1 This is a list of changes that have been made since the 11/16/88 version
4 1. New features in 12/28/93
5 a. Associative arrays. The new version of ksh supports both
6 associate arrays and the older indexed arrays with the same
7 array syntax. A new -A option of typeset is used to declare
8 an array to be associative. As with indexed arrays, $name is
9 equivalent to ${name[0]}. The prefix operator ! was added
10 to the parameter expansion syntax to expand to the list of
11 indices. For example, ${!name[@]} expands to the list of array
12 indices for variable name.
14 b. Several additions have been made to shell arithmetic:
15 1. The shell now performs floating point arithmetic. The
16 typeset options -F and -E have been added for floating
17 point and scientific notation respectively.
18 2. The prefix and postfix ++ and -- operators.
19 3. The comma and ?: operators.
20 4. The math library functions.
21 5. An arithmetic for statement of the form
22 for ((expr1; expr2; expr3))
25 6. Integer arithmetic extended up to base 64.
27 c. Some additions to the macro expansion syntax have been made
28 to specify substrings and sub-arrays:
29 1. ${name:expr} expands to the substring of ${name} starting at
30 the character position defined by arithmetic expression expr.
31 2. ${name:expr1:expr2} expands to the substring of ${name} starting
32 at expr1 and consisting of at most expr2 characters.
33 3. ${name[@]:expr} expands to the values of ${name[@]} starting at
34 the element defined by arithmetic expression expr.
35 4. ${name[@]:expr1:expr2} expands to at most expr2 values of
36 ${name} starting at expr1.
37 5. ${@:expr} expands the positional parameters starting at expr.
38 6. ${@:expr1:expr2} expands to at most expr2 positional parameters
40 7. ${!name} expands to the name of the variable named by name.
41 It will expand to name unless name is reference variable.
42 8. ${!name[sub]} expands to the name of the subscript of the
43 given variable. If sub is @ or * the list of subscripts
45 9. ${!prefix*} and ${!prefix@} expand to the list of variable
46 names beginning with prefix.
47 10. The substring operators, # and % can be now be applied
48 with aggregates (@ or *) and are applied to each.
49 11. ${name/pattern/string} expands to the value of name with
50 the first occurrence of pattern replaced by string.
51 With aggregates (@ or *) this operation is applied to each.
52 12. ${name/#pattern/string} Same as above but the pattern
53 to be replaced must match at the beginning.
54 13. ${name/%pattern/string} Same as above but the pattern
55 to be replaced must match at the end.
56 14. ${name//pattern/string} expands to the value of name with
57 the each occurrence of pattern replaced by string.
58 With aggregates (@ or *) this operation is applied to each.
60 d. The name space for variables has been extended. The character '.'
61 can be used at the beginning of a name, and to separate identifiers
62 within a name. However, to create a name of the form, foo.bar,
63 the variable foo must exist. The namespace starting with .sh
64 is reserved for shell implementation variables. Exported
65 variable cannot contain a '.'.
67 e. Compound assignments. The assignment syntax, varname=value,
68 has been extended to allow assignments of the form
69 varname=(assignment_list). As elsewhere in the shell
70 spaces or tabs are optional around the parentheses, and
71 no space is permitted between the varname and the =. The
72 assignment_list can be one of the following:
73 1. A list of words. In this case each word is expanded as
74 in a for list and the resulting items become elements
75 of the indexed array varname.
76 2. A list of subscript assignments in the form
77 [subscript]=value. In this, these elements become
78 elements of the associative array varname.
79 3. A list of assignments; simple or compound. In this
80 case, each assignment is made to varname.name, where
81 name is the name of the enclosed assignment.
82 4. Assignments in the form of readonly or typeset
83 statements. In this case each assignment is made as
84 in 3 above, and the attributes are given to the
85 corresponding variable.
86 In case 3 and 4 above, the value of "$varname" after
87 the above assignment is (assignment_list), where the
88 assignment_list produced would reproduce all of the
89 variables under varname.*.
91 f. Function names of the form variable.action (called discipline
92 functions) can be defined where variable is any valid variable
93 name and action is get, set, or unset. The function variable.get
94 is invoked each time the variable is referenced. The set
95 discipline is invoked each time the variable is assigned to.
