Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 721, Mon Nov 6 14:03:39 2000 UTC	revision 777, Fri Jan 12 12:17:38 2001 UTC
#	Line 11	Line 11
11	Date:	Date:
12	Tag: <post-commit CVS tag>	Tag: <post-commit CVS tag>
13	Description:	Description:
14		----------------------------------------------------------------------
15		Name: Matthias Blume
16		Date: 2001/01/12 21:20:00 JST
17		Tag: blume-20010112-sync
18		Description:
19
20		I am just flushing out some minor changes that had accumulated in
21		my private branch in order to sync with the main tree.  (This is
22		mainly because I had CVS trouble when trying to merge _into_ my
23		private branch.)
24
25		Most people should be completely unaffected by this.
26
27		----------------------------------------------------------------------
28		Name: Allen Leung
29		Date: Thu Jan 11 21:03:00 EST 2001
30		Tag: leunga-20010111-labexp=mltree
31		Description:
32
33		1.  Removed the type LabelExp and replace it by MLTree.
34		2.  Rewritten mltree-simplify with the pattern matcher tool.
35		3.  There were some bugs in alpha code generator which would break
36		64-bit code generation.
37		4.  Redo the tools to generate code with the
38		5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
39		are now generated by perl scripts.
40
41		----------------------------------------------------------------------
42		Name: Matthias Blume
43		Date: 2001/01/10 21:55:00 JST
44		Tag: blume-20010110-rcc
45		Description:
46
47		The RCC stuff now seems to work (but only on the x86).
48		This required hacking of the c-calls interface (and -implementation) in
49		MLRISC.
50
51		Normal compiler users should be unaffected.
52
53		----------------------------------------------------------------------
54		Name: Matthias Blume
55		Date: 2001/01/09 01:20:00 JST
56		Tag: blume-20010109-rcc
57		Description:
58
59		This is a fairly big patch, flushing out a large number of pending
60		changes that I made to my development copy over the last couple of days.
61
62		Of practical relevance at this moment is a workaround for a pickling
63		bug that Allen ran into the other day.  The cause of the bug itself is
64		still unknown and it might be hard to fix it properly, but the
65		workaround has some merits of its own (namely somewhat reducing pickling
66		overhead for certain libraries).  Therefore, I think this solution should
67		be satisfactory at this time.
68
69		The rest of the changes (i.e., the vast majority) has to do with my
70		ongoing efforts of providing direct support for C function calls from
71		ML.  At the moment there is a new primop "RAW_CCALL", typing magic
72		in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
73		case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
74		translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
75		handle RCC.
76
77		The last part (the changes to mlriscGen.sml) are still known to be
78		wrong on the x86 and not implemented on all other architectures.  But
79		the infrastructure is in place. I had to change a few functor
80		signatures in the backend to be able to route the CCalls interface
81		from MLRISC there, and I had to specialize the mltree type (on the
82		x86) to include the necessary extensions. (The extensions themselves
83		were already there and redy to go in MLRISC/x86).
84
85		Everything should be very happy as soon as someone helps me with
86		mlriscGen.sml...
87
88		In any case, nothing of this should matter to anyone as long as the
89		new primop is not being used (which is going to be the case unless you
90		find it where I hid it :). The rest of the compiler is completely
91		unaffected.
92
93		----------------------------------------------------------------------
94		Name: Matthias Blume
95		Date: 2001/01/05 00:30:00 JST
96		Tag: blume-20010105-primops
97		Description:
98
99		Added some experimental support for work that I am doing right now.
100		These changes mostly concern added primops, but there is also a new
101		experimental C library in the runtime system (but currently not enabled
102		anywhere except on Linux/X86).
103
104		In the course of adding primops (and playing with them), I discovered that
105		Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
106		broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
107		To recover, I made type information in INL_PRIM mandatory and changed
108		prim.sml as well as built-in.sml accordingly.  The InLine structure now
109		has complete, correct type information (i.e., no bottom types).
110
111		Since all these changes mean that we need new binfiles, I also bumped the
112		version number to 110.32.1.
113
114		----------------------------------------------------------------------
115		Name: Matthias Blume
116		Date: 2000/12/30 22:10:00 JST
117		Tag: blume-20001230-various
118		Description:
119
120		Added proxy libraries for MLRISC and let MLRISC libraries refer
121		to each other using path anchors.  (See CM manual for explanation.)
