Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 972, Sun Oct 28 00:37:02 2001 UTC	revision 988, Wed Nov 21 21:37:47 2001 UTC
#	Line 13	Line 13
13	Description:	Description:
14
15	----------------------------------------------------------------------	----------------------------------------------------------------------
16		Name: Matthias Blume
17		Date: 2001/11/21 16:35:00 EST
18		Tag: blume-20011121-foot-in-mouth
19		Description:
20
21		Removed the "Release_110_37" tag because of a serious bug.
22		This will be re-tagged once the bug is fixed.
23
24		----------------------------------------------------------------------
25		Name: Matthias Blume
26		Date: 2001/11/21 16:14:00 EST
27		Tag: blume-20011121-forgottenfile
28		Description:
29
30		Forgot to add a file.  (Just a .tex-file -- part of
31		the CM manual source.)
32
33		----------------------------------------------------------------------
34		Name: Matthias Blume
35		Date: 2001/11/21 16:10:00 EST
36		Tag: blume-20011121-invalid_110_37
37		Description:
38
39		Note: I removed the original tag "Release_110_37" from this commit
40		because we found a serious bug in all non-x86 backends.
41		- Matthias
42
43		1. Modifications to the SML/NJ code generator and to the runtime system
44		so that code object name strings are directly inserted into code
45		objects at code generation time.  The only business the runtime system
46		has with this is now to read the name strings on occasions.
47		(The encoding of the name string has also changed somewhat.)
48
49		2. CM now implements a simple "set calculus" for specifying export lists.
50		In particular, it is now possible to refer to the export lists of
51		other libraries/groups/sources and form unions as well as differences.
52		See the latest CM manual for details.
53
54		3. An separate notion of "proxy" libraries has again be eliminated from
55		CM's model.  (Proxy libraries are now simply a special case of using
56		the export list calculus.)
57
58		4. Some of the existing libraries now take advantage of the new set
59		calculus.
60		(Notice that not all libraries have been converted because some
61		of the existing .cm-files are supposed to be backward compatible
62		with 110.0.x.)
63
64		5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
65		"General.exnMessage"!  The former relies on a certain hook to be
66		initialized, and that often does not happen in the stand-alone case.)
67
68		----------------------------------------------------------------------
69		Name: Lal George
70		Date: 2001/11/21  13:56:18 EST
71		Tag: george-2001121-pseudo-ops
72		Description:
73
74		Implemented a complete redesign of MLRISC pseudo-ops. Now there
75		ought to never be any question of incompatabilities with
76		pseudo-op syntax expected by host assemblers.
77
78		For now, only modules supporting GAS syntax are implemented
79		but more should follow, such as MASM, and vendor assembler
80		syntax, e.g. IBM as, Sun as, etc.
81
82		----------------------------------------------------------------------
83		Name: Matthias Blume
84		Date: 2001/11/14 11:52:00 EST
85		Tag: blume-20011114-srcname
86		Description:
87
88		1. Routed the name of the current source file to mlriscgen where it
89		should be directly emitted into the code object.  (This last part
90		is yet to be done.)
91
92		2. Some cleanup of the pgraph code to make it match the proposal that
93		I put out the other day.  (The proposal notwithstanding, things are
94		still in flux here.)
95
96		----------------------------------------------------------------------
97		Name: Lal George
98		Date: 2001/11/14 09:44:04 EST
99		Tag:
100		Description:
101
102		Fix for a backpatching bug reported by Allen.
103
104		Because the boundary between short and long span-dependent
105		instructions is +/- 128, there are an astounding number of
106		span-dependent instructions whose size is over estimated.
107
108		Allen came up with the idea of letting the size of span
109		dependent instructions be non-monotonic, for a maxIter
110		number of times, after which the size must be monotonically
111		increasing.
112
113		This table shows the number of span-dependent instructions
114		whose size was over-estimated as a function of maxIter, for the
115		file Parse/parse/ml.grm.sml:
116
117		maxIter            # of instructions:
118		10                      687
119		20                      438
120		30                      198
121		40                        0
122
123		In compiling the compiler, there is no significant difference in
124		compilation speed between maxIter=10 and maxIter=40. Actually,
125		my measurements showed that maxIter=40 was a tad faster than
126		maxIter=10! Also 96% of the  files in the compiler reach a fix
127		point within 13 iterations, so fixing maxIter at 40, while high,
128		is okay.
129
130		----------------------------------------------------------------------
131		Name: Matthias Blume
132		Date: 2001/10/31 15:25:00 EST
133		Tag: blume-20011031-pgraph
134		Description:
135
136		CKIT:
137		* Changed the "Function" constructor of type Ast.ctype to carry optional
138		argument identifiers.
139		* Changed the return type of TypeUtil.getFunction accordingly.
140		* Type equality ignores the argument names.
141		* TypeUtil.composite tries to preserve argument names but gives up quickly
142		if there is a mismatch.
143
144		installation script:
145		* attempts to use "curl" if available (unless "wget" is available as well)
146
147		CM:
148		* has an experimental implementation of "portable graphs" which I will
149		soon propose as an implementation-independent library format
150		* there are also new libraries $/pgraph.cm and $/pgraph-util.cm
151
152		NLFFI-LIB:
153		* some cleanup (all cosmetic)
154
155		NLFFIGEN:
156		* temporarily disabled the mechanism that suppresses ML output for
157		C definitions whose identifiers start with an underscore character
158		* generate val bindings for enum constants
159		* user can request that only one style (light or heavy) is being used;
160		default is to use both (command-line arguments: -heavy and -light)
161		* fixed bug in handling of function types involving incomplete pointers
162		* generate ML entry points that take record arguments (i.e., using
163		named arguments) for C functions that have a prototype with named
164		arguments
165		(see changes to CKIT)
166
167		----------------------------------------------------------------------
168	Name: Allen Leung	Name: Allen Leung
169	Date: 2001/10/27 20:34:00 EDT	Date: 2001/10/27 20:34:00 EDT
170	Tag: leunga-20011027-x86-fast-fp-call	Tag: leunga-20011027-x86-fast-fp-call

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.972   changed lines   Added in v.988

---

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