forked from sasagawa888/eisl
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathNEWS
executable file
·179 lines (124 loc) · 5.94 KB
/
NEWS
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
* Release 3.0
Early work on a github release for the masses.
* Release 2.5
Added in various documentation fixes from Michael Ernst (BTW: has anyone
considered getting slave labour or University Students doing technical
writing in CS to do proofreading as an exercise for GNU documents. At
least from the teaching point of view this would seem to me to be a good
thing).
Added in various fixes from Eugen Dedu which include re-releasing
the perf/ performance measurement library and various other fixes.
* Release 2.4
Built in a new test system which may or may not work :-) but aims to
handle separate build/install directories properly. The complete set
of tests has not been ported to the new test system yet. That can wait
for the next revision.
Added in Paul Suggs fixes.
Added in Phil Bleckers patches for 2.3, in particular:
o Check for gettimeofday in configure.in, and blindly assumed that if
it's not found then ftime is available.
o Changes to now.c are included to use ftime. (This allowed me to
cross-compile from Unix to MS-Windows using mingw32.)
o Check to see if stdin is defined around the prototype to L_buffer_dump
so that L_buffer.h can be included by itself in C++ files without nana.h.
o Remove the hard-coded CFLAGS from a couple of Makefile.am.
Redid the changes for cross compiling suggested by Alex Lancaster.
* Release 2.3
Eiffel.h has been modified to provide DO .. END macros for invariant
checking at the suggestion of Bertrand Meyer (Eiffels author).
* Release 2.2
Fixes for cross compiling supplied by Alex Lancaster applied, minor
changes to configuration system.
* Release 2.1
0. Added cycles.h which does cycle level timing using the
CPU specific instructions or registers. (This is currently
only supported on Pentium compatible machines using the
RDTSC instruction and ./configure --enable-rdtsc).
1. Minor fixes to the test system.
* Release 2.0
Added a few extra tools and some more documentation, in particular:
0. Nana Shortform Generator -- generates a HTML summary of your codes
interface.
1. Nana Performance Measurement -- measures the space/time overhead
of the nana library (or anything else).
2. Minor bug fixes/introductions.
3. Modified the test system to use the gnuware package
to keep track of version, host, target, changes.
* Release 1.14
Minor fixes for egcs and a bit of documentation tidying up.
* Release 1.13
Fixes to documentation and problems with compiling under Solaris.
* Release 1.12
Added WITHOUT_NANA macro to allow you to compile without any
nana features. You must use the VL((...)) version of macros
for any macro with a variable number of arguments.
[Hopefully this solution is better than the previous one]
* Release 1.11
Minor fixes to documentation.
Added functions for handling the variable number of arguments
to macros problem (well sort of, see VL(...) in L.h). This was
suggested by Gordon Matzigkeit.
* Release 1.10
Added a flag, etc to L_buffer.* so that the writing to circular buffers
can have wraparound disabled (thus preserving the first (oldest) k bytes
of messages, rather than the default last (newest) k bytes).
A L_time component has been added that logs a time stamp and char*
into a circular buffer. It should be a bit faster than L_buffer since
it doesn't use vsnprintf formatting.
A majordomo mail archive has been setup at ([email protected]).
Send "subscribe nana <your-address>" to [email protected]
After some discussions with Richard M. Stallman (rms) Nana will be
released as a GNU tool. Note that the release conditions remain the
same, i.e. you can still use nana in a commercial project.
[Thanks to Richard for all his work over the years]
* Release 1.09
Added a call list modules (calls) which is a container for
function calls. This can be used to print all the objects
in a program or to check internal consistency.
This version should be released on the FreeBSD 2.2.5R CDROM.
* Release 1.08
Some minor fixes to the check-mail target.
Modified L_buffer_puts to allow a NULL pointer as the argument, this maps
to "(null)" (suggested by Phil Blecker).
* Release 1.07
Added examples and documentation for the use of Eric Allman's syslog
with nana.
Fixed an embarassing bug in L_buffer.c.
* Release 1.06
A fix for a small autoconfigure problem under FreeBSD plus some documentation
changes. The FreeBSD ports mechanism defines INSTALL_SCRIPT which
seems to confuse the autoconf 2.12 generated configure script.
* Release 1.04/1.05
This fixes a bug that was introduced into L_buffer.c at 1.03
to allow for machines without vsnprintf. Thanks to Neal Becker
and Oliver Elphick.
* Release 1.03
Minor changes to documentation, fixes to tests and fixes to various
portability problems.
* Release 1.02
Some minor changes to the testing mechanisms and libraries.
The check target now submits the result to a mail alias where
I can check the results.
* Release 1.01
L_buffer.h - improved documentation for circular buffer logging etc.
eiffel.h - support for eiffel like assertions (class invariants, method
pre and postconditions).
Qstl.h - support for Standard Template Library (STL) iterators and quantifiers.
(only for C++ programmers)
L_buffer.h - documented and tidied the logging to a circular buffer library.
This keeps the last X messages in an in-core buffer.
(this is ideal for those "what was the system doing before it
broke" questions).
Some hopefully useful changes to configuration, in particular:
% I_DEFAULT=fast ./configure
which uses single instructions for assert failure
vs
% I_DEFAULT=verbose ./configure
which prints out a nice message (and then dumps core).
Documentation updated and improved (still a way to go).
Tests have been rewritten.
Updated to automake-1.1o.
* Release 1.00
I've removed the Ada support from the system and made up a FreeBSD
port and package of the system. I've also implemented some of the
suggestions made by George K.Bronnikov and Alexander Samoilov.