122
123		Updated CM documentation.
124
125		Fixed some bugs in CM.
126
127		Implemented "proxy" libraries (= syntactic sugar for CM).
128
129		Added "-quiet" option to makeml and changed runtime system accordingly.
130
131		Added cleanup handler for exportML to reset timers and compiler stats.
132
133		----------------------------------------------------------------------
134		Name: Lal George
135		Date: 2000/12/22 22:22:58 EST 2000
136		Tag: Release_110_32
137		Description:
138
139		Infinite precision used throughout MLRISC.
140		see MLRISC/mltree/machine-int.sig
141
142		----------------------------------------------------------------------
143		Name: Matthias Blume
144		Date: 2000/12/22 23:16:00 JST
145		Tag: blume-20001222-warn
146		Description:
147
148		Corrected wording and formatting of some CM warning message which I
149		broke in my previous patch.
150
151		----------------------------------------------------------------------
152		Name: Matthias Blume
153		Date: 2000/12/22 21:20:00 JST
154		Tag: blume-20001222-anchorenv
155		Description:
156
157		Fixed CM's handling of anchor environments in connection with CMB.make.
158
159		----------------------------------------------------------------------
160		Name: Matthias Blume
161		Date: 2000/12/22 13:15:00 JST
162		Tag: blume-20001222-cleanup
163		Description:
164
165		Removed src/cm/ffi which does not (and did not) belong here.
166
167		----------------------------------------------------------------------
168		Name: Matthias Blume
169		Date: 2000/12/21 23:55:00 JST
170		Tag: blume-20001221-exn
171		Description:
172
173		Probably most important: CM no longer silently swallows all exceptions
174		in the compiler.
175		Plus: some other minor CM changes.  For example, CM now reports some
176		sizes for generated binfiles (code, data, envpickle, lambdapickle).
177
178		----------------------------------------------------------------------
179		Name: Matthias Blume
180		Date: 2000/12/15 00:01:05 JST
181		Tag: blume-20001215-dirtool
182		Description:
183
184		- "dir" tool added.
185		- improvements and cleanup to Tools structure
186		- documentation updates
187
188		----------------------------------------------------------------------
189		Name: Allen Leung
190		Date: Thu Dec 14 03:45:24 EST 2000
191		Description:
192		Tag:  leunga-20001214-int-inf
193		Description:
194
195		In IntInf, added these standard functions, which are missing from our
196		implementation:
197
198		andb : int * int -> int
199		xorb : int * int -> int
200		orb  : int * int -> int
201		notb : int -> int
202		<<   : int * word -> int
203		~>>  : int * word -> int
204
205		Not tested, I hope they are correct.
206
207		----------------------------------------------------------------------
208		Name: Allen Leung
209		Date: Fri Dec  8 19:23:26 EST 2000
210		Description:
211		Tag:  leunga-20001208-nowhere
212		Description:
213
214		Slight improvements to the 'nowhere' tool to handle OR-patterns,
215		to generate better error messages etc.  Plus a brief manual.
216
217		----------------------------------------------------------------------
218		Name: Lal George
219		Date: 2000/12/08 09:54:02 EST 2000
220		Tag: Release_110_31
221		Description:
222
223		- Version 110.31
224		----------------------------------------------------------------------
225		Name: Allen Leung
226		Date: Thu Dec  7 22:01:04 EST 2000
227		Tag:  leunga-20001207-cell-monster-hack
228		Description:
229
230		Major MLRISC internal changes.  Affect all clients.
231		Summary:
232
233		1.  Type CELLS.cell = int is now replaced by a datatype.
234		As a result, the old regmap is now gone.  Almost all interfaces
235		in MLRISC change as a consequence.
236
237		2.  A new brand version of machine description tool (v3.0) that generates
238		modules expecting the new interface.  The old version is removed.
239
240		3.  The RA interface has been further abstracted into two new functors.
241		RISC_RA and X86RA.  These functors have much simpler interfaces.
242		[See also directory MLRISC/demo.]
243
244		4.  Some other new source->source code generation tools are available:
245
246		a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
247		b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
248		I use this tool to generate the peephole optimizers---with the new
249		cell type changes, peephole rules are becoming difficult to write
250		without conditional pattern matching.
251
252		5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
253		cover the entire MLRISC source tree so many things broke.
254
255		6.  CM files have been moved to the subdirectory MLRISC/cm.
256		They are moved because there are a lot of them and they clutter up the
257		root dir.
258
259		7.  More detailed documentation to come...
260
261		NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
262		a makeml -rebuild first.  This is because of other other
263		changes that Matthias has made (see below).
