Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 1048, Thu Jan 31 06:36:17 2002 UTC	revision 1067, Fri Feb 15 17:08:17 2002 UTC
#	Line 13	Line 13
13	Description:	Description:
14
15	----------------------------------------------------------------------	----------------------------------------------------------------------
16		Name: Matthias Blume
17		Date: 2002/02/15 12:08:00 EST
18		Tag: blume-20020215-iptrs
19		Description:
20
21		Two improvements to ml-nlffigen:
22
23		1. Write files only if they do not exist or if their current contents
24		do not coincide with what's being written.  (That is, avoid messing
25		with the time stamps unless absolutely necessary.)
26
27		2. Implement a "repository" mechanism for generated files related
28		to "incomplete pointer types".   See the README file for details.
29
30		----------------------------------------------------------------------
31		Name: Matthias Blume
32		Date: 2002/02/14 11:50:00 EST
33		Tag: blume-20020214-quote
34		Description:
35
36		Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
37		because of the new and improved tag generation scheme.  (Thanks to Allen
38		Leung for pointing it out.)
39
40		----------------------------------------------------------------------
41		Name: Lal George
42		Date: 2002/02/14 09:55:27 EST 2002
43		Tag: george-20020214-isabelle-bug
44		Description:
45
46		Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
47		of Isabelle on the x86.
48
49		From Allen:
50		-----------
51		I've found the problem:
52
53		in ra-core.sml, I use the counter "blocked" to keep track of the
54		true number of elements in the freeze queue.  When the counter goes
55		to zero, I skip examining the queue.  But I've messed up the
56		bookkeeping in combine():
57
58		else ();
59		case !ucol of
60		PSEUDO => (if !cntv > 0 then
61		(if !cntu > 0 then blocked := !blocked - 1 else ();
62		^^^^^^^^^^^^^^^^^^^^^^^
63		moveu := mergeMoveList(!movev, !moveu)
64		)
65		else ();
66
67		combine() is called to coalesce two nodes u and v.
68		I think I was thinking that if the move counts of u and v are both
69		greater than zero then after they are coalesced then one node is
70		removed from the freeze queue.  Apparently I was thinking that
71		both u and v are of low degree, but that's clearly not necessarily true.
72
73
74		02/12/2002:
75		Here's the patch.  HOL now compiles.
76
77		I don't know how this impact on performance (compile
78		time or runtime).  This bug caused the RA (especially on the x86)
79		to go thru the potential spill phase when there are still nodes on the
80		freeze queue.
81
82
83
84
85		----------------------------------------------------------------------
86		Name: Matthias Blume
87		Date: 2002/02/13 22:40:00 EST
88		Tag: blume-20020213-fptr-rtti
89		Description:
90
91		Fixed a bug in ml-nlffigen that was introduced with one of the previous
92		updates.
93
94		----------------------------------------------------------------------
95		Name: Matthias Blume
96		Date: 2002/02/13 16:41:00 EST
97		Tag: blume-20020213-cmlpq
98		Description:
99
100		Added new priority queue export symbols (which have just been added to
101		smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
102		not compile and the installer would choke.)
103
104		----------------------------------------------------------------------
105		Name: Matthias Blume
106		Date: 2002/02/13 16:15:00 EST
107		Tag: blume-20020213-various
108		Description:
109
110		1. More tweaks to ml-nlffigen:
111
112		- better internal datastructures (resulting in slight speedup)
113		- "-match" option requires exact match
114		- "localized" gensym counters (untagged structs/unions nested within
115		other structs/unions or within typedefs get a fresh counter; their
116		tag will be prefixed by a concatenation of their parents' tags)
117		- bug fixes (related to calculation of transitive closure of types
118		to be included in the output)
119
120		2. Minor Basis updates:
121
122		- added implementations for List.collate and Option.app
123
124		----------------------------------------------------------------------
125		Name: Matthias Blume
126		Date: 2002/02/11 15:55:00 EST
127		Tag: blume-20020211-gensym
128		Description:
129
130		Added a "-gensym" option to command line of ml-nlffigen.  This can be
131		used to specify a "stem" -- a string that is inserted in all "gensym'd"
132		names (ML structure names that correspond to unnamed C structs, unions,
133		and enums), so that separate runs of ml-nlffigen do not clash.
134
135		----------------------------------------------------------------------
136		Name: Matthias Blume
137		Date: 2002/02/11 12:05:00 EST
