGetIMeth (intro in #276), affects only versions
	 compiled with #undef USEASM
	-tyarecode.c
	 * replaced some CB(),CL() with CompTriggerLinkageError()
	 * fixed trigger of NullPointerException on _empty_ void
	   methods on JDK1.2 (using Shudo's demo file, cmp #274)

	file #282
	******  released as TYA1.5  *******
------------------------------------------------------------------------
Sat Sep 25 20:25:19 1999  Suresh Srinivas <ssuresh@cthulhu.engr.sgi.com>
	-tya.c
	 no more using ``bswap'' inline assembler: tableswitch was 
	 broken (was affecting java/util/Properties) due to getint32
	-tyarecode.c, tyaruntime.c
	 re-activated CompTriggerOutOfMemoryException() calls
	 plus one extra for ``new'' opcode
	 but removed out_of_memory() call in MyArrayAlloc()
	 plus CompTriggerLinkageError call in ``new'' opcode
	-tyarecode.c
	 extended access flag test for ``new'' opcode
	-tyaexc.c
	 added CompTriggerLinkageError(): makes code more readable

Sat Sep 25 20:25:19 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaexc.c,tyarecode.c
	 moved EAX check from IncompatibleClassChangeError()
	 to caller's side for common interfacing with
	 Suresh Srinivas patch, added param check for NULL in
	 CreateExceptObject(), case 9: .... due different interfacing

	file #281 (merged with #280 & commented by A.K.)
----------------------------------------------------------------
Fri Sep 24 21:40:50 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c
	 fixed broken disabling of JIT Compiler, but it's
	 further not used by default: switch this feature
	 via #define IGNORE_DISABLE.  (Swing1.1 uses this.)
	-tya.c 
	 commented out ``internal test'' lines from #277
	-InitTest.java, *.class
	 added Artur's InitTest into TYA distribution.
	 Tnx to Artur Biesiadowski <abies@pg.gda.pl>
	 Once again: ONLY for 1.2 usage - maybe we should
	 consider 1.1 as outdated.

	file #280 [iz]
---------------------------------------------------------------
Wed Sep 22 17:29:00 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c,tyarecode.c
	 continued #277, making swing work again by controlling 
	 some needed CompileClass_Hook() calls via class->
	 reserved_for_jit field.
	 
	file #279
---------------------------------------------------------------
Tue Sep 21 20:33:00 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c,tya.h,tyarecode.c
	 simplified #276: no more additionaly parameters 
	 for GetBlock needed

	file #278
---------------------------------------------------------------
Mon Sep 20 20:38:44 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c,tyarecode,c
	 tried more compatible class resolving: breaks
	 some apps (swing) but passes Artur's InitTest.java (1998),
	 (experimental stuff: switch via #define EXT_COMPAT.
	 This feature is currently ONLY for JDK1.2 available!

	file #277
---------------------------------------------------------------
Sat Sep 18 10:33:26 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya*
	 added check for some kind of IncompatibleClassChangeError:
	 catching later changed interface implementor classes
	 (could be switched on/off by #define EXT_ERR)
	-tya.c
	 increased value of cinfo->resspace (compare bugfix #270:
	 more spaced needed for compiling catch frame)

	file #276
----------------------------------------------------------------
Thu Sep 16 20:47:09 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya*
	 added check for some kind of IllegalAccessError:
	 writing on final vars
	 (could be switched on/off by #define EXT_ERR)

	file #275
----------------------------------------------------------------
Tue Sep 14 21:49:33 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c
	 rewrote NullPointerException check for invocations
	 tnx for a whole serie of bug reports to Kazuyuki Shudo 
	 <shudoh@muraoka.info.waseda.ac.jp>

	file #274
----------------------------------------------------------------
Sun Sep 12 22:18:46 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c
	 bugfix signal handling signals != SIGFPE, SIGSEGV
	 tnx for bug report to Kazuyuki Shudo 
	 <shudoh@muraoka.info.waseda.ac.jp>

	file #273
----------------------------------------------------------------
Thu Sep  9 18:59:32 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarechelp.c
	 bugfix: missing stack pop in d2i,d2l if NaN
	 tnx report to Michael G. Ross <mgross@ai.mit.edu>

	file #272
-----------------------------------------------------------------
Tue Sep  7 17:17:10 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c
	 some more restricted INLINING due classloading
	 problems in Jdk1.2: for example if loading native code
	 after libtya. Now we pass Mo Dejong's Jacl-1.2.3
	 test suite on 1.2 too :) /jl.loadLibrary

	file #271
-----------------------------------------------------------------
Sun Sep  5 19:08:30 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya*
	 some cleanup
	 
	file #270
	******  released as TYA1.4v2  *******
----------------------------------------------------------------	
Sun Aug 29 20:37:03 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaexc.c
	 bugfix for exception/catch (if more than 
	 92 byte local variable space: was register reload
	 problem). 
	 Thanks bugreport Mo Dejong <dejong@cs.umn.edu>
	-config.h
	 removed from tgz archive

	file #270pre
----------------------------------------------------------------
Sat Aug 21 17:26:26 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c
	 bugfix in opcode combination optimizer DF_const2

	file #269
-----------------------------------------------------------------
Mon Jul  5 19:24:19 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-config*
	 removed older JDK support -all except mainstream
	 from blackdown.org and freebsd.org
	-tyarecode.c
	 bugfix in irem for a special x86 problem, cmp #266
	-tya.c, tya.h, tyarecode.c
	 use common GetBlock() also for new operator
	-calling InitClass()

	file #268
-----------------------------------------------------------------
Sun Jul  4 19:57:43 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c
	 experimental RESOLVE_FT up to 4 (for workaround only)
	-tyarecode.c
	 bugfix in running empty static ()V methods

	file #267
------------------------------------------------------------------
Sat Jul  3 18:07:41 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c
	 bugfix in idiv for a special x86 problem
	 tnx bug report Matt Welsh <mdw@cs.berkeley.edu>
	-tyaasm_freeBSD.S 
	 added forgotten underscore to EE
	 tnx patch Alexander V. Romanyuk and Nicolas Niclausse
	-configure.in
	 added option -green -Djava.compiler=NONE
	 tnx hint David.O.Martinez@bankamerica.com
	-tya.c
	 wrapped both now() calls with #ifndef JDK12

	file #266
-------------------------------------------------------------------
Sat Jun 19 21:45:25 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya*
	 some cleanup / removed some TYA verbose messages
	-tyaasm_freeBSD.S
	 recreated again another new version for a.out systems: 
	 !!untested!!

	file #265
	******  released as TYA1.4  *******
------------------------------------------------------------------
Sun Jun 13 10:45:23 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaruntime.c
	 reintroduced file from build #251 (w/o FAST_NATIVE102STYLE)
	 due some reported problems in #252
	-tyarecode.c: cleanup 

	file #264
------------------------------------------------------------------
Fri Jun 11 19:56:03 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c, tyarechelp.c, tya.h
	 * added another opcode combination optimizers 
	   indicated by DF_const0, DF_const1, DF_const2
	 * replaced/moved asm stack and call related opcodes
	   for better pairing
	 * optimized lload* java-ops by directly pushing 
	   data to stack avoiding register stalls,
	   changed Comp_PUSH_LocalVarToStack() void to int
	-tya.h	  
	 increased #define TYA_JSTACK value

	file #263
-------------------------------------------------------------------
Sun Jun  6 19:16:03 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c, tyaasm.S
	 continued experiments with pairing opcodes
	 
	file #262
-------------------------------------------------------------------
Sat Jun  5 13:10:53 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c,tyarechelp.c
	 * i2l opcode: replaced cdq with two pairing ops
	 * removed CCONST and optflag_2048 stuff (from #25x)
	-tyaexc.c,tyarecode.c
	 continued #260 for array load and store changing
	 CompTriggerArrOutOfBoundsException()

	file #261
-------------------------------------------------------------------
Fri Jun  4 18:40:43 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaasm.S,tyarecode.c,tyarechelp.c
	 * rewrote invocation to use more pairing and prevent AGIs

	file #260
--------------------------------------------------------------------
Mon May 31 19:43:02 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarechelp.c,tyarecode.c,tya.h
	 * use DF_DD opt for java.lang.Math natives
	 * reactivated compileNOJFRset optimizer for jdk12 too
	 * sqeezed one byte per entry in lookupswitch opcode
	 * easier handling write cache local variables except objects
	 * leave out some useless NullPtr checks in invocation opcode,
	   but NOT in getfield,putfield

