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
|
Wed Jan 25 09:11:34 1995 Christopher Lott <lott@bogner.informatik.uni-kl.de>
* Version 1.14
* scan.l: made code clean for C++ (tested with g++) by handling
flex's redefinition of input() -> yyinput() if the symbol
__cplusplus is defined. No other changes were necessary.
g++ still complains about a missing prototype for getopt(),
but adding one results in a link error, curious.
Tue Oct 18 18:54:51 1994 Christopher Lott <lott@bogner.informatik.uni-kl.de>
* Version 1.13
* scan.l: changed data type on two instances of variable 'c',
which is used to catch the value from input(), from char to int
after a bug report from an AIX user. I suspect char's are
unsigned on his machine; clearly they are signed on the Sun, my
development platform. Annoyingly, I had it right in one routine
but wrong in two others.
* Finally documented my revision control process so that I'll
stick to it. Generated a patch 1.12 -> 1.13.
Tue Oct 11 16:25:41 1994 Christopher Lott <lott@bogner.informatik.uni-kl.de>
* Version 1.12
* Makefile.in: removed the test directory's dependency on the
tarred/compressed/uuencoded file due to a loop. The test
directory may be unpacked unnecessarily, but this is not as
serious as failing to generate a new tarred/compressed/uuencoded
file after the test data changes.
Fri Sep 2 08:37:36 1994 Christopher Lott (lott@bogner.informatik.uni-kl.de)
* csize.man: added mention that trigraphs are not supported, and
neither dollar signs nor at signs are accepted in identifiers.
Changed discussion of handling comments on preprocessor lines.
Thu Sep 1 14:54:43 1994 Christopher Lott (lott@bogner.informatik.uni-kl.de)
* Version 1.11
* Responded to bug report from Sven-Ove Westberg <sow@cad.luth.se>
who correctly pointed out that csize was not properly handling
comments that occur on the same line as a preprocessor
directive. Extended function pp_dir() to handle comments.
Sun Aug 21 13:16:15 1994 Christopher Lott (lott@bogner.informatik.uni-kl.de)
* Version 1.10
* Shortened file names to stay within the 14-character constraint
imposed by some sytems.
* Added a missing closing quotation mark in the test directory
Makefile
* Adapted to version 2.4.7 of flex.
* Updated the README to admit that I do not fully utilize the
capabilities of the scanner.
* Reworked the changelog into Gnu's accepted format
* Moved function names in C files to column 1 to help ansi2knr
* Moved Makefile to Makefile.in, reworked it for autoconf
* Created configure.in so that I can use autoconf
* Copied aclocal.m4 from GNUm4-1.2 directory
* Created configure using autoconf
* Updated the INSTALL file to state clearly the dependence on UNIX
systems, and to reflect the addition of configure.
* Copied the install.sh file from GNUm4-1.2 for SYSV people.
Sun Jul 24 10:00:45 1994 Christopher Lott (lott@bogner.informatik.uni-kl.de)
* Version 1.9.
* Failure: the string literal "\\" was not recognized properly.
Removed fault by reworking code to process backslashes in
string literals more intelligently.
* Did the same for pp dirs, where the same fault was lurking.
Wed Jul 29 07:12.49 1994 Christopher Lott (lott@bogner.informatik.uni-kl.de)
* Version 1.8
* Split the README into a README and an INSTALL file.
* Renamed version.h to patchlevel.h
* Added a target to the Makefile for unpacking the tests.
Mon Jun 13 05:19:40 1994 Christopher Lott (lott@bogner.informatik.uni-kl.de)
* Version 1.7
* Added code to handle the cases where a totally blank line
appears in a comment or a preprocessor directive. Lines with
white-space characters were already being treated correctly,
but lines with nothing but a newline were not. In the cases
of comments or preprocessor directives, the line is not counted
in the blank-line category but instead in the comment or nonblank
noncomment line category, respectively.
* Updated the man page to explain these counting rules.
* Added test cases (v-*) for these conditions. Because counting
the blank lines depends on their surrounding context, I can't use
a grep/wc combination to verify the counts automatically.
Sun Jun 12 09:37:20 1994 Christopher Lott (lott@bogner.informatik.uni-kl.de)
* Version 1.6
* Removed fault that caused string literal continuation lines
not to be counted correctly as nonblank noncomment lines.
This was difficult and resulted in much reorganization of
the code that counts nonblank noncomment lines, newlines,
comments, and pp dirs to be more modular.
* Added test cases for continuation lines in string literals
and in preprocessor directives.
* Added -v flag == print version information.
Updated the man page accordingly.
Added file version.h, updated the Makefile.
* Updated the README to warn about different implementations
of getopt() that cause mismatches in the test ouput. This
is not a serious problem.
* Added many comments to explain the inner workings of the
counting functions.
Fri Jun 10 08:44:26 1994 Christopher Lott (lott@bogner.informatik.uni-kl.de)
* Version 1.5
* Put files under revision control
* Added preprocessor directives:
include_next, import, warning, sccs, ident, assert, unassert.
I can't find documentation of #indent in K&R 1/e, K&R 2/e,
or even in Stroustrup _C++_ 2/e. However, GCC's info file
mentions it, and it's supported both by gcc and Sun's cc.
The rest came from reading version 2.5.8 of GCC's cccp.c.
Changed the man page to reflect the additions.
* removed fault that caused string literals with embedded
backslashes and newlines to be handled improperly.
Added function string_lit() to process all string literals.
* Removed fault that caused backslashes outside of string
literals to generate a "bad character" message. Added a lex
rule. Now they are recognized, but the backslash processing
is still not sophisticated.
Changes between versions 1.3 and 1.4:
* Removed faults that caused multi-line comments to threw off
the newline count and to echo comments improperly.
Both were removed by switching back to the original
comment-eating code; dropped use of the code from flexdoc(1).
* Removed fault that caused nonempty whitespace lines not to be
echoed verbatim
* Removed fault that caused blank line immediately after pp
directive to be ignored. Changed pp-eating code to give
back the newline that terminates the pp directive; this
allows the scanner to recognize the beginning of a line (^)
in the pattern for whitespace lines.
* Removed ^ before pattern for preprocessor directives; i.e.,
they will be recognized even if non-whitespace preceeds
them. This deviates from the spec in K&R 2/e but matches
gcc's behavior. Changed the man page to reflect this.
* Aadded more cases to the regression test suite to detect faults.
Changes between versions 1.2 and 1.3:
* Upgraded to flex version 2.4.6; forced me to work around a
bug in flex - use yyrestart(FILE *) to set up the scanner.
* Removed fault that caused the the closing */ of a comment
to be omitted in the echoed output.
Changes between versions 1.1 and 1.2:
* Replaced comment-eating code with example from flexdoc(1) man page
* lex could not count pp dirs properly, so I gave up on lex
and switched to using flex.
* Cleaned up man page and Makefile
* Added scan.c.flex to the distribution for those w/o flex
Release 1.1 of csize:
* First release to interested parties.
|