96 The unset discipline is invoked when a variable is unset.
97 The new variables .sh.name, .sh.subscript, and .sh.value are
98 defined inside the function body. Other shell extensions
99 may have their own set of discipline functions.
101 g. The compound command !, which negates the return value of the
102 following pipeline, has been added.
104 h. On systems that support dynamic loading with dlopen(), it is
105 now possible to add built-in commands at runtime with the
106 a builtin command named builtin.
108 i. The following builtins have been added:
109 1. command name [ ... ]
110 2. sleep [decimal-seconds]
111 3. builtin [-ds] [-f file] [name...]
112 4. getconf name [pathname]
115 j. An addition format for literal strings, $'....' can
116 be used where ever literal strings are valid. The string
117 inside the single quotes will be converted using the ANSI-C
118 escape conventions. Additionally, the escape sequence \E
119 expands to the escape character (default \033) whenever ANSI-C
120 escape sequences are recognized.
122 k. A typeset -n option has been added which causes the value of a
123 variable to be treated as a reference to another variable so that
124 variables can be indirectly named. For example, if $1 contains
125 the name of a variable, then typeset -n foo=$1 causes the variable
126 foo to be synonymous with the variable whose name is $1. A builtin
127 alias, nameref='typeset -n' has been added to aid mnemonics.
128 Reference names cannot contain a '.'. Whenever that portion of
129 a variable up to the first '.' matches a reference name, the
130 reference value is substituted. For example, with nameref foo=.top,
131 then ${foo.bar} is equivalent to ${.top.bar}. When used as the
132 index of a for or select loop, each assignment causes a
133 new name reference to occur.
135 l. The KEYBD trap has been added which is triggered when a key
136 or escape sequence is typed while reading from the keyboard
137 in an edit mode. This, combined with some new variables
138 makes it possible to program your key bindings in ksh.
140 m. New variables have been added:
141 1. FIGNORE defines a set of file names to be ignored in each
142 directory when performing pathname expansion, replacing
143 the rule that requires that a leading . be matched explicitly.
144 2. Variable sh.edchar contains the value of the keyboard character
145 that has been entered when processing a KEYBD trap. If the value
146 is changed as part of the trap action, then the new value replaces
147 the key or keys that caused the trap.
148 3. Variable sh.edcol is set to the character position of the cursor
149 within the input buffer during a KEYBD trap.
150 4. Variable sh.edmode is set to the escape character when in vi
152 5. Variable sh.edtext is set to the contents of the input buffer
154 6. HISTEDIT is checked before FCEDIT. FCEDIT is obsolete.
155 7. HISTCMD is the number of the current command in the history
157 8. Variable .sh.version is set to the version string for
159 9. Variable .sh.name is set to the name of the variable
160 that that was referenced or assigned to when executing a get
161 or set discipline function.
162 10. Variable .sh.subscript is set to the subscript for the variable
163 that was referenced or assign to when executing a get or
164 set discipline function.
165 11. Variable .sh.value is set to the new value for the variable
166 that was assigned to when executing the set discipline function.
168 n. New invocation and set -o options have been added:
169 1. set -o notify (or set -b) causes background completion messages
170 to be displayed as soon as the job completes.
171 2. There is a compile time option named KIA which enables
172 creation of a relational database for commands, variables
173 and functions defined and referenced by a script. The
174 option -I <filename>, causes the database to be generated
175 in <filename>. The database format can be queried via
177 o. ksh93 can read and evaluate pre-compiled scripts generated by
178 a separate program called shcomp.
179 p. More work on internationalization has been added:
180 1. The decimal point character is processed per locale
181 2. A $ can be placed in front of each string to indicate
182 that the string needs translation but is otherwise ignored.
183 This means that if a message catalog of all $"..." strings
184 is generated, then a program such as print $"hello world"
185 could display "bonjour monde" in the french locale.
186 q. Backreferences have been added to pattern matching. The sequence
187 \d, where d is a digit from 1-9, matches the same string as
188 the d-th previous parenthesis group. Backreferences
189 can be used within patterns, and within replacement strings
190 with any of the ${name/...} operators.
192 2. Changes made in 12/28/93
193 a. The output format of many commands has changed as follows:
194 1. System error messages are displayed whenever a failure
195 is caused by a system call.