	file #259
---------------------------------------------------------------------
Sun May 30 20:47:29 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c,tyarecode.c
	 bugfix [ipcnt].isdest now stored at 2nd op (old was at 1st op,
	 so have to init cinfo->iptab with 0 before compiling)
	 (is important for checkopt alert)
	-tyarecode.c
	 * exp. coding a pair istore/iload (later rejected)
	 * exp. coding a pair aload_0/invocation (later rejected)
	 * exp. coding using ``setge al'' in Comp87Vergleich()
	 * opcode 0x92 (int2char) and eax,ffff with movzx eax,ax
	 * opcode 0x93 (int2short) cwde with movsx eax,ax
	 * added shortcut for lcmp paired with ifne or ifeq
	 * added code for pair of fcmpl/fcmpg with float dest
	 * squeezed some byte form tableswitch
	-tya.h 
	 * wrote some doc about new opcode pair coding

	file #258
--------------------------------------------------------------------
Mon May 24 17:30:26 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c
	 continued COMBINEOP optimization for opcode pairs like
	 ``dadd AND dastore'' etc.
	-tyaasm_FreeBSD.S
	 after confirmation it works ok, I removed warning 
	 notice for compilation problems

	file #254..#257 (some steps, finished in #258)
--------------------------------------------------------------------
Sat May 22 17:50:27 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c
	 more relaxed INLINE decision
	-tyarechelp.c
	 initial COMBINEOP optimization for opcode pairs like
	 ``dcmpl AND ifle'' etc.
	-tyarecode.c
	 * replaced 2 byte JMP with JMP1 return opcodes
	 * added exp opt for Math.abs(D)D

	file #253
---------------------------------------------------------------------
Sun May 16 16:25:55 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaruntime.c
	 another complete rewrite of JNI invocation
	 by using dynolink() again (was already used up to #214)
	-tyaasm.S
	 bugfix in FastInvNoCheck64() added forgotten line in #250
	 ``movl 20(%edx),%edx''
	-tyaasm_FreeBSD.S
	 recreated a new version for a.out systems: !!untested!!

	 file #252
	 ******  released as TYA1.3v2  *******
---------------------------------------------------------------------
Fri May 14 19:36:40 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaruntime.c
	 JNI: continued #249 / simplified initial ``if'- condition
	 define via #ifdef EXP
	-COPYRIGHT: changed headline notice from
	 ``not free software'' to ``not public domain software''	 

	file #251
---------------------------------------------------------------------
Thu May 13 19:00:24 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarechelp.c
	 bugfix ftoi,dtol,... etc. numeric overflow handling
	 TNX for bug hunting to
	 Wolfgang Muees <wolfgang@wmsickte.escape.de>
	 Good job! Makes swingset1.1 demo HTML pane work.
	-tya*.c
	 removed some old JDK1.0.2 stuff: 
	 in fact build #181 was the last TYA running on JDK1.0.2
	-tyaruntime.c, tyaasm.S
	 made javastack overflow test more sensitive

	file #250 [nb2]
--------------------------------------------------------------------
Mon May  3 19:08:14 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaruntime.c
	 rewrote JNI invocation code directly using
	 their JNI (and NMI) invokers if possible
	 
	file #248,#249 (archived some experimental sub builds)
--------------------------------------------------------------------
Fri Apr 30 21:10:34 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.h, tyarecode.c, tyarechelp.c
	 added another approach for java.langMath:
	 function calls w/o much overhead
	 [ use via #define FAST_FPARITH ]

	file #247
--------------------------------------------------------------------
Sat Apr 24 20:23:00 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.h, tyarecode.c, tyarechelp.c
	 added some experimental shortcuts for
	 inline Math.sin, Math.cos, Math.sqrt (--do not use,
	 except you know why / good for some Fourier Analysis)
	 [ use via #define INLINE_FPARITM ]
	-tyarecode.c
	 squeezed some bytes in opcode 0x09,0x0A,0x0E

	file #246
--------------------------------------------------------------------
Thu Apr 22 20:54:15 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-some clean up
	-switch off EXCEPTION_BY_SIGNAL for JDK12

	file #245
	******  released as TYA1.3v1  *******
--------------------------------------------------------------------
Fri Apr  9 21:44:36 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c
	 * bugix in drem and frem opcode: was missing floating point
	   stack pop / GOOD for java2D textures stuff
	 * activated compileNOJFRset optimizer for jdk12 too

	file #244
---------------------------------------------------------------------
Sat Apr  3 17:01:02 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c,tyaruntime.c,tyaasm.S
	 simplified getting space for JavaFrame (plus some cleanup)
	-tya*.h
	 moved RESOLVE_FT and INIT0 to tya.h
	 
Sat Apr  3 17:01:02 1999  Steve Price <sprice@hiwaay.net>
	tya.S: patch for running on FreeBSD/ELF systems

	file #243
---------------------------------------------------------------------
Fri Apr  2 10:42:43 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c
	 reorganized JITCompileMethod(): getting java-stack-space
	-tyarecode.c
	 squeezed some bytes in caload by using a ``movzwl'' statement

	file #242
----------------------------------------------------------------------
Wed Mar 31 19:45:31 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-fixed an older bug in FAST invocation of NON-compiled methods:
	 (common for jdk1.1.7 and 1.2)
	  * tyarecode.c: using one common way for getting local var
	    space on stack (returned by FastInvCheck* in %edx reg.
	    [ no more CW(SUB_SP_BYTE);CB(compileLOCALspace); ]
	  * tyaruntime.c: changed type of FastInvCheck* to long long
	    because of additional returning on mbpcalled->nlocals*4 
	    in %edx
	  * tyaasm.s: just like in tyaruntime.c versions the asm-versions
	    of FastInvCheck are returning local stack space in %edx
	-added macro DEF_CreateNewJavaStack for different function
	 arguments in tyaruntime.c

	file #241
----------------------------------------------------------------------
Mon Mar 29 17:31:16 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c
	 * in JITCompileMethod() : introduced better way calling
	   CompiliereMethode() / solves the gc problem mentioned
	   in #239
	 * added two another hooks: CompiledFramePrev_hook() and
	   CompiledFrameID_hook()

	file #240
----------------------------------------------------------------------
Fri Mar 26 18:18:53 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaasm.s
	 initial try on JavaStack overflow handler in asm part
	 running applet and some jfc jdk 1.2 demo stuff 
	 (needs using -Xms4M / -J-Xms4M switches because of gc)
	-tyaruntime.c
	 exclude useless #ifdef FAST_NATIVE102STYLE for JDK1.2 

	file #239
----------------------------------------------------------------------
Thu Mar 25 20:31:06 1999  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c
	-added JavaStack overflow handler in #4: CreateNewJavaStack

	file #238
----------------------------------------------------------------------
Thu Mar 25 18:16:00 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-removed superflous stuff in GetBlock(), added assertation
	
	file #237A
----------------------------------------------------------------------	
Wed Mar 24 20:39:13 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-an version TYA1.3pre for JDK1.2 running also some AWT stuff
	 (Iview etc.)
	-compileable both for JDK1.1.7 and JDK1.2 (Linux only tested)
	-bugfix in hook #4 for 64bit results
	
	file #237
----------------------------------------------------------------------	
Mon Mar 22 20:44:03 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-a special version for JDK1.2 running Sy demo stuff
	 under certain conditions (with some enabled options in 
	 tyaconfig.h)
	-adjusted value returning in CodeRunner64()
	-added include signal.h for EXCEPTION_BY_SIGNALS
	-bugfix do not treat short int w sparse bytes as int:
	 (mbp->locals now proc via some SHL/SHR_BX) *CARE*
	 
	file #236
----------------------------------------------------------------------	
Tue Mar 16 22:00:00 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-an initial hit for JDK1.2 running some trivial stuff,
	 but no special options possible (and jar unpacked)
	-switched to using traditional resolver function via macro
	-fixed another changed data offsets in data structures (*.S,*.c)
	 for example: ee->current_frame via #define CFR_OFF