138		Tag: blume-20020211-gensml
139		Description:
140
141		A quick fix for a problem with GenSML (in the pgraph-util library):
142		Make generation of toplevel "local" optional.  (Strictly speaking,
143		signature definitions within "local" are not legal SML.)
144
145		Other than that: updates to INSTALL and cm/TODO.
146
147		----------------------------------------------------------------------
148		Name: Matthias Blume
149		Date: 2002/02/08 15:00:00 EST
150		Tag: blume-20020208-uniquepid
151		Description:
152
153		0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
154
155		1. The installer (config/install.sh) has gotten smarter:
156
157		- Configuration options are a bit easier to specify now
158		(in config/targets).
159		- Bug in recognizing .tar.bz2 files fixed.
160		- Installer automatically resolves dependencies between
161		configuration options (e.g., if you ask for eXene, you will
162		also get cml -- regardless whether you asked for it or not).
163		- Installer can run in "quieter mode" by setting the environment
164		variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
165		"completely silent", though.
166		- Build HashCons library as part of smlnj-lib.
167
168		2. A new scheme for assigning persistent identifiers to compilation
169		units (and, by extension, to types etc.) has been put into place.
170		This fixes a long-standing bug where types and even dynamic values
171		can get internally confused, thereby compromising type safety
172		(abstraction) and dynamic correctness.  See
173
174		http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
175
176		for an example of how things could go wrong until now.
177
178		The downside of the new scheme is that pids are not quite as
179		persistent as they used to be: CM will generate a fresh pid
180		for every compilation unit that it thinks it sees for the first
181		time.  That means that if you compile starting from a clean, fresh
182		source tree at two different times, you end up with different
183		binaries.
184
185		Cutoff recompilation, however, has not been compromised because
186		CM keeps pid information in special caches between runs.
187
188		----------------------------------------------------------------------
189		Name: Lal George
190		Date: 2002/02/07 15:34:13 EST 2002
191		Tag: <none>
192		Description:
193
194		Compilers that generate assembly code may produce  global labels
195		whose value is resolved at link time. The various peephole optimization
196		modules did not take this in account.
197
198		TODO. The Labels.addrOf function should really return an option
199		type so that clients are forced to deal with this issue, rather
200		than an exception being raised.
201
202		----------------------------------------------------------------------
203		Name: Lal George
204		Date: 2002/02/06 13:55:02 EST
205		Tag: george-20020206-ra-breakup
206		Description:
207
208		1. A bug fix from Allen.
209
210		A typo causes extra fstp %st(0)'s to be generated at compensation
211		edges, which might cause stack underflow traps at runtime.  This
212		occurs in fft where there are extraneous fstps right before the 'into'
213		trap instruction (in this case they are harmless since none of the
214		integers overflow.)
215
216		2. Pulled out various utility modules that were embedded in the modules
217		of the register allocator. I need these modules for other purposes, but
218		they are not complete enough to put into a library (just yet).
219		----------------------------------------------------------------------
220		Name: Matthias Blume
221		Date: 2002/01/31 16:05:00 EST
222		Tag: blume-20020131-sparc-ccalls
223		Description:
224
225		1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
226		of extra stack space by mistake.  Fixed.
227
228		2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
229
230		----------------------------------------------------------------------
231	Name: Allen Leung	Name: Allen Leung
232	Date: 2002/01/30	Date: 2002/01/30
233	Tag: leunga-20020130-nowhere-bug-fix	Tag: leunga-20020130-nowhere-bug-fix
#	Line 622	Line 837
837
838	----------------------------------------------------------------------	----------------------------------------------------------------------
839	Name: Matthias Blume	Name: Matthias Blume
	>>>>>>> 1.169
840	Date: 2001/09/18 15:35:00 EDT	Date: 2001/09/18 15:35:00 EDT
841	Tag: blume-20010918-readme11036	Tag: blume-20010918-readme11036
842	Description:	Description:

---

Colored Diff Long Colored Diff Unidiff Context Diff Side by Side

Legend: Removed from v.1048   changed lines   Added in v.1067

---

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