196 2. The exit status has changed in many cases:
197 a. USAGE messages cause an exit status of 2.
198 b. Commands not found cause exit - 127.
199 c. Command found, but not executable - 126.
200 d. Terminated because of signal - 256+sig
201 3. The output of values from built-ins that contain special
202 characters are quoted in a manner that then can be re-input.
203 4. The trace output puts quotes around the output so that it
204 can be reused as input.
205 5. The output for trap is in a format that can be reinput the
206 the shell to restore the traps.
207 6. kill -l lists the signal names without numbers as
208 required by the POSIX standard.
210 b. The following changes have been made to shell functions:
211 1. The semantics of functions declared with name() has changed
212 to conform with the IEEE-POSIX 1003.2 standard. In particular,
213 these functions are executed in a dot script environment rather
214 than a separated function environment so that there are no
215 local variables and no scoping for traps.
216 2. Functions declared as function name, preserve the old ksh
217 semantics can be also used as the first argument to the dot (.)
218 command to have them executed in a dot script environment.
220 c. The command search rules have changed as follows:
221 1. Special built-ins (those with a dagger in front of them) are
223 2. Functions are executed next.
224 3. Other built-ins that do not require an executable version
225 (for example cd and read) come next.
226 4. If the command name contains a slash, the pathname corresponding
227 to the command name is executed.
228 5. If name corresponds to a previously encountered pathname
229 on the PATH variable, the corresponding command is executed.
230 6. If the command name does not contain a slash, then the PATH
231 variable is used to find an executable by that name. If
232 the directory that the command is found is also contained in
233 the FPATH variable, then the command treated as a function.
234 If the shell has a built-in version of the command corresponding
235 to this command, then the built-in version of this command
236 is executed. Otherwise, the shell remembers that pathname
237 corresponding to this command name and executes this pathname.
238 7. If the name is not found on PATH, then the directories in
239 FPATH are searched. If found, then the command is executed
242 d. Built-in commands options now conform to the IEEE-POSIX 1003.2
243 conventions with some additions. In particular,
245 will now print a Usage line for name, except for true, false,
246 colon, login, newgrp, echo, [, and command.
248 e. Tilde expansion is now performed as part of the word expansions.
249 The effect of this is that if word begins with ~ in ${name op word},
250 it will be expanded unless escaped.
252 f. Pathname expansion is no longer performed on redirection words
253 unless the shell is interactive.
255 g. Changes to shell and options:
256 1. The -n option has been enhanced to produce more warning and
257 portability messages.
258 2. The -C option is equivalent to -o noclobber. Files are
259 created with O_EXCL when -C is on.
261 h. The following changes have been made to [[...]]:
262 1. A string by itself is equivalent to -n string.
263 2. -e has been added as equivalent to -a.
264 3. == has been added as equivalent =.
265 4. -a and = are now considered obsolete.
266 5. Arithmetic comparisons are now considered obsolete.
268 i. kill has been changed as follows:
269 1. Signal names can be upper case or lower case.
270 2. Numerical arguments to kill -l cause the given signal names to
272 3. String arguments to kill -l cause the given signal numbers to
274 4. Synopsis changed for getopts conformance.
276 j. print has a -f format option which is equivalent to
277 the IEEE POSIX printf. Both print -f format, and
278 printf have the following extensions from IEEE POSIX:
279 1. Floating point formats are supported.
280 2. Size and precision specifications can be *.
281 3. The %d option can take an argument after precision to
282 specify the base that the number will be displayed.
283 4. A %q format can be used to output a string quoted so
284 that it can be re-input to the shell.
285 5. A %P format can be used to output the shell pattern which
286 corresponds to the give extended regular expression.
287 6. For numerical fields, the arguments can be arithmetic
288 expressions which will be evaluated.
289 7. The %n format works as described in ANSI-C.
291 k. The following changes have been made to fc:
292 1. It has been renamed hist. fc is now a predefined alias.
293 2. hist uses ${HISTEDIT:-$FCEDIT}. FCEDIT is obsolete.
294 3. A new -s option is equivalent to the obsolete -e -.
295 4. If the first argument refers to a command earlier than the
296 first accessible command, it now implies the first accessible
297 command, so that hist -l 1 lists all accessible history commands.