	file #235
----------------------------------------------------------------------
Mon Mar 15 21:14:25 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-changes in javaframe setting (hook4#)
	 something like ((char*)JF)+=((xx+mb->yyyyyy)*SIS4)
	-added code for opcodes 0xD4 and 0xDA (with msg for op3 !=0)
	-fixed changed data offsets in data structures (*.S,*.c)
	 for example: ee->exceptionkind

	file #234 / not ready for run
----------------------------------------------------------------------
Sun Mar 14 19:34:03 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-experimental study using new resolve, self calling <clinit>
	 methods / suspended in #235 / for later continuation
	-adjusted value returning in CodeRunner64()
	-substituted hashID processing, used a new macro instead
	-removed *lockClass() calls, ditto getObjConstructorHash stuff
	 (1.2 only)
	-some adjustments in java_lang_Compiler_start for 
	 JITInterface6 data structure in jit.h (1.2 only)
	 
	file #233 / not ready for run
----------------------------------------------------------------------
Fri Mar  5 17:33:26 1999  Albrecht Kleine  <kleine@ak.sax.de>
	invoke interface: squeezed some bytes
	
	file #231, #232==final for 1.1 
	moved to archive as TYA1.2v5, but not released
----------------------------------------------------------------------
Wed Mar  3 20:46:07 1999  Albrecht Kleine  <kleine@ak.sax.de>
	astore others opcodes: squeezed some movs
	file #230

	aastore + exception squeezed
	file #229 

	#define INIT0: leave out some initializers with 0 (this came
	from jdk 1.0.2)
	
	file #228 
----------------------------------------------------------------------
Mon Feb 15 20:25:00 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya*
	 * replaced some more OR_regreg with TEST_regreg
	 * removed a NOP in opcode int2byte
	 * some optimisation: using registers now via 
	   write cache if we don't have to care about GC and 
	   exception register reload

	file #227
	**** TYA1.2v4 released ****
----------------------------------------------------------------------
Sun Feb 14 20:25:16 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c, tyarecode.c
	 * no more useless calls of JavaLangObject<init>
	 * some optimisation: leave out maintaining of a complete 
	   JavaFrame in some special cases (statics etc.)

	file #226
----------------------------------------------------------------------
Sun Feb 14 10:25:16 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya*.S, tyaruntime.c, tyarecode.c
	 some minor optimization around FastInvNoCheck call

Mon, 8 Feb 1999 17:37:53 C. Scott Ananian <cananian@lcs.mit.edu>
	-tyarecode.c
	 * added 0xCC opcode
	 * fixed some opcode names in the comments ( ldc* )

	file #225
---------------------------------------------------------------------
Mon Feb  1 16:35:00 1999  Kazuyuki Shudo <shudoh@muraoka.info.waseda.ac.jp>
	-tya.c, config*.*
	 added a robust way to find `struct sigcontext' and
	 made a patch on TYA 1.2v3. Now, a macro `USER_DS' is not
	 needed. The patch make EXCEPTIONS_BY_SIGNAL work with 
	 kernel 2.2 or 2.1. I've tested on kernel 2.2.1 and not 
	 tested on 2.0.X and 2.1.X.

Sun Jan 31 18:59:02 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaruntime.c
	 temp disabled optimizer patch from #207
	 (detected some JNI problems / 
	 tnx to  Denis Riedijk <ried@si.tn.tudelft.nl> )

	file #224
---------------------------------------------------------------------
Thu Jan 21 18:59:02 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c
	 improvement bugfix RI-string length (cmp #222)
	-tya*.c
	 replaced OR_regreg with TEST_regreg,
	 some MOV_AXxx replaced by XCHG_AX --- if possible
	 because better pairing / less code

Thu, 21 Jan 1999 07:05:41  Peter van Heusden <pvh@leftside.wcape.school.za>
	patch for TYA 1.2v3 configure.in - this adds a check based on
	'uname -s' to see if it is on FreeBSD. It also checks that 
	java actually exists on the system (by running 'java -version').

	file #223
-------------------------------------------------------------------
Sun Jan 17 19:30:07 1999  Kazuyuki Shudo <shudoh@muraoka.info.waseda.ac.jp>
	-tya.c
	 prevent that java_lang_Compiler_start(...) 
	 runs more than once (i.e. at startup only)  [due to GC!]

Sun Jan 17 20:19:15 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c
	 added bugfix RI-string end address [was introduced in #220]
	 tnx bugreport Lukas Knecht <knecht@infox.ch>

	file #222
	*** patch for TYA1.2v3 mailed on blackdown list ***
-------------------------------------------------------------------
Sun Jan 17 09:47:34 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya* : some common cleanup

	file #221
	**** released as TYA1.2v2 ****
--------------------------------------------------------------------
Tue Jan 12 19:57:02 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c, tyaruntime.c
	 * in hook #4 check at first for void methods (speed up
	   constructors)
	 * set invoker directly to hook #4, don't go
	   via JVM's own invoker (who will call our's)
	 * store length of RI-string in bit 8..15 of CompiledCodeFlags
	-tyarechelp.c
	 removed unused function CompAddFieldOffset()
	-tyarecode.c, tyaruntime.c
	 removed some 1.0.2 stuff in instanceof and checkcast 
	 opcodes [ resides in MyIsInstanceOf() ]

	file #220
--------------------------------------------------------------------
Sun Jan  3 12:00:48 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c
	 * added more tests in hook #5 (sig handler) for more 
	   robustness against sig 8 and sig 11 in native code
	 * set hook #5 only if #ifdef EXCEPTION_BY_SIGNALS
	 * no more panic() in hook#5 but simply return to caller

	file #219
---------------------------------------------------------------------
Sat Jan  2 19:28:58 1999  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaexc.c 
	 simplified CompTriggerExceptionFromInvoked()
	-tyaruntime.c: 
	 removed old native invoker: now
	 only JNINATIVENEW available, changed result 
	 of CodeRunner32() from long long to int

Mon Dec 28 20:40:06 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c
	 removed an useless getting EE into ebx

	file #218
---------------------------------------------------------------------
Tue Dec 15 22:20:06 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c 
	 swiched #define IGNORE_DISABLE on (again)
	-tyarecode.c 
	 rejected <clinit> optimizer patch from #193

	file #217 
---------------------------------------------------------------------
Fri Dec 11 19:52:58 1998  Albrecht Kleine  <kleine@ak.sax.de>

	******************************
	*****  released TYA 1.2  *****
	******************************

	file #216
-----------------------------------------------------------------------
Wed Dec  9 19:34:07 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-cfg files 
	 added configuration example for jdk1.1.5 FreeBSD
	 port (it's outdated, but was an intersting experience)
	-tya.S, tyaasm_FreeBSD.S
	 added new files for the FreeBSD port using some basics
	 by Keith Ball <KBall@encanto.com>
	-tya.c, tya.h, tyaexc.c
	 fixing bug for moved code during realloc(), 
	 good for FreeBSD port

	file #216
-----------------------------------------------------------------------
Sat Nov 21 20:07:17 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c, tya.c
	 * minor improvements in inlining
	 * bugfix from #214 for compling using undefined 
	   TRY_FASTINVOKE (store EE still in register EBX)
	 * some experiments with clinits (rejected)
	 * removed an old FIXME in resolving process 
	   (was address const cmp, good for FreeBSD port)

	file #215
-----------------------------------------------------------------------
Thu Nov 19 21:44:15 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c, tya.c, tya.h, tyaexc.c, tyaasm.S
	 * removed one extra parameter in fast_invocation process,
	   squeezing some bytes there
	 * minor better output in VERBOSE_ASM86 mode

	file #214
-----------------------------------------------------------------------
Sun Nov 15 21:23:03 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c
	 bugfix in goto_w opcode (CW vs CB macro)
	-tya.c, tyarecode.c, tyaexc.h, tya.h. tyaconfig.h
	 extended RESOLVE_FT level up to 3 by NOT discarding
	 the whole method, but generating a runtime LinkageError:
	 replaced ``RESERVED1'' exception place by placing
	 CompPrepareLinkageError() at compiled code @offset 0
	-demo/
	 added ExceptionsTest.java for checking proper
	 catching of null pointer exceptions via sig 11

	file #213
----------------------------------------------------------------------
Sat Nov 14 18:15:39 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c
	 * re-introduced Disable_hook(),
	 * minor clean-up from #211
	-tya.c, tyarecode.c, tya.h
	 moved recognition of possible method inlining
	 from recode() into CompileClass_Hook() using a new flag
	-tyaruntime.c
	 some ``#ifdef'' changes for proper compiling using undefined
	 FAST_NATIVE102STYLE
	 
