File: changes.txt

package info (click to toggle)
cccc 1:3.1.4-4
  • links: PTS
  • area: main
  • in suites: wheezy
  • size: 3,956 kB
  • sloc: ansic: 33,244; cpp: 10,527; java: 622; makefile: 156; sh: 11
file content (100 lines) | stat: -rw-r--r-- 4,514 bytes parent folder | download | duplicates (6)
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
CHANGE LOG FOR CCCC

This change log was begun at the suggestion of Kenneth H. Cox.

A number of early beta releases were done for CCCC in 1999, but this
file only tracks changes after the first public beta release of
2000, which was version 3.pre36.


version 3.pre37
Fixed problems prn1 and prn2 reported by Kenneth H. Cox.
These related to the processing of preprocessor commands.
Regression tests prn1.cc and prn2.cc added to the distribution.
Information added to the "About CCCC" section of the generated report
to nag people to participate in TJL's research project.
Change log added

version 3.pre40
Many changes relating to making code compile using Borland CBuilder.
Changes include:
- modified pccts/dlg/output.c to generate code in CLexer.cpp and CParser.cpp
which CBuilder can compile
- changed tracein(char*), traceout(char*) to tracein(const char*), traceout
(const char*) to eliminate CBuilder warning
- modified flag buffer length in cccc_utl.cc to eliminate overrun write 
detected by CodeGuard
- fixed nasty bug in Win32 filename globbing logic which would have caused
files with directories in their names not to be found
- added support for CBuilder to makefile and ccccdist.dsw
NB Making the code compile under CBuilder has been done as an exercise in 
portability, and has helped identify real problems in the code which can 
be addressed and will improve quality on all platforms.  The CBuilder port
is not recommended as a production platform at this time.

version 3.pre48
This version was the latest packaged version at SourceForge for over 
a year but I note to my shame that I didn't record any changes against 
it.

version 3.pre54
Relative to 3.pre48, this version has the following features:
redid build technology to rely on .bat files under win32 rather than
shipping a prebuilt binary of GNU make;
added build technology to build using Borland command line C++ compiler
(this version crashes in some of the unit tests at present, patches
to prevent this would be very welcome);  
fixed minor syntax problems to enable the code to compile under GCC 
2.95.3 (3.pre48 was released when EGCS 2.91.66 was current).

version 3.pre55
Various changes to accommodate GCC 3.0.2, including: hacking in the PCCTS
source code to prevent #endif directives having trailing information 
(commented or otherwise); fixing the type of an iterator to std::vector;
introducing a using directive from strstream.  When GCC 2.95 becomes old
enough that I don't need to support it any more I will need to modify
the strstream stuff to use stringstream instead.

version 3.pre63
Added support for a setup program for Win32 using Inno Setup.
Fixed a lot of parse issues with the Java parser.  All example programs in 
the Sun Java tutorial now parse without (mentioning) errors.
Modified the support URLs in various places to point to the email address
of the SourceForge cccc-users list rather than my own email address.

version 3.pre81
Fixed the diff.bat script to reflect the true outcome of the comparison.
Made cccc_met.cc round down to eliminate a difference in the output 
on prn14.
Added XML format output.
Fixed command line environment for Windows XP.  
Modified URLs to research project to reflect new naming at ecu.edu.au.

version 3.pre82
Implemented workaround for prn16.  This was a problem reported by Colin 
Watson, giving different test case output on different architectures.
The problem turns out to be due to unexpected (and so far unexplained)
actions being invoked from the parser, the workaround consists of filtering
the record_XXXXXX_extent functions so that they take no action if 
parameters which are expected to be non-empty strings turn out to be empty.
This change has modified the expected output for many of the test cases:
on inspection the new behaviour appears to be at least as valid in each
case, however users should be aware that this fix changes CCCC's opinion
on what is a module to exclude the names of types which only appear in
parameter lists.

version 3.pre84
Recoded use of strstreams to use stringstreams instead and eliminate 
warnings with GCC 3.2.
Added prn16 into list of tests which get executed and corrected reference
output files.

version 3.1.2
Modified version numbering convention.
Reworked build system for Win32 to use the free Microsoft Visual C++ Toolkit 2003.
Reworked Win32 installer to create development environment.
Reworked documentation.
Released under GPL.
Documentation modified to make it clear that the project is now dormant.