12 |
Tag: <post-commit CVS tag> |
Tag: <post-commit CVS tag> |
13 |
Description: |
Description: |
14 |
---------------------------------------------------------------------- |
---------------------------------------------------------------------- |
15 |
|
Name: Matthias Blume |
16 |
|
Date: 2000/05/15 16:20:00 JST |
17 |
|
Tag: blume-20000515-lightrebuild |
18 |
|
Description: |
19 |
|
|
20 |
|
1. I added an alternative to "-rebuild" to "makeml". The difference is |
21 |
|
that prior to calling CMB.make' the CM-variable "LIGHT" will be |
22 |
|
defined. In effect, the command will not build any cross-compiler |
23 |
|
backends and therefore finish more quickly. |
24 |
|
|
25 |
|
The "fixpt" script also takes a "-light" switch to be able to use |
26 |
|
this new facility while compiling for a fixpoint. |
27 |
|
|
28 |
|
2. I replaced all mentions of anchored paths in group owner specifications |
29 |
|
with simple relative paths (usually starting with ".."). |
30 |
|
The rationale is that a library's internal workings should not be |
31 |
|
compromised by the lack of some anchor. (An anchor is necessary |
32 |
|
for someone who wants to refer to the library by an anchored path, |
33 |
|
but it should not be necessary to build the same library in the first |
34 |
|
place.) |
35 |
|
|
36 |
|
3. I changed the way CM's tool mechanism determines the shell command |
37 |
|
string used for things like ml-yacc etc. so that it does not break |
38 |
|
when CM.Control.implicit_anchors is turned off. |
39 |
|
|
40 |
|
---------------------------------------------------------------------- |
41 |
|
Name: Matthias Blume |
42 |
|
Date: 2000/05/12 18:20:00 JST |
43 |
|
Tag: blume-20000512-ml-build |
44 |
|
Description: |
45 |
|
|
46 |
|
Fixed a bug in config/_ml-build that prevented ml-yacc and ml-lex from |
47 |
|
getting installed properly (by config/install.sh). |
48 |
|
|
49 |
|
---------------------------------------------------------------------- |
50 |
|
Name: Matthias Blume |
51 |
|
Date: 2000/05/12 17:30:00 JST |
52 |
|
Tag: blume-20000512-anchors |
53 |
|
Description: |
54 |
|
|
55 |
|
!!! NEW BOOT FILES !!! |
56 |
|
|
57 |
|
This change is in preparation of fading out support for "implicitly |
58 |
|
anchored path names". I went through all sources and used the |
59 |
|
explicit (and relatively new) $-notation. See system/README and the |
60 |
|
CM manual for more info on this. |
61 |
|
|
62 |
|
I also modified the anchoring scheme for some things such as "smlnj", |
63 |
|
"MLRISC", "cm", etc. to take advantage of the fact that explicit |
64 |
|
anchors are more expressive: anchor name and first arc do not have to |
65 |
|
coincide. This entails the following user-visible change: |
66 |
|
|
67 |
|
You have to write $smlnj/foo/bar instead of smlnj/foo/bar. In |
68 |
|
particular, when you fire up sml with a command-line argument, say, |
69 |
|
e.g.: |
70 |
|
|
71 |
|
sml '$smlnj/cmb.cm' |
72 |
|
|
73 |
|
At the ML toplevel prompt: |
74 |
|
|
75 |
|
CM.autoload "$smlnj/cmb.cm"; |
76 |
|
|
77 |
|
There is also a new controller in CM.Control that can be used to turn |
78 |
|
off all remaining support for implicit anchors by saying: |
79 |
|
|
80 |
|
CM.autoload "$smlnj/ |
81 |
|
#set CM.Control.implicit_anchors false; |
82 |
|
|
83 |
|
This causes CM to reject implicitly anchored paths. This is (for the |
84 |
|
time being) less permissive than the "final" version where there will |
85 |
|
be no more such implicit anchors and relative paths will be just that: |
86 |
|
relative. |
87 |
|
|
88 |
|
The next step (version after next version?) will be to make the |
89 |
|
default for CM.Control.implicit_anchors false. After the dust has |
90 |
|
settled, I can then produce the "final" version of this... |
91 |
|
|
92 |
|
Note: Since bootstrapping is a bit tricky, I provided new boot files. |
93 |
|
|
94 |
|
---------------------------------------------------------------------- |
95 |
|
Name: Matthias Blume |
96 |
|
Date: 2000/05/11 16:30:00 JST |
97 |
|
Tag: blume-20000511-sources |
98 |
|
Description: |
99 |
|
|
100 |
|
The main change is that I added function CM.sources as a generalized |
101 |
|
version of the earlier CM.makedepend. This entails the following |
102 |
|
additional changes: |
103 |
|
|
104 |
|
- CM.makedepend has been dropped. |
105 |
|
|
106 |
|
- CM manual has been updated. |
107 |
|
|
108 |
|
- TOOLS signature and API have been changed. |
109 |
|
|
110 |
|
---------------------------------------------------------------------- |
111 |
|
Name: Allen Leung |
112 |
|
Date: 2000/05/10 21:17 EDT |
113 |
|
Tag: leunga-20000510-moby-c--ssa |
114 |
|
Description: |
115 |
|
|
116 |
|
Various bug fixes and new features for C--, Moby and MLRISC optimizations. |
117 |
|
None of these affect SML/NJ. |
118 |
|
|
119 |
|
1. Register Allocation |
120 |
|
|
121 |
|
a. A new ra spilling module (ra/ra-spill-with-renaming) is implemented. |
122 |
|
This module tries to remove local (i.e. basic block level) redundancies |
123 |
|
during spilling. |
124 |
|
|
125 |
|
b. A new framework for performing region based register allocation. |
126 |
|
Not yet entirely functional. |
127 |
|
|
128 |
|
2. X86 |
129 |
|
|
130 |
|
a. DefUse for POP was missing the stack pointer [found by Lal] |
131 |
|
b. Reload for CALL was incorrect in X86Spill [found by John] |
132 |
|
c. Various fixes in X86Spill so that it can be used correctly for |
133 |
|
the new spilling module. |
134 |
|
|
135 |
|
3. SSA/IR |
136 |
|
|
137 |
|
a. New module ir/dj-dataflow.sml implements elimination based |
138 |
|
data flow analysis. |
139 |
|
|
140 |
|
4. MLRiscGen |
141 |
|
|
142 |
|
a. Fix for gc type annotation |
143 |
|
|
144 |
|
5. MDGen |
145 |
|
|
146 |
|
Various fixes for machine description -> ml code translation. For ssa |
147 |
|
only. |
148 |
|
|
149 |
|
---------------------------------------------------------------------- |
150 |
Name: Allen Leung |
Name: Allen Leung |
151 |
Date: 2000/05/08 22:17 EDT |
Date: 2000/05/08 22:17 EDT |
152 |
Tag: leunga-20000508-labexp |
Tag: leunga-20000508-labexp |