	file #212 [zg]
----------------------------------------------------------------------
Fri Nov 13 20:29:47 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c,tya.c
	 bugfixes & new strategy for errorhandling exhausted 
	 memory due inline coding: ``continue, but use no inlining'',
	 and use a limitation for size of inlined method

	file #211 [tj]
----------------------------------------------------------------------
Mon Nov  9 18:00:00 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tya*.*
	 some cleanup / moved tyaprintstats to tyautil.c
	-tya.c
	 optimisation in InvokeCompiledMethod_Hook() :
	 check the 32bit result invokers at first
	-cfg files
	 added 1.1.7, removed 1.0.2 config stuff

	file #210
	**** released as TYA1.1v4 ****
----------------------------------------------------------------------
Sun Nov  1 20:06:29 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaruntime.c, tyaasm.S, tya.h
	 introduced better java stack overflow handling by
	 creating up to 10 JavaStacks: both for defined and 
	 undefined USEASM
	-tya.c
	 now including interface classes into running CompileClass_Hook
	 by removing an ``if'' in InitializeForCompiler_Hook(). 
	 This fixes a problem introduced in #203 with wrong compilation
	 of storage of invocation results for void and 64bit result
	 methods.
	-demo/JNI/*
	 added JNI benchmark by Artur Biesiadowski <abies@pg.gda.pl>
         (sources only)
	 
	file #209
---------------------------------------------------------------------
Sat Oct 31 16:10:51 1998  Artur Biesiadowski <abies@pg.gda.pl>
	-tyarecode.c, tya.h, tyaconfig.h
	 introduced GATHER_STATS: every opcode executed is counted up.
	 Use this for checking how much invokes of various types are
	 executed for various programs.
	 To get statistics you have to recompile with GATHER_STATS
	 defined, then run the java program and attach to it with gdb, 
	 for example:
	 > gdb /usr/local/java/bin/i586/green_threads/java process_pid
	 Then in gdb directly check stats_opcodes table or use:
	 $ print tyaprintstats()
	  ( You may need to set LD_LIBRARY_PATH to include
	    /usr/local/java/lib/i586/green_threads/ )
	    
	file #208
-------------------------------------------------------------------
Sat Oct 31 15:13:09 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaruntime.c, tya.c, tya.h, tyaconfig.h
	 continued new JNINATIVENEW coding [depends on invoker!]
	 * rewrote step A for CodeRunner32_withDummies() nomore 
	   using directly calls of dynolink/dynolinkJNI   (step B)
	 * extended  to three other CodeRunners           (step C)
	 * calling MakeStackRevInstruction() now in an earlier
	   phase: in function CompileClass_Hook()
	 * added special handling of short float args     (step D)

	files #205,#206,#207 (steps B,C,D)
-------------------------------------------------------------------
Wed Oct 28 20:31:38 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaruntime.c
	 started new approach to native method invoking
	 1. CodeRunnerXX_withDummies: adjusted, because we don't
	    have to expect usage for invocation of static methods:
	    always  mb->fb.access & ACC_STATIC == 0
	 2. special handling SYNCHRONIZED native methods
	    [later removed]
	 3. an experimental (giant) step to a later 
	    JNINATIVENEW coding

	file #204 (step A)
-------------------------------------------------------------------
Fri Oct 23 20:28:26 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tya*.*
	 1.reorganized usage of CompiledCodeFlags and -Info
	   (using own _flags_ controling invokers and result types
	   and removed FIXME from #201)
	 2.removed hook used only #ifdef DEBUG_INV (cmp #161)
	 
	file #203
-------------------------------------------------------------------
Sun Oct 18 21:18:02 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c
	 Changed conditions for compiling call of FastInvNoCheckXX,
	 this makes Netb startup better; cmp line 2 in build #187
	-tyarechelp.c
	 fixed NAN conversion bug (temp)
	 [tnx bugreport to
	  Kazuyuki Shudo <shudoh@muraoka.info.waseda.ac.jp>]
	-tyaexc.c
	 added MakeClassSticky call for StOvExc-class

	file #202
	**** released as TYA1.1v3 ****
-----------------------------------------------------------------
Thu Oct 15 21:23:39 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c
	 fixed athrow bug: more independence of EXCEPTION_BY_SIGNALS 
	 by CompTriggerNullPointerException2
	 [tnx bugreport to
	  Kazuyuki Shudo <shudoh@muraoka.info.waseda.ac.jp>]

	-tyarecode.c
	-tya.c
	 rejected a patch from #184 (as quick workaround
	 for jdk1.1.6v5)
	 [``different approach for getting signal context'']
	 
Sun Oct 11 17:18:25 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c
	 rewrote method return value recognition in
	 InvokeCompiledMethod_Hook() misusing ACC_DOCED flag:
	 marked ``FIXME'' == subject of change

	file #201 / patch mailed on blackdown list
-----------------------------------------------------------------
Sun Oct 11 11:32:25 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c,tyarecode.c
	 changed the conditions for possible method inlining
	
Sat Oct 10 17:50:04 1998  Ondrej Popp <ondrej@geocities.com>
Sat Oct 10 17:50:04 1998  Alexander Davydenko <alex@Javad.Ru>
	-tya.c
	 avoid a incompatible types in assignment 
	 compiler error
	
Sat Oct 10 17:50:04 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c
	 changed my typo from ``long'' to ``long long''

	file #200A
-----------------------------------------------------------------
Fri Oct  9 20:25:46 1998  Albrecht Kleine  <kleine@ak.sax.de>
	******************************
	*****  released TYA 1.1  *****
	******************************
	
	tya*.* / common clean up in all tya files
	file #200
-----------------------------------------------------------------
Tue Oct  6 18:56:59 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaruntime.c
	 workaround for JNI code similar disabling 
	 FAST_NATIVE102SYTLE - but less strictly

	file #199
-----------------------------------------------------------------
Sun Oct  4 21:00:02 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c, tya.c
	 move away code writing the result into optop[]
	 from *return opcodes (0xAX..0xB0) into 
	 InvokeCompiledMethod_Hook() / JIT compiled code called
	 by other JIT compiled code doesn't need that stuff
	 (similar was used up to file #55)

	file #198	
----------------------------------------------------------------
Sun Oct  4 16:34:02 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c
	 non statics invocation - handling a special case: stac==4
	-tya.h
	 bugfix was a buggy #define MOV_MCX_8_DX: cmp #195

	file #197
----------------------------------------------------------------
Wed Sep 30 16:57:19 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c, tyaasm.S
	 give param in directly in %eax to asm version of 
	 MyObjalloc() for opcode 0xBB,0xDD
	 
	-tyarecode.c, tyaruntime.c
	 opcodes new,newarray,anewarray: 0xBB,0xDD,.....
	 handle memory problems now via out_of_memory() 
	
	file #196 [nb] ..some public beta testing, but no reply :(
----------------------------------------------------------------
Mon Sep 28 17:48:47 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c
	 adjusted estimation formula for cinfo->resspace
	 (cmp changes in #186,#190)
	-tyarecode.c 
	 squeezed some bytes in *return opcodes
	 
	file #195
--------------------------------------------------------------
Sat Sep 26 09:52:53 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c, tyaruntime.c, tya.h
	 completed #192 for ``#ifndef USEASM'' (only useful
	 for debugging purpose)
	-tyarecode.c
	 sqeezed a POPBX after invocation 
	-tyaasm.S
	 added a short test for native code: do not call
	 JITCompileMethod() each time

	file #194
--------------------------------------------------------------
Thu Sep 24 19:57:11 1998  Artur Biesiadowski <abies@pg.gda.pl>
	-tyarecode.c
	 better handling of static finals in opcode 0xB2,0xD2,
	 added test file in demo/ directory: TestFinalStatic.java

	file #193
--------------------------------------------------------------
Thu Sep 24 19:56:57 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c,tyaasm.S,tyarecode.c
	 rewrote invocation code: any code which is not JIT
	 compiled runs via CodeRunnerXX_withDummies()
	 (The both dummies are needed for a common parameter 
	 structure with JIT compiled code)

	file #192  @ first anniversary of TYA 0.1 release date ;)
--------------------------------------------------------------
Thu Sep 17 21:48:38 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c
	 moved CompTriggerNullPointerException2 to start of 
	 invocation opcodes where mbp is known
	 [cmp files #176,#177,#180]
	-tyaasm.S
	 some experiments with a code self modifier [#190A,del.]

