Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 958, Fri Oct 12 03:54:43 2001 UTC	revision 977, Wed Nov 14 16:53:16 2001 UTC
#	Line 13	Line 13
13	Description:	Description:
14
15	----------------------------------------------------------------------	----------------------------------------------------------------------
16		Name: Matthias Blume
17		Date: 2001/11/14 11:52:00 EST
18		Tag: blume-20011114-srcname
19		Description:
20
21		1. Routed the name of the current source file to mlriscgen where it
22		should be directly emitted into the code object.  (This last part
23		is yet to be done.)
24
25		2. Some cleanup of the pgraph code to make it match the proposal that
26		I put out the other day.  (The proposal notwithstanding, things are
27		still in flux here.)
28
29		----------------------------------------------------------------------
30		Name: Lal George
31		Date: 2001/11/14 09:44:04 EST
32		Tag:
33		Description:
34
35		Fix for a backpatching bug reported by Allen.
36
37		Because the boundary between short and long span-dependent
38		instructions is +/- 128, there are an astounding number of
39		span-dependent instructions whose size is over estimated.
40
41		Allen came up with the idea of letting the size of span
42		dependent instructions be non-monotonic, for a maxIter
43		number of times, after which the size must be monotonically
44		increasing.
45
46		This table shows the number of span-dependent instructions
47		whose size was over-estimated as a function of maxIter, for the
48		file Parse/parse/ml.grm.sml:
49
50		maxIter            # of instructions:
51		10                      687
52		20                      438
53		30                      198
54		40                        0
55
56		In compiling the compiler, there is no significant difference in
57		compilation speed between maxIter=10 and maxIter=40. Actually,
58		my measurements showed that maxIter=40 was a tad faster than
59		maxIter=10! Also 96% of the  files in the compiler reach a fix
60		point within 13 iterations, so fixing maxIter at 40, while high,
61		is okay.
62
63		----------------------------------------------------------------------
64		Name: Matthias Blume
65		Date: 2001/10/31 15:25:00 EST
66		Tag: blume-20011031-pgraph
67		Description:
68
69		CKIT:
70		* Changed the "Function" constructor of type Ast.ctype to carry optional
71		argument identifiers.
72		* Changed the return type of TypeUtil.getFunction accordingly.
73		* Type equality ignores the argument names.
74		* TypeUtil.composite tries to preserve argument names but gives up quickly
75		if there is a mismatch.
76
77		installation script:
78		* attempts to use "curl" if available (unless "wget" is available as well)
79
80		CM:
81		* has an experimental implementation of "portable graphs" which I will
82		soon propose as an implementation-independent library format
83		* there are also new libraries $/pgraph.cm and $/pgraph-util.cm
84
85		NLFFI-LIB:
86		* some cleanup (all cosmetic)
87
88		NLFFIGEN:
89		* temporarily disabled the mechanism that suppresses ML output for
90		C definitions whose identifiers start with an underscore character
91		* generate val bindings for enum constants
92		* user can request that only one style (light or heavy) is being used;
93		default is to use both (command-line arguments: -heavy and -light)
94		* fixed bug in handling of function types involving incomplete pointers
95		* generate ML entry points that take record arguments (i.e., using
96		named arguments) for C functions that have a prototype with named
97		arguments
98		(see changes to CKIT)
99
100		----------------------------------------------------------------------
101		Name: Allen Leung
102		Date: 2001/10/27 20:34:00 EDT
103		Tag: leunga-20011027-x86-fast-fp-call
104		Description:
105
106		Fixed the bug described in blume-20010920-slowfp.
107
108		The fix involves
109		1. generating FCOPYs in FSTP in ia32-svid
110		2. marking a CALL with the appropriate annotation
111
112		----------------------------------------------------------------------
113		Name: Matthias Blume
114		Date: 2001/10/16 11:32:00 EDT
115		Tag: blume-20011016-netbsd
116		Description:
117
118		Underscore patch from Chris Richards (fixing problem with compiling
119		runtime system under recent NetBSD).
120
121		----------------------------------------------------------------------
122		Name: Allen Leung
123		Date: 2001/10/12 17:18:32 EDT 2001
124		Tag: leung-20011012-x86-printflowgraph
125		Description:
126
127		X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
128
129		----------------------------------------------------------------------
130	Name: Lal George	Name: Lal George
131	Date: 2001/10/11 23:51:34 EDT	Date: 2001/10/11 23:51:34 EDT
132	Tag: george-20011011-too-many-instrs	Tag: george-20011011-too-many-instrs

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.958   changed lines   Added in v.977

---

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