299 l. The dot command (.) has changed as follows:
300 1. The argument can be the name of a function declared as
301 function name. The function will execute without creating a
303 2. If there are arguments to the given script or function,
304 the positional parameters are restored to their original
305 value when . completes.
307 m. The read built-in has been changed as follows:
308 1. A -A option to read has been added to allow the fields to be
309 read into an indexed array.
310 2. A -t n option has been added which causes read to
311 timeout after n seconds when reading from a slow device.
312 3. A -d char option has been added which causes the read
313 to terminate at char rather than at new-line.
315 n. The trap command has been changed as follows:
316 1. Trap names can be either upper case or lower case.
317 2. Trap -p cause only the specified trap values to be displayed.
318 3. The value of trap in a subshell will be the value in the parent
319 shell until a call to trap which changes the trap settings has
320 been made. Thus, savetraps=$(trap) works as required by the
323 o. The exec command has been extended as follows:
324 1. The -c option clears the environment first.
325 2. The -a name option sets argv[0] to name for the program.
327 p. true and false are built-ins, not aliases to built-ins.
329 q. test has been modified to conform to the IEEE-POSIX 1003.2
330 standard when there are three or less arguments.
332 r. umask -S option displays the mask in a symbolic format.
334 s. wait now returns the correct exit status of any previous
335 background job that has not been waited for, not just
338 t. The whence built-in has an option -a which causes all
339 uses for the given command name to be reported.
341 u. unalias has -a option to clear all the aliases.
343 v. The times built-in command has been removed. The time
344 reserved word, without a command, gives time cumulative
345 time for the shell and its children. A built-in alias
346 for times should enable scripts using times to continue
349 w. Command substitution and arithmetic substitution will now be
350 performed for PS1, ENV, and PS4 evaluation in addition to
353 x. The SECONDS variable now displays elapsed time in floating
354 point seconds with 3 places after the decimal point by
357 y. The getopts built-in now handles the complete libast optget
358 functionality. If any errors have occurred with getopts
359 when it has reached the end of arguments, then the Usage
360 message will be generated from the option string and the
361 exit status from getopts will be 2 rather than 1. The
362 usage message will be stored in the OPTARG variable if
363 the option string contains a leading colon; otherwise
364 it will be printed on standard error automatically.
366 z. THE ENV file is only processed for interactive shell
367 invocations. In addition, the -x attributes for
368 aliases and functions is ignored.
370 aa. The built-in edit modes have been changed as follows:
371 1. The pathname completion and pathname listing options
372 now perform command completion and command listing
373 when applied to a word in the command position.
374 2. In emacs mode ^N as the first related command after
375 the prompt will move to the next command relative to the
376 last known history position.
377 3. In emacs mode, successive kill and delete commands will
378 accumulate their data in the kill buffer, by appending or
379 prepending as appropriate. This mode will be reset by any
380 command not adding something to the kill buffer.
381 4. The control-T of emacs mode has been changed to behave like
382 control-T in gnu-emacs.
383 bb. The TMOUT variable also sets a limit for select timeouts
384 and default timeouts for read.
387 4. The source code has undergone significant modification.
388 a. Much of the code has been rewritten, In many cases this has
389 resulted in significant performance improvement.
391 b. The code is organized differently. See the README files
394 c. Most configuration parameters now get generated using
395 the FEATURE mechanism of nmake. Other options are set
398 c. The are several new compile time options. See the README
399 file for details. Some of the old ones have been removed.
401 d. The install script is a Mamfile that is generated by
402 nmake and processed by a script that comes with the
405 e. There are far fewer global names. This should make it
406 must easier to add built-in commands without worrying
409 f. The code uses the sfio library which makes it possible
412 g. The code is written in ANSI C with full prototypes.
413 The code is based on the IEEE POSIX 1003.1 standard.
414 The code can be compiled with K&R C and with C++ by
415 using the ANSI cpp that comes with nmake or running
416 the code through the proto filter before pre-processing.
417 This happens automatically with our shipping system.
419 h. There is a programming interface for capturing references
420 and assignment to shell variables. It is also possible
421 to intercept variable creation and supply the array processing
422 function for that variable. See nval.3 for a description.