xref: /netbsd-src/external/bsd/tradcpp/dist/CHANGES (revision 31615c9617fab4df7f5e221552df7da87f14320d)
1release 0.5.3 (20190121)
2   - Fix markup typo in the man page.
3   - Abort on line numbering or column numbering overflow. Line
4     numbers are limited to values that fit in "unsigned int". Also
5     reject input lines longer than 2^32-1 characters. It seems
6     reasonable to presume that any input that violates these
7     constraints is someone screwing around and not a serious attempt
8     to compile or preprocess anything useful. Done in response to
9     n2129, but without getting into any of the silliness found there.
10   - Recognize __ia64__ for IA64 builds.
11   - Recognize __aarch64__ for 64-bit ARM builds, as sent in by
12     various people.
13   - Recognize __riscv__ and __riscv64__ for risc-v builds.
14
15release 0.5.2 (20160904)
16   - Fix typo in -U usage message, noticed by Joerg.
17   - Add a -debuglog option to send an execution trace to a file.
18     Intended to be used when debugging imake templates and other
19     complex input, not for debugging tradcpp itself.
20
21release 0.5.1 (20150612)
22   - Fix a stupid regression in 0.5 that causes it to not recognize a
23     pile of options.
24   - Fix output corruption caused by mishandling which macros are
25     currently in use. In particular, "curmacro" is only valid while
26     we're parsing a macro name and arguments, and can change once we
27     start expanding, so don't use it to clear the in-use flag. This
28     problem has been around all along but was only just exposed.
29   - Also don't set curmacro to null after calling expand_domacro as
30     that can cause us to think a macro name we just read is defined().
31     This one was introduced in 0.5.
32   - Don't use "remove" as a local variable as gcc 4.1 gets upset
33     about it vs. remove(3) in stdio.h.
34
35release 0.5 (20150612)
36   - Don't report unclosed comments as "No newline at end of file".
37   - Don't rely on <stdbool.h> existing, as (predictably) it doesn't
38     work on Solaris.
39   - Similarly, don't rely on C11 anonymous unions as the Solaris
40     compiler vomits on them.
41   - Typo fix in man page from Jason McIntyre; and change "Usage" to
42     "usage" in usage for pedantic reasons, from Igor Sobrado.
43   - Accept "-" as either input or output file name to mean stdin or
44     stdout respectively. Suggested by Jonathan Gray.
45   - Fix output spacing behavior to match gcc when newlines appear in or
46     while looking for macro arguments. Partly from Joerg Sonnenberger.
47   - Implement __FILE__ and __LINE__ macros. Mostly from Joerg Sonnenberger.
48   - Implement #line. Partly from Joerg Sonnenberger.
49   - Declare usage() with PF(). From wiz.
50
51release 0.4 (20130713)
52   - Fix stupid build problem introduced in 0.3.1.
53   - Accept and ignore -m32, which imake issues willy-nilly on a bunch
54     of platforms. I thought this had already been done, but apparently
55     not.
56   - Don't use the <err.h> functions. There are still people out there
57     using legacy systems missing them.
58   - Sort out some more issues pertaining to handling quoted strings.
59   - Add some more tests.
60
61release 0.3.1 (20130709)
62   - Don't leak memory and assert if a bad command-line option comes
63     after a -D or a -include foo.
64   - Since imake is a principal application for tradcpp and imake carefully
65     hides what it's doing when you run it, when rejecting an invalid option
66     be sure to report *what* that option is.
67
68release 0.3 (20130616)
69   - Don't eval the control expression of the first #if of a block when
70     already in a false block; it might not be valid. Reported by
71     Baptiste Daroussin.
72   - Don't recognize comments within character constants.
73   - Don't recognize macro argument parentheses or commas within strings,
74     or within character constants either.
75
76release 0.2 (20130611)
77   - auto-recognize more builtin PowerPC and mips macros
78   - pass -Wunused (partly from Baptiste Daroussin)
79   - allow absolute paths in include files (partly from Baptiste Daroussin)
80   - don't use getprogname() in the name of portability
81   - add tests arising from December 2010 tech-toolchain thread (one
82     from der Mouse, one of mine)
83   - clean out usage of sys/cdefs.h macros and don't use the implementation
84     namespace
85   - make -Wcomment work again
86   - fix handling of relative includes
87   - provide a man page
88   - other minor improvements
89
90release 0.1 (20130610)
91   - first release, works with at least some imake templates
92