	file #191
--------------------------------------------------------------
Wed Sep 16 20:37:21 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c,tya.c,tyaruntime.c
	 new approach for getting mbp if invoke_interface:
	 A) via interfaceklass->imethodtable             (#188)
	 B) use more facts known during compilation time (#189)
	 C) use heavy assembler optimized inlined code   (#190)
	
	files #188-190 (steps A,B,C)
---------------------------------------------------------------
Fri Sep 11 18:19:05 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyarecode.c
	 1. sqeezed an ``add ebx,4'' from virtual_invocation
	 2. better handling of direct recursive invocation
	-tyarechelp.c
	 currently no more write cache for inc/dec operations:
	 is suboptimal but needed if exceptions are to catch
	 [tnx bugreport Fulco <phongsri@tin.it>]

	file #187	 
----------------------------------------------------------------
Sat Sep  5 19:30:46 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaexc.c
	 restore esi,edi registers if exceptions are to catch
	 [tnx bugreport Fulco <phongsri@tin.it>]
	 
Thu Sep  3 19:55:02 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-added demo/Makefile (forgotten in #168 - A.K.)
	-tya.c 
	 fixed typo - replaced ``#ifndef GUA_WORKAROUND''
	 with ``#ifdef NOGUA_WORKAROUND'',  cmp #175
	 This solves a problem with classes compiled by jikes.
	 [tnx bugreport Giovanni Lagorio <gio@libertyline.com>]

	file #186
----------------------------------------------------------------
Thu Aug 27 13:45:51 1998  Artur Biesiadowski <abies@pg.gda.pl>
	-tyarecode.c
	 patch corrects wide handling for _load and _store opcodes

	file #185
----------------------------------------------------------------
Sun Aug 16 11:11:09 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-makefile.in
	 added forgotten tyarechelp.c + added dest "tya.prep"
	-tya.c
	 different approach for getting signal context (depeding
	 on stackpointer, no more search loop)

	file #184
----------------------------------------------------------------
Mon Aug 10 19:52:09 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tyaruntime.c
	 replaced dynolink with dynolinkJNI call / this solves
	 some heavy problems running native code

	file #183
----------------------------------------------------------------
Sat Aug  8 20:52:40 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c, tyarecode.c, tyaconfig.h
	 introduced RESOLVE_FT (fault tolerance) level: do
	 NOT discard the complete caller method in case of
	 some resolver problems
	-tyaasm.S, tyarecode.c
	 squeezed a byte in opcode 0xBB

Sat Aug  8 11:36:07 1998  Rick van Rein <vanrein@cs.utwende.nl>
	-tya*
	 redirect stderr output to syslog if defined USE_SYSLOG
	 (fprintf(stderr,...  ---> iprintf)
	-FAQ+=2

	file #182
-----------------------------------------------------------------
Wed Aug  5 20:32:07 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-tya.c 
	 added switch IGNORE_DISABLE to ignore Disable_hook call 
	 (because trouble with a certain installer program)
	-tya.S
	 replaced call %ebx with call *%ebx
	 (reported by Rick van Rein and Jason Gilbert)
	-tya.h, tyaruntime.h
	 bugfix for running together with jdk1.0.2
	 (do not use USEASM and unhand stuff)

Wed Aug  5 20:32:07 1998  Rick van Rein <vanrein@cs.utwende.nl>
	-Makefile.in
	 reduced compiler switch -O6 down to -O3 because some gcc 
	 seems to run in trouble, added CSOURCES variable

	file #181
----------------------------------------------------------------
Sat Jul 25 18:00:00 1998  Albrecht Kleine  <kleine@ak.sax.de>
	******************************
	*****  released TYA 1.0  *****
	******************************
	
	file #180 (TYA 1.0 ex TYA 0.8pre)
----------------------------------------------------------------
Sat Jul 25 13:40:15 1998  Albrecht Kleine  <kleine@ak.sax.de>
	common clean up in all tya files

Sun Jul 12 21:08:40 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tyarecode.c
	-re-introduced CompTriggerNullPointerException2
	 at start of invocation opcodes (was first 
	 introduced in files #176,#177, but later removed)
	 [tnx bugreport Cristian Bogdan <cristi@nada.kth.se>]

	file #180 (started prerelease beta test)
----------------------------------------------------------------
Sun Jul 12 12:52:05 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tyarechelp.c, tyarecode.c, tya.h
	-moved recode helpers from tya.c into separate file
	-removed a FIXME: better checking for compilation buffer
	 overflow

	file #179
----------------------------------------------------------------
Wed Jul  8 20:39:49 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tyaconfig.h,tya*
	 -some experiments with FAST_NATIVE102STYLE
	 -fixed exception handler part 2 data type (old was int*)
	Makefile.in
	 -added destination "tya.list" for some assembler listing
	
Wed Jul  8 19:48:10 1998  Eric Howe  <mu@clio.trends.ca>
	tya*.c
	-added logging mechanism via lprint using a logfile
	 from environment variable TYA_LOGFILE
	 [ comment by A.K.: I've changed this somewhat:
	   1.any DEBUG or VERBOSE messages are written via stderr
	   2.if _no_ TYA_LOGFILE is set we're writing via stderr ]
	
	file #178	
-------------------------------------------------------------
Sun Jul  5 21:02:43 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tyarecode.c
	 -moved nullpointer exception check at start of 
	  invocation opcodes into native methods part
	  [later removed again / Jul  8 1998]

	file #177
-------------------------------------------------------------
Sun Jul  5 11:02:43 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c,tyarecode.c
	 -use panic() instead of abort()
	tyaasm.S, tya.c, tyaexc.c, tyaruntime.c, 
	 -removed an old FIXME: added check for StackOverflow 
	  in FastInvocationCheck (both C and ASM versions) 
	  using a pre-generated StackOverflowException done
	  in PrepareExceptions()
	tyaruntime.c
	 -temp bugfix: added a push %eax for stack reserve in 
	  void methods: solves a problem in WmfTest.

Wed Jul  1 20:50:53 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tyarecode.c, tyaexc.c
	 -bugfix: re-introduced nullpointer exception check 
	  at start of invocation opcodes: needed for catching 
	  native methods of null objects: added new function
	  CompTriggerNullPointerException2, [but is suboptimal])
	  This solves problem running a bdk demo.
	 
	file #176
------------------------------------------------------------
Tue Jun 30 21:33:51 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tyarecode.c, tya.c, tyaexc.c, tyaconfig.h
	-workaround in hook #11,#12 for better appearance
	 of exception stack trace: don't dump sublasses of 
	 classJavaLangThrowable
	-removed #define GUA_WORKAROUND (now default enabled)
	-added %esp correction into catch frame code
	-fixed a bug in backpatching (#173, #174)
	-added cfg for jdk1.1.6/v2/sbb

	file #175
------------------------------------------------------------
Fri Jun 26 18:46:54 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tyaexc.c, dasm.c
	-moved dasm.c to tyautil.c + added some funcs from tya.c
	-new code layout in memory (cmp sheet in tyaexc.c)
	-save jump to code, but use defined code ENTRYPOINT

	file #174 [!]
-------------------------------------------------------------
Thu Jun 25 22:06:12 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya*
	-introduced easy_exception_handler in tyaasm.S
	-rearranged genral exception_handler at end of code
	-using a backpatch to tell preparation code where 
	 to jump: easy exc handler or general exc handler

	file #173 [!]
-------------------------------------------------------------
Thu Jun 25 18:03:18 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tyaexc.c, tyarecode.c
	-handle ``athrow'' exceptions via common handler
	
	file #172
-------------------------------------------------------------	
Wed Jun 24 22:00:00 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c,tyaexc.c
	-handle exceptions from invoked via common handler
	 
	file #171
-------------------------------------------------------------
Tue Jun 23 20:45:17 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya*, dasm.c
	-moved some exception stuff to new file tyaexc.c
	-moved some VERBOSE_ASM86 stuff to dasm.c
	-introduced #define LOCALDATA,EXTRAARGS to make function 
	 InvokeCompiledMethod_Hook() better readable (I hope.)
	
