File: NEWS

package info (click to toggle)
numdiff 5.9.0-2
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid, trixie
  • size: 2,680 kB
  • sloc: ansic: 10,808; sh: 3,383; makefile: 276
file content (254 lines) | stat: -rw-r--r-- 9,528 bytes parent folder | download | duplicates (2)
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
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
5.8.1 --> 5.9.0 (Ivano Primi <ivprimi(at)libero(dot)it>) :

*   New option --raw for printing the differences between the compared
    files in raw format (the raw format is not convenient for humans
    but is very compact and can be easily parsed by other programs).

*   The output of the option -S includes now the positions of
    the first occurrences in the compared files of the largest
    absolute/relative error.

*   Thanks to Elias Pipping, the issue with long execution times and overloading
    of the processor in case Numdiff had been built against GNU MP with
    optimization enabled is now solved.

*   Documentation updated and corrected where needed.

5.8.0 --> 5.8.1 (Ivano Primi <ivprimi(at)libero(dot)it>) :

*   The distribution terms of the documentation have been slightly changed
    to comply with Debian policies.

5.6.1 --> 5.8.0 (Ivano Primi <ivprimi(at)libero(dot)it>) :

*   Ndselect recognizes now the following additional options:

    # -O to specify the string to use as separator 
         while writing the selected fields to the standard output,
    # -D to specify the set of strings to use as delimiters
         while splitting the input lines into fields. 

*   The long option --separator of Numdiff/Ndselect has been
    changed to --separators and its help message has been improved.

*   The options -s of Numdiff and -S of Ndselect recognize and accept
    now a larger set of escape sequences. 

*   Numdiff can detect now if any of the files to compare is binary
    and return a suitable error message.

*   The short form of Numdiff option --dummy is not -D anymore, but -U.

*   Numdiff accepts now the option --delimiters (short form: -D) to  
    specify the set of strings to use as delimiters
    while splitting the input lines into fields. 

*   The option --currency of Numdiff now recognizes and accepts escape 
    sequences in its argument.

5.6.0 --> 5.6.1 (Ivano Primi <ivprimi(at)libero(dot)it>) :

*   Added instruction to Makefile.in to create the directory
    for the man pages in case it does not exist yet.

5.2.1 --> 5.6.0 (Ivano Primi <ivprimi(at)libero(dot)it>) :

*   Several bugs have been corrected:

    # Fixed improper handling of the last line of a file
      when no trailing newline is present at its end.
    # Array overrun in function find_and_hash_each_file.
    # Fixed two bugs which prevented Numdiff to be built on Cygwin.
      Thanks to Satish Balay for the report.
    # Numdiff is not linked anymore against the
      GNU Multiple Precision Arithmetic Library if --disable-gmp 
      is passed to the configure script.

*   Numdiff recognizes now the following additional options:

    # -O for the "overview" mode,
    # -B to treat the compared files as binary files 
         (only meaningful  under  MSDog/MSWindog),
    # -c to set the currency name for the two files to compare,
    # -F to change the formula used to compute the relative errors.

*   Added new entries to Numdiff help which explain how to use the new options.

*   The options -a and -r accept now an extended syntax which allows
    to specify to which fields the given threshold applies.
    The option -r does not accept anymore the prefixes 1: and 2: (use
    -F instead).

*   Several help messages of Numdiff have been improved.

*   Several error and warning messages of Numdiff have been improved. 

*   ndselect allows now to select not only the lines but also
    the fields that have to be printed. 
    It recognizes then the following additional options:

    # -F to set the number of the first field to print,
    # -L to set the number of the last field that can be printed,
    # -I to set the increment to use when selecting the fields to print,
    # -S to specify the set of characters to use to split 
         the input lines into fields,
    # -x to avoid to print empty lines.

*   Added new entries to ndselect help which explain how to use the new options.

*   Several help messages of ndselect have been improved.

*   Added man pages for numdiff and ndselect.

*   Documentation, i18n and italian l10n have been updated.

5.2.0 --> 5.2.1 (Ivano Primi <ivprimi(at)libero(dot)it>) :

*   Corrected a bug which prevented Numdiff to be built on MacOS X(R).
    Thanks to Antoine Dechaume for the report.