264
265
266		----------------------------------------------------------------------
267		Name: Matthias Blume
268		Date: 2000/11/30 23:12:00 JST
269		Tag: blume-20001130-filereorg
270		Description:
271
272		Some manual updates and some file reorganizations in CM.
273
274		----------------------------------------------------------------------
275		Name: Matthias Blume
276		Date: 2000/11/24 17:45:00 JST
277		Tag: blume-20001124-link
278		Description:
279
280		Drastically improved link traversal code for the case that the dynamic
281		value was already loaded at bootstrap time.  As a result, CM and CMB
282		now both load blazingly fast -- even on a very slow machine.  Also,
283		memory consumption has been further reduced by this.
284
285		Warning: The format of the PIDMAP file has changed.  THerefore, to
286		bootstrap you have to do this:
287
288		1. Run CMB.make
289		2. Make a symbolic link for the boot directory:
290		ln -s sml.boot.ARCH-OS xxx
291		3. "Rebuild" the boot directory:
292		./makeml -boot xxx -rebuild sml ; rm xxx
293		4. Boot normally:
294		./makeml
295
296		----------------------------------------------------------------------
297		Name: Matthias Blume
298		Date: 2000/11/21 21:20:00 JST
299		Tag: blume-20001121-tools
300		Description:
301
302		Continued hacking on autoloading problem -- with success this time.
303		Also changed tool-plugin mechanism.  See new CM manual.
304
305		----------------------------------------------------------------------
306		Name: Matthias Blume
307		Date: 2000/11/19 14:30:00 JST
308		Tag:  blume-20001119-autoload
309		Description:
310
311		Some hacking to make autoloading faster.  Success for CMB, no success
312		so far for CM.  There is a reduced structure CM' that autoloads faster.
313		(This is a temporary, non-documented hack to be eliminated again when
314		the general problem is solved.)
315
316		----------------------------------------------------------------------
317		Name: Matthias Blume
318		Date: 2000/11/17 14:10:00 JST
319		Tag: blume-20001117-pickle-lib
320		Description:
321
322		1. Eliminated comp-lib.cm
323		2. Made pickle-lib.cm
324		3. Eliminated all uses of intset.sml (from comp-lib.cm)
325		4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
326		equivalent constructs from smlnj-lib.cm (INtHashTable).
327		5. Point 4. also goes for those uses of intmap.* in MLRISC.
328		Duplicated intmap modules thrown out.
329		6. Hunted down all duplicated SCC code and replaced it with
330		equivalent stuff (GraphSCCFn from smlnj-lib.cm).
331		7. Rewrote Feedback module.
332		8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
333		should be thrown out and equivalent modules from smlnj-lib.cm
334		should be used (IntRedBlackSet, IntListSet, ...).
335
336		Confirmed that compiler compiles to fixpoint.
337
338		----------------------------------------------------------------------
339		Name: Allen Leung
340		Date: 2000/11/10 18:00:00
341		Tag: leunga-20001110-new-x86-fp
342
343		A new x86 floating point code generator has been added.
344		By default this is turned off.  To turn this on, do:
345
346		CM.autoload "$smlnj/compiler.cm";
347		Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
348
349		Changes:
350
351		1.  Changed FTAN to FPTAN so that the assembly output is correct.
352		2.  Changed the extension callback for FTANGENT to generate:
353
354		fptan
355		fstp  %st(0)
356		instead of
357		fptan
358		fstpl ftempmem
359
360		3.  Numerous assembly fixes for x86.
361
362		5.  Cleaned up the machine code output module x86/x86MC.sml and added
363		support for a whole bunch of instructions and addressing modes:
364
365		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
366		faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
367		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
368		fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
369		fxch %st(n)
370		fld %st(n)
371		fst %st(n)
372		fst mem
373		fstp %st(n)
374		fucom %st(n)
375		fucomp %st(n)
376
377		All these are now generated when the fast fp mode is turned on.
378
379		6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
380
381		----------------------------------------------------------------------
382		Name: Matthias Blume
383		Date: 2000/11/09 11:20:00 JST
384		Tag: blume-20001109-scc
385		Description:
386
387		Eliminated some code duplication:
388
389		1. Added "where" clause to GraphSCCFn in SML/NJ Library.
390		(Otherwise the functor is useless.)
391		2. Used GraphSCCFn where SCCUtilFun was used previously.
392		3. Got rid of SCCUtilFun (in comp-lib.cm).
393
394	----------------------------------------------------------------------	----------------------------------------------------------------------
395	Name: Lal George	Name: Lal George

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.721   changed lines   Added in v.777

---

root@smlnj-gforge.cs.uchicago.edu	ViewVC Help
Powered by ViewVC 1.0.0