	file #170
-------------------------------------------------------------
Tue Jun 23 20:22:14 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tyarecode.c, tya.h
	-changed ClassCastException handling: keep Object in eax
	 but check excepton state in ebx
	-more common exception preparation: saves about 100 byte
	 per compiled method: 
	-moved exception_generator to C code: better DEBUG facilities

	file #169
-------------------------------------------------------------
Fri Jun 19 19:16:08 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tya.h
	-bugfix in ClassCastException handling / removed a FIXME
	 plus re-adjusted start_of_code: aligned@4
	 (tnx to Artur for his testing stuff)

Thu Jun 18 14:30:00 1998  Bengt Kleberg <bengt@softwell.se>
	makefile.in, demo/makefile
	-added a test goal to the tya (make test)
	
	file #168
-------------------------------------------------------------
Mon Jun 15 22:07:13 1998  Artur Biesiadowski  <abies@pg.gda.pl>
	1.skipped lldiv/llrem wrappers and used directly
	  __divdi3 and__moddi3 
	  (control via #define INLINE_LONGARITM)
	2. better VERBOSE_ASM86 output for MOV_R_R.. x86 opcodes

Sun Jun 14 21:38:15 1998  Artur Biesiadowski  <abies@pg.gda.pl>
	configure.in, configure
	new option --with-jdk:	Specify where main java directory lies 
	(can be ommitted for autodetect)

	file #167
--------------------------------------------------------------
Sat Jun 13 20:13:56 1998  Albrecht Kleine  <kleine@ak.sax.de>
	config*
	-added config for Sergey_Nikitin_libc5_port 
	 dated June 8 1998
	tyarecode.c
	-added an #ifdef TRY_FAST_INVOKE (to avoid compiling
	 of some unused stuff) in invocation

Thu Jun 11 12:10:31 1998  Olivier Refalo <refalo@mail.dotcom.fr>
	added new file: dasm.c . This contains java_bytecode disasm,
	used if set #ifdef VERBOSE_ASM86

	file #166
---------------------------------------------------------------
Wed Jun 10 18:28:34 1998  Albrecht Kleine  <kleine@ak.sax.de>
	*****
	***** seventh public release (TYA 0.7) *****
	*****
	tya.c, tyarecode.c
	-common cleanup in program files and README
	-discarded 2 patches (reset exception status) from #160
	 
	file #165
---------------------------------------------------------------
Wed Jun 10 18:28:34 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tyarecode.c,tya.c
	-fixed another `mega' bug: handling compilation errors
	 in invocation / stack cleaning
	-fprintf (C)-message now to stderr
	
	file #164
---------------------------------------------------------------
Sun Jun  7 21:15:53 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c,tyarecode.c
	-replaced struct sigcontext_struct with TYA_SIGCONTEXT
	 (was handled in config but forgotten in tya.c)
	-added some msgs for VERBOSE_ASM86
	-fixed another `mega' bug: triggering exceptions vs.
	 stack cleaning in void methods (tnx to Eric S. Fraga for
	 bug report)

	file #163
---------------------------------------------------------------
Thu Jun  4 20:16:01 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tyarecode.c, *config*
	-temp bugfix: replaced alloca again with sysMalloc
	-rejected opt in MyIsInstanceOf()  [cmp file #156]
	(both because problems with a certain applet)
	-added configuration for jdk1.1.6-v1
	
	file #162  (started prerelease beta test)
---------------------------------------------------------------
Mon Jun  1 16:19:36 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c,tyarecode.c,tya.h
	-important bugfix in errorhandling (was problem if TYA 
	 can't compile a method caused by NoSuchMethodError etc.),
	 now we additionally save EBX on stack, so LOCSTART is now @40
	-extended DEBUG facilities: added new hook, 
	 but using only #ifdef DEBUG_INV

	file #161
---------------------------------------------------------------
Thu May 28 18:19:31 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c,tya.h:
	- fixed error handling in ProcessExceptionTab()
	  (it was possible to free an unalloced ptr)
	- changed error handling when compiling inline code
	  (now trigger a 2nd try without inlining)
	- reset exception status after compilation error
	tyarecode.c: 
	- inline 2nd try check + returning new error value
	- invocation added check if really CompiledCode is present
	  (removing a FIXME)
	- reset exception status after ResolveClassConstantFromClass
	  (opcode 0xBB: if resolving new classes)
	tyaruntime.c: 
	- additional check for proper invoker before going the 
	  fast_native_way (important if JIT compiling was aborted)

	file #160
--------------------------------------------------------------
Sun May 24 10:31:12 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tyarecode.c, tya.h, tyaasm.S
	-additional setting jf->optop in JITCompileMethod()
	 (compare files #144, #149, #158)
	-bugfix oom-exception handling in new opcode via
	 new function MyObjAlloc(): wrote C and ASM version
	 (also for MyArrAlloc(), but later discarded)

	file #159
-------------------------------------------------------------
Thu May 21 19:13:27 1998  Rene Schmit <rene@bss.lu>
	tya.c,tyarecode.c
	-improved mem allocation (for inlining and cinfo->...)
	-removed useless memory copy from JITCompileMethod()
	 but allocating more mem   --->compare file #144, #149
	-including MemDebug if neccessary

	file #158
-------------------------------------------------------------
Thu May 21 10:25:12 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tyarecode.c, tya.h
	-using sib coding in anewarray opcode
	-invoke opcode: introduced special handling for methods 
	 with stac==4 (one arg): eax contains already object

	file #157
-------------------------------------------------------------
Mon May 18 20:13:58 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c / tyaruntime.c  /  tyarecode.c, tyaasm.S
	-moved some byte in compiling of invoke*-opcodes
	 to ASM-version of FastInvokeCheck()
	-experimental peephole opt for a dcmp/if* combination
	  [later removed / Wed May 20 20:20:59 1998]
	-optimized some if-statements in MyIsInstanceOf()

	file #156
-------------------------------------------------------------
Sat May 16 19:14:18 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya* (touched all files except tya.c)
	-introduced an additional FastInvCheck() function written
	 fully in asm / switch via #define USEASM in tyaconfig.h
	 (good for 10% speedup in highly recursive programs,
	 but difficult to maintain)

Mon May 11 21:00.00 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tyarecode.c
	-removed a NOP from goto opcode

Sun May 10 09:40:27 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tyarecode.c, config*
	-fixed a mega bug: was using a wrong etype in invocation
	 leading to stack corruption (tnx for bug report to 
	 Relu <rpatrasc@purki.cis.uoguelph.ca>)
	-added cfg for jdk115v7 by sbb
	
	file #155
--------------------------------------------------------------
Tue May  5 21:13:12 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya* (touched all files)
	-removed compiling of extra local buffer (TEMPBUFINTS),
	 but alloced 3 extra ints for helper data
	-adjusted to new storage places [bp+LOCSTART+x], 
	 (for details compare new local data layout draft
	 in InvokeCompiledMethod_Hook() )
	-rewrote multiarrayalloc opcode (usable < 5 dimensions)
	 using a buffer on stack for dimensions array
	-squeezed again some byte from invocation opcode

	file #154
--------------------------------------------------------------
Sun May  3 19:54:08 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tyaruntime.c, tyarecode.c
	-in FastInvCheck(),CodeRunner() removed parameter "mbcaller",
	 ditto adjusted invoke section of recode()
	-in FastInvCheck() getting space on JavaStack now via JF++
	-squeezed another 2 byte in invoke section of recode(),
	 [ this is possible because FastInvCheck(),CodeRunner() 
	  are compiled preserving ebx ]

	file #153
---------------------------------------------------------------
Sat May  2 19:17:29 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tyarecode.c, tyaruntime.c, tya.h
	-ret,ret_w opcodes: added special Comp_PUSH_LocalVarToStack()
	-long div,mod: removed an useless PUSH/POP combination
	-invoke: added FastInvCheck2() [if we don't need have 
	 a runtime test whether compiled code is present]
	-invoke: sqeezed 3 byte for add esp,12 [if compileFASTinvoke]
	
	file #152
---------------------------------------------------------------
Mon Apr 27 21:02:11 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tya.h, tyarecode.c, tyaruntime.c
	-sqeeezed some bytes:
	 * 1 from CheckCast: need to change param order MyIsInstanceOf()
	 * 2 from each ireturn, lreturn (for synchronized methods)
	 * 4 from newarray (using a PUSHLONG2 + saving a storage in edx)
	 * 2 from invocation (removed a useless MOV_BXAX +
	 * 3 from invocation (if stack usage <128)
	-changed in anewarray opcode: order of 
	 CompTriggerOutOfMemoryException and CB(POPDX): 
	-moved MakeStackRevInstruction() from tya.c to tyaruntime.c

	file #151
-------------------------------------------------------------
Sun Apr 26 17:05:51 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tya.h, tyarecode.h
	-changed recode() interface using Classjava_lang_Class struct:
	 this saves often usage of unhand() macro, affects also
	 GetMethodNumber() and GetBlock(), 
	 hold compliance to jdk102 by using some #ifdef
	-merged RunCompiledCode() and InvokeCompiledMethod_Hook()
	-some cleanup in messages if #ifdef DEBUG  

