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
|
$OpenBSD: PROJECTS,v 1.1 1996/08/14 06:19:10 downsj Exp $
Things to be done in pdksh (see also the NOTES file):
* builtin utilities:
pdksh has most if not all POSIX/at&t ksh builtins, but they need to
be checked that they conform to POSIX/at&t manual. Part of the
process is changing the builtins to use the ksh_getopt() routine.
The following builtins, which are defined by POSIX, haven't been
examined:
eval time
The first pass has been done on the following commands:
. : alias bg break cd continue echo exec exit export false fc fg
getopts jobs kill pwd read readonly return set shift trap true umask
unalias unset wait
The second pass (ie, believed to be completely POSIX) has been done on
the following commands:
test
(ulimit also needs to be examined to check it fits the posix style)
* test suite
Ideally, as the builtin utilities are being POSIXized, short tests
should be written to be used in regression testing. The tests
directory contains some tests, but many more need to be written.
* internationalization
Need to handle with the LANG and LC_* environment variables. This
involves changes to ensure <ctype.h> macros are being used (currently
uses its own macros in many places), figuring out how to deal with
bases (for integer arithmetic, eg, 12#1A), and (the nasty one) doing
string look ups for error messages, etc.. It probably isn't worth
translating strings to other languages yet as the code is likely
to change a lot in the near future, but it would be good to have the
code set up so string tables can be used.
* trap code
* add the DEBUG trap.
* fix up signal handling code. In particular, fatal vs tty signals,
have single routine to call to check for pending/fatal traps, etc.
* parsing
* the time keyword needs to be hacked to accept options (!) since
POSIX says it shall accept the -p option and must skip a -- argument
(end of options). Yuck.
* lexing
the lexing may need a re-write since it currently doesn't parse $( .. ),
$(( .. )), (( ... )) properly.
* need to ignore contents of quoted strings (and escaped chars?)
inside $( .. ) and $(( .. )) when counting parentheses.
* need to put bounds check on states[] array (if it still exists after
the re-write)
* variables
* The "struct tbl" that is currently used for variables needs work since
more information (eg, array stuff, fields) are needed for variables
but not for the other things that use "struct tbl".
* Arrays need to be implemented differently: currently does a linear
search of a linked list to find element i; the linked list is not
freed when a variable is unset.
* functions
POSIX and at&t ksh functions are different in that POSIX functions
don't change disable/restore traps and option parsing (OPTIND/OPTARG
plus internal state) isn't saved/restored. The suggestion made in
POSIX.2 rationale is to have x() { .. } do the POSIX thing, and have
function x { ..} do the at&t ksh thing. So, should have two types of
functions.
* history
There are two versions of the history code, COMPLEX_HISTORY and
EASY_HISTORY, which need to be merged. COMPLEX does at&t style history
where the history file is written after each command and checked when
ever looking through the history (in case another shell has added
something). EASY simply reads the history file at startup and writes
it before exiting.
* re-write the COMPLEX_HISTORY code so mmap() not needed (currently
can't be used on machines without mmap()).
* Add multiline knowledge to COMPLEX_HISTORY (see EASY_HISTORY
stuff).
* change COMPLEX_HISTORY code so concurrent history files are
controlled by an option (set -o history-concurrent?). Delete
the EASY_HISTORY code.
* bring history code up to POSIX standards (see POSIX description
of fc, etc.).
* documentation
Some sort of tutorial with examples would be good. Texinfo is probably
the best medium for this. Also, the man page could be converted to
texinfo (if the tutorial and man page are put in the same texinfo
page, they should be somewhat distinct - i.e., the tutorial should
be a separate thread - but there should be cross references between the
two).
* miscellaneous
* POSIX specifies what happens when various kinds of errors occur
in special built-ins commands vs regular commands (builtin or
otherwise) (see POSIX.2:3.8.1). Some of this has been taken
care of, but more needs doing.
* POSIX says if an exec fails, the exit code should be 127 (not found)
or 126 (not executable)...
* remove static limits created by fixed sized arrays
(eg, getsc_(line[]), ident[], heres[], PATH, states(lex.c),
buffer size in emacs/vi code)
* merge the emacs and vi code (should reduce the size of the shell and
make maintenance easier).
[John Rochester is working on the merge]
* add POSIX globbing (eg, [[:alnum:]]), see POSIX.2:2.8.3.2.
* catch SIGWINCH and update the COLUMNS and LINES parameters (also,
need to let the command line editor know of change - ideally this
would work even if the editor was currently reading commands).
* teach shf_vfprintf() about long long's (%lld); also make %p use
long longs if appropriate.
* add \[...\] parsing to prompt printing (don't count width of chars
inside the \[..\] - used to keep escape sequences in prompts from
messing up command-line-editor's idea of where the cursor is)
* file(command) completion list in vi/emacs: change so a number-prefix
picks one of the possibilities (eg, if in vi: foo^[= lists fooa, foob
and fooc as possible completions, ^[2= would choose the second
possibility (foob)).
|