5.0.0 --> 5.2.0 (Ivano Primi <ivprimi(at)libero(dot)it>) :

*   Numdiff uses now the GNU Multiple Precision Arithmetic Library
    (also called GNU MP or GMP) if this library is available 
    at build-time.
    The old internal support for multiple precision arithmetic
    is a fall-back in case GNU MP is absent.
    However it is possible to use the internal support for 
    multiple precision arithmetic even when GNU MP is available:
    it is sufficient to pass the option --disable-gmp
    (or --enable-gmp=no) to the  configure  script
    before building the program.
    After building and installing Numdiff,
    it is possible to know whether it has been
    linked against GNU MP or not just by calling the program
    with the option -v. In case of link against GMP
    the program tells also which version of
    GNU MP it uses to perform its computations.

5.0.0 beta4 --> 5.0.0 (Ivano Primi <ivprimi(at)libero(dot)it>) :

*   The filter has been improved, now it gets no more confused if
    the same value is represented in scientific notation in one file
    and in common decimal notation in the other file.

*   The Makefile.in (and consequently the Makefile) has
    been modified to conform to the GNU Coding Standards.

*   Now numdiff and ndselect accept long options too
    (for instance --separator='\n\t %' instead of -s '\n\t %').

*   The option -r accepts now an additional parameter which tells 
    to compute the relative differences always with respect 
    to the values in one file, instead of using the classic 
    formula for the relative error. 

*   The argument of the option -f has been made optional.

*   Documentation, i18n and italian l10n have been updated.

5.0.0 beta1 --> 5.0.0 beta4 (Ivano Primi <ivprimi(at)libero(dot)it>) : 

*   Added additional information to the report printed by
    the option -S.

*   Now the program can accept the options with argument

    -z INT, -Z INT, -X INT

    as short equivalent forms of 

    -z 1:INT -z 2:INT, 
    -Z 1:INT -Z 2:INT  and
    -X 1:INT -X 2:INT  respectively.

*   The option  -R  has been removed since not useful.
	
*   The option  -s  has now an extended syntax allowing to specify
different sets of field delimiters for the two files to compare.

*   Adoption of the version 1.3 of the GNU Free Documentation License
for the TeXinfo documentation.
	
*   Now the test programs are not anymore distributed together with
    the source code of Numdiff.    
	
4.2.0 --> 5.0.0 beta1 (Ivano Primi <ivprimi(at)libero(dot)it>) : 

*   Removed the support for double, long double 
    and high precision arithmetic. Now only multiple
    precision arithmetic is available (but it is not a limitation :) ).

*   Removed the option -L, since the tool ndselect made it obsolete.

*   Implemented a synchronization procedure to detect insertions/deletions
    of lines. To accomplish this, source code from GNU diff (diffutils 2.8.1)
    has been used (with the necessary modifications).
    Added the options -f, -z, -Z, -R, -m, -H, -T.

*   Removed the option -F, -X is now used in place of it.	

*   Added new entries to the help which explain how to use the new options.

*   The output format has been slightly changed.

*   Makefile and configuration scripts have been changed, since
    the code has become much more complex than it was.

*   Added more tests to check that the new features are correctly working.
	
*   Change of the license: from GPL version 2 or later to GPL version 3 or later.
    The copyright statement of my files has been consequently updated.
    Also the message shown by 'numdiff -v'.

4.0.0 --> 4.2.0 (Ivano Primi <ivprimi(at)libero(dot)it>) : 

*   Changed the way the options -V and -b interact when they are both set.

*   Added new tests.

*   Now, when the options -v and -h are both set, Numdiff displays before
    the information about version, Copyright and NO-Warranty and then
    the help message.

*   The Copyright has been updated.

*   Added the tools `ndfilter' and `ndselect'.

*   i18n and italian l10n updated.  

*   The documentation has been consequently updated.

3.0.2 --> 4.0.0 (Ivano Primi <ivprimi(at)libero(dot)it>) : 

*   New output format. 
    At the same time the code printing the output has been cleaned.

*   Added the -V option for the "verbose" mode.

*   Better handling of the last line of a file.

*   i18n, l10n and documentation have been consequently
    updated.

3.0.1 --> 3.0.2 (Ivano Primi <ivprimi(at)libero(dot)it>) : 

*   The code of the function Cabs() has been improved
    for both double and long-double precision.

*   Fixed a compilation error which occurred when one
    chose to activate the double, the long-double or the high
    precision instead of the multiple precision arithmetic.
    This compilation error was introduced just before releasing
    the version 3.0 of Numdiff .

3.0   --> 3.0.1 (Ivano Primi <ivprimi(at)libero(dot)it>) : 

*   The code of the function Cabs() has been modified
    in order to speed up the execution when working with
    real numbers (the change concerned only the code
    for multiple precision arithmetic).

3.0 (Ivano Primi <ivprimi(at)libero(dot)it>) : 

*   First public release