Thu Apr 23 18:31:13 1998  Artur Biesiadowski  <abies@pg.gda.pl>
	tyarecode.c, tya.h
	-improved coding of getfield/setfield opcodes,
	 TestOpcodeField.java included :)

	file #150
--------------------------------------------------------------
Wed Apr 22 20:29:47 1998  Artur Biesiadowski  <abies@pg.gda.pl>
	tyarecode.c, tya.h, makefile.in
	-improved coding of putstatic and getstatic opcodes
	-added two scripts in demo/, to test new speedup and 
	 javacspeed to test javac startup and compilation speeds.
	-added make distclean to clean directory before 
	 distribution/patching.
	 
Mon Apr 20 19:58:08 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, aconfig.h, configure.in
	-temp bugfix: restored state of #144 in JITCompileMethod()
	 (because problems #undef TRY_FAST_INVOKE and 
	  disappeared main menu in appletviewer)
	-added jdk113v2 in configuration files
	
	file #149
--------------------------------------------------------------
Sun Apr 19 10:48:15 1998  Albrecht Kleine  <kleine@ak.sax.de>
	sixth public release (TYA 0.6), highlights are:
	-including configuration tool
	-bugfixes
	-supporting the	glibc jdk-ports again

Fri Apr 17 18:57:49 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-prepared new release 0.6: adapted release notes
	-bugfix in CompTriggerExceptionFromInvoked() [cmp #139]
	-common cleanup

	file #148
-------------------------------------------------------------
Tue Apr 14 21:25:07 1998  Artur Biesiadowski  <abies@pg.gda.pl>
	NEW: introduced configure script for autoconf,
	recognizing installed jdk + optional dest directory

Tue Apr 14 21:25:07 1998  Albrecht Kleine  <kleine@ak.sax.de>
	some adjustments in *.h for Artur's automagic configure

Mon Apr 13 11:00:57 1998   Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tyaruntime.c, tya.h, tyarecode.c
	-squeezed some bytes in Comp87Vergleich compiling
	-added test for quick opcode in GetBlock()
	-help cc optimizer somewhat in InvokeCompiledMethod_Hook()
	-removed useless arg in JITCompileMethod()
	-re-introduced goto_w, jsr_w, ret_w opcodes
	-added NULL ptr check in inline invocation coding
	
	file #147
------------------------------------------------------------
Sat Apr 11, AK: I confirmed a prerelease for the Debian people
	call it TYA 0.5A, based on patched #140, but
	running with glibc ports
	
------------------------------------------------------------
Tue Apr  7 19:37:05 1998  Albrecht Kleine  <kleine@ak.sax.de>
	started to split tya.c into some parts: 1st tyaruntime.c

Tue Apr  7 19:23:23 1998  Artur Biesiadowski  <abies@pg.gda.pl>
	rename tyaasm.s to tyaasm.S because make problem
	using gcc -pipe switch
	
	file #146
-------------------------------------------------------------
Mon Apr  6 20:59:40 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tya.h, tyarecode.c
	-renamed llmul,lldiv,llrem to llmul_wrapper,.. etc
	 solves a problem using TYA together with the
	 glibc2.0.x compiled JDK-ports
	 (success widely reported by glibc users)
	 
	-moved #define dprintf to tya.h 
	 (because some stdio.h have this defined)

	file #145
-------------------------------------------------------------
Sun Apr  5 16:09:04 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tyarecode.c
	-moved tyaconfig.h as FIRST included file
	-removed LIBSHORTCUT experiment

Sun Apr  5 09:08:39 1998  Rene Schmit <rene@bss.lu>
	tya.c
	-improved profiling the JIT compilation
	-new hexdump function
	-removed useless memory copy from JITCompileMethod()

	file #144
--------------------------------------------------------------
Sat Mar 28 20:42:07 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c
	-sqeezing some bytes in CompExceptionHandler
	-better calculation of start_of_code address
	 (less wasted space / alignment@4)

	file #143
--------------------------------------------------------------
Fri Mar 27 20:45:05 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tyaasm.s
	-removed inline asm from CodeRunner[32|64]
	-introduced new asm written funtions CRun[32|64]asm()

	file #142
--------------------------------------------------------------
Sun Mar 22 11:00:19 1998  Albrecht Kleine  <kleine@ak.sax.de>
	tya.c, tya.h, tyaasm.s, Makefile
	-removed CompTriggerArrayStoreExceptionSub
	 merged caller with former ArrEleIsInstanceOf() to 
	 an asm written function into new tyaasm.s file
	-local var offset range bugfixed (80 to 128)

Wed Mar 18 13:14:00 1998  Eugen N.Vasilchenko <eugen@inter-soft.com.ru>
	tyarecode.c
	-bugfix saload/baload in sign extension
	-use SIB addressing for ?astore opcodes
	
	file #141
--------------------------------------------------------------
Sun Mar 15 12:35:58 1998  Albrecht Kleine  <kleine@ak.sax.de>
	fifth public release (TYA 0.5) / highlights are:
	* faster code
	* smaller size
	* bugfixes
	* new concepts
	+ merged all texts to one README (12 kbyte)
	
Sun Mar 15 12:30:13 1998  Rene Schmit <rene@bss.lu>
	-Makefile: added green_threads to INSTDIR
	-allocate 1 extra byte in MakeStackRevInstruction()
	 [and changed alloc call to sysMalloc call  A.K.]

Thu Mar 13 21:32:02 1998  Albrecht Kleine  <kleine@ak.sax.de>
	removed CONSIDER_PRE stuff in invocation code
	(because trouble using a certain applet)

Wed Mar 11 20:06:19 1998  Albrecht Kleine  <kleine@ak.sax.de>
	commented out broken ldiv/lrem EXCEPTION_BY_SIGNALS 
	(all others are okay)
	 
Mon Mar  9 21:54:56 1998  Artur Biesiadowski  <abies@pg.gda.pl>
	baload/bastore speeedup / forced sib-encoding

Mon Mar  9 20:03:51 1998  Albrecht Kleine  <kleine@ak.sax.de>
	bugfix invocation code (eax vs. ebx)
	
	[all merged to #140 AK]
-------------------------------------------------------------
Sun Mar  8 20:10:32 1998
	-CLEANED up & rearranged files into:
	 1. tyaconfig.h	-configuration settings
	 2. tya.c	-all except recode()
	 3. tyarecode.c	-function recode() == the main switch
	 4. tya.h	-prototypes and machine code defines
	                 [currently removed all static attributes]
	-minor optimization in ReverseCopyViaReversedScript()
	-changed AND_CX_ opcode in CompTriggerExceptionFromInvoked()
	-fprintf messages now prefixed with "TYA: "-string,
	 so everyone knows _who_ is writing to stderr ;)
	-optimization in invocation code: squeezed some MOVs
	-removed INITALIGN stuff again
	-started mini prerelease beta test

	file #139
--------------------------------------------------------------
Sat Mar  7 19:56:31 1998   Albrecht Kleine  <kleine@ak.sax.de>
	-decreased TEMPBUFOFFS down to 40 and renamed to
	 define TEMPBUFINTS 10 (integers in buffer)
	-optimization in getfield,putfield opcodes
	-dynamic search for sigcontext_struct for JDK!=113v2
	-increasing double ops due initial stack alignment to 8
	 at method startup code (control via #define INITALIGN)
	file #138

---------------------------------------------------------------
Thu Mar  5 18:18:47 1998   Albrecht Kleine  <kleine@ak.sax.de>
	-optimisation in handling opcodes: checkcast, anewarray
	 and arraylength, putstatic
	-in FastInvCheck(): deleted  JF->vars++;
	-changed #define JDK113_2 stuff
	file #137
	 
--------------------------------------------------------------
Mon Mar  2 21:05:39 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-special functions for running code returning 32 bits 
	 or 64 bits: CodeRunner[32|64]
	-changed Didier's shift opcodes from byte data to TYA style
	file #136
	
--------------------------------------------------------------
Sun Mar  1 11:35:01 1998  Albrecht Kleine  <kleine@ak.sax.de>
	special handling of compiling local space: this depends
	whether the mb is known during compilation or not
	(now compileLOCALspace==-1 means ``unknown stack usage'')
	file #135
	
--------------------------------------------------------------
Sat Feb 28 16:56:08 1998  Didier Gautheron <dgautheron@magic.fr>
	improved long shift and compare operations
	file #134
	
--------------------------------------------------------------	
Thu Feb 26 21:52:52 1998  Albrecht Kleine  <kleine@ak.sax.de>
	invocation code: introduced common parameters for
	both FastInvCheck() and CodeRunner(): this saves a lot of
	bytes for handling stack operations
	file #133
	
--------------------------------------------------------------
Wed Feb 25 21:11:21 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-removed old MyExecWrapper functions, removed CompStackReversion() 
	 from invocation code (saves a lot of space)
	-stack reversion is now done in CodeRunner procedure
	 (using an instruction string in another ReverseCopyViaScript()
	 function )
	file #132
	
--------------------------------------------------------------
Tue Feb 24 18:18:20 1998  Albrecht Kleine  <kleine@ak.sax.de>
	melted My[Static]ExecWrapper and nativeCodeRunner into one:
	 -saves space in invocaton code (for 64 bit results)
	 -fixes bug in handling linkage errors
	 -so we have only one common invocation way for all 
	  invocations (except fast_invocation)

Mon 23 Feb 1998 02:46:55  Didier Gautheron <dgautheron@magic.fr>
	bugfix: ANDed shift counter with 0x3F for 3 long shift opcodes

Mon Feb 23 20:52:35 1998  Albrecht Kleine  <kleine@ak.sax.de>
	bugfix: in RunCompiledCode [semicolon after if(...) ]
	(typo came in file #116)

	[ all merged to #131  AK ]
	
----------------------------------------------------------------
Sat Feb 14 14:57:37 1998  Albrecht Kleine  <kleine@ak.sax.de>
	-introduced an optimization in invocation:
	 first time considering the opcode before current opcode
	 control via #define CONSIDER_PRE
	-a minor change in land-opcode
	-some more comments
	-minor change in signal handler for JDK113-1 (version 16th August)

Sun Feb  8 17:16:17 1998  Olaf Flebbe <flebbe@science-computing.de>
	tya.h: additional macros CB,CW,CL: printing produced machine code

	[ both merged to #130: old stackmachine no longer supported  AK]
	
---------------------------------------------------------------
Sat Feb  7 22:44:55 1998  Olaf Flebbe <flebbe@science-computing.de>
	improved address calc in tya-pendpush.c
	useful for array access
	
Sun Feb  7 21:53:45 1998  Albrecht Kleine  <kleine@ak.sax.de>
	in CompExceptionHandler() use ExecEnv value 
	instead of NULL parameter for is_subclass_of() call

	[ both merged to #129B  AK ]
	
----------------------------------------------------------------
Sat Feb  7 15:15:05 1998  David Lucas <ddlucas@lse.com>
	more compliance with egcs 1.01/gnu-C 2.8
	local[1].h=(JHandle*)ee;  

Fri, 6 Feb 1998 17:27:40  Artur Biesiadowski  <abies@pg.gda.pl>
	1) Use temporary registers for ArrayOutOfBound if available
	2) Use mov eax, [eax*4+ebx] for iastore and iaload

Feb 2, 98 08:37:47  Olaf Flebbe <flebbe@science-computing.de>
	Optimisation in *load,*store,iconst_1
	
Fri Feb  6 20:50:37 1998  Albrecht Kleine  <kleine@ak.sax.de>
	bugfix in tableswitch (repairs broken INLINING)

	[ merged 4 above to #129 (A.K.) ]

-------------------------------------------------------------
Thu Feb  5 20:59:14 1998  Albrecht Kleine  <kleine@ak.sax.de>
	splitted tya.c into separate parts for
	old and new: pure stackmachine vs. pending push sm.
	initial draft: #128
	
-------------------------------------------------------------
Thu Feb  5 19:14:49 1998  Albrecht Kleine  <kleine@ak.sax.de>
	added bugfixes to improve ``pending push'' 
	(jsr/ret/exceptions) + clean up src code
	file #127 (trouble INLINING vs. PENDPUSH)
	
-------------------------------------------------------------
Tue Feb  3 xx.yy.zz 1998  Artur Biesiadowski  <abies@pg.gda.pl>
	extended ``pending push'' 
	file #126
	
-------------------------------------------------------------
Mon Feb  2 xx.yy.zz 1998  Artur Biesiadowski  <abies@pg.gda.pl>
	first try on ``pending push'' 
	NEW: CPU register eax is the logical stacktop
	file #125
	
------------------------------------------------------------
Sun Feb  1 16:20:00 1998  Albrecht Kleine  <kleine@ak.sax.de>
	file #124A,B,C: more optimization, 
	but later discontinued by #125
	
------------------------------------------------------------	
Sat Jan 31 16:38:00 1998  Albrecht Kleine  <kleine@ak.sax.de>
	file #124 merged #123 with Artur's exception handling,
	plus bugfixes for catch{} and for jdk1.1.3-1

------------------------------------------------------------
Wed Jan 28 19:40:00 1998  Albrecht Kleine  <kleine@ak.sax.de>
	file #123: changed opcode #defines, opt. in INLINING
	file #122: replaced FP addr mode [ebx] to [esp]
	
------------------------------------------------------------	
Tue Jan 27 xx:yy:zz 1998  Artur Biesiadowski  <abies@pg.gda.pl>
	catch some exceptions by signalhandler

------------------------------------------------------------	
Sun Jan 25 xx.yy.zz 1998  Olaf Flebbe <flebbe@science-computing.de>
	in TinyPeepholeOpt: eliminate stack operation for 
	doubles and more ecx/eax
	
-------------------------------------------------------------
Sat Jan 24 11:00:00 1998  Albrecht Kleine  <kleine@ak.sax.de>
	replaced fourth public release (TYA 0.4)
	because trouble with jdk113-2
	file #121
	
-------------------------------------------------------------
Thu Jan 22 18:17:06 1998  Albrecht Kleine  <kleine@ak.sax.de>
	fourth public release (TYA 0.4): much optmization
	faster code / smaller size / bugfixes / new concepts,
	fileversion #120
-------------------------------------------------------------
// OLDER CHANGELOGs (file #1...#119)
// ARE PRIVATE PROPERTY OF ALBRECHT KLEINE
// The public Changelog starts with file version #120.
// Below I've noticed only release dates and contributions.
-------------------------------------------------------------
Tue Jan 20 20:28:20 1998  Artur Biesiadowski  <abies@pg.gda.pl>
        added VERBOSE compile time stuff by Artur Biesiadowski
	<abies@pg.gda.pl>
	
-------------------------------------------------------------
Sun Dec  7 19:55:16 1997  Albrecht Kleine  <kleine@ak.sax.de>
	third public release (version 0.3),  fileversion #98
	Highlights are remarkable speed-up and some solved gc 
	problems.
	
-------------------------------------------------------------
Sat Dec  6 21:30:35 1997  Albrecht Kleine  <kleine@ak.sax.de>
	Replaced makefile following some hints by
	Jason Gilbert <jason@scott.net> and
	Artur Skawina <skawina@usa.net>
	
-------------------------------------------------------------
Wed Nov 19 11:36:43 1997  Albrecht Kleine  <kleine@ak.sax.de>
	second public release (version 0.2),  fileversion #93
	
-------------------------------------------------------------
Wed Sep 24 19:30:37 1997  Albrecht Kleine  <kleine@ak.sax.de>
	first public release alpha code version 0.1,  
	fileversion #71
