Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 1 | /**************************************************************** |
| 2 | Copyright (C) Lucent Technologies 1997 |
| 3 | All Rights Reserved |
| 4 | |
| 5 | Permission to use, copy, modify, and distribute this software and |
| 6 | its documentation for any purpose and without fee is hereby |
| 7 | granted, provided that the above copyright notice appear in all |
| 8 | copies and that both that the copyright notice and this |
| 9 | permission notice and warranty disclaimer appear in supporting |
| 10 | documentation, and that the name Lucent Technologies or any of |
| 11 | its entities not be used in advertising or publicity pertaining |
| 12 | to distribution of the software without specific, written prior |
| 13 | permission. |
| 14 | |
| 15 | LUCENT DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, |
| 16 | INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. |
| 17 | IN NO EVENT SHALL LUCENT OR ANY OF ITS ENTITIES BE LIABLE FOR ANY |
| 18 | SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES |
| 19 | WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER |
| 20 | IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, |
| 21 | ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF |
| 22 | THIS SOFTWARE. |
| 23 | ****************************************************************/ |
| 24 | |
| 25 | This file lists all bug fixes, changes, etc., made since the AWK book |
| 26 | was sent to the printers in August, 1987. |
| 27 | |
Arnold D. Robbins | 5068d20 | 2020-02-06 22:27:31 +0200 | [diff] [blame] | 28 | February 6, 2020: |
| 29 | Additional small cleanups from Christos Zoulas. awk is now |
| 30 | a little more robust about reporting I/O errors upon exit. |
| 31 | |
Arnold D. Robbins | 78c79c0 | 2020-01-31 08:40:11 +0200 | [diff] [blame] | 32 | January 31, 2020: |
| 33 | Merge PR #70, which avoids use of variable length arrays. Thanks |
| 34 | to GitHub user michaelforney. Fix issue #60 ({0} in interval |
Arnold D. Robbins | 768d6b5 | 2020-01-31 08:54:10 +0200 | [diff] [blame] | 35 | expressions doesn't work). Also get all tests working again. |
| 36 | Thanks to Arnold Robbins. |
Arnold D. Robbins | 78c79c0 | 2020-01-31 08:40:11 +0200 | [diff] [blame] | 37 | |
Arnold D. Robbins | 4d9b129 | 2020-01-24 11:15:30 +0200 | [diff] [blame] | 38 | January 24, 2020: |
| 39 | A number of small cleanups from Christos Zoulas. Add the close |
| 40 | on exec flag to files/pipes opened for redirection; courtesy of |
| 41 | Arnold Robbins. |
| 42 | |
Arnold D. Robbins | de6284e | 2020-01-19 20:37:33 +0200 | [diff] [blame] | 43 | January 19, 2020: |
| 44 | If POSIXLY_CORRECT is set in the environment, then sub and gsub |
| 45 | use POSIX rules for multiple backslashes. This fixes Issue #66, |
| 46 | while maintaining backwards compatibility. |
| 47 | |
Martijn Dekker | fed1a56 | 2020-01-17 13:02:57 +0100 | [diff] [blame] | 48 | January 9, 2020: |
| 49 | Input/output errors on closing files are now fatal instead of |
| 50 | mere warnings. Thanks to Martijn Dekker <martijn@inlv.org>. |
| 51 | |
Arnold D. Robbins | c7eeb57 | 2020-01-05 21:18:36 +0200 | [diff] [blame] | 52 | January 5, 2020: |
| 53 | Fix a bug in the concatentation of two string constants into |
| 54 | one done in the grammar. Fixes GitHub issue #61. Thanks |
| 55 | to GitHub user awkfan77 for pointing out the direction for |
| 56 | the fix. New test T.concat added to the test suite. |
| 57 | Fix a few memory leaks reported by valgrind, as well. |
| 58 | |
Arnold D. Robbins | 7db55ba | 2019-12-27 12:02:52 +0200 | [diff] [blame] | 59 | December 27, 2019: |
| 60 | Fix a bug whereby a{0,3} could match four a's. Thanks to |
| 61 | "Anonymous AWK fan" for the report. |
| 62 | |
Arnold D. Robbins | 0b82bc6 | 2019-12-11 09:24:38 +0200 | [diff] [blame] | 63 | December 11, 2019: |
| 64 | Further printf-related fixes for 32 bit systems. |
| 65 | Thanks again to Christos Zoulas. |
| 66 | |
Arnold D. Robbins | 416c6db | 2019-12-08 21:43:32 +0200 | [diff] [blame] | 67 | December 8, 2019: |
| 68 | Fix the return value of sprintf("%d") on 32 bit systems. |
| 69 | Thanks to Jim Lowe for the report and to Christos Zoulas |
| 70 | for the fix. |
| 71 | |
Arnold D. Robbins | 108224b | 2019-11-10 21:19:18 +0200 | [diff] [blame] | 72 | November 10, 2019: |
| 73 | Convert a number of Boolean integer variables into |
| 74 | actual bools. Convert compile_time variable into an |
| 75 | enum and simplify some of the related code. Thanks |
| 76 | to Arnold Robbins. |
| 77 | |
Arnold D. Robbins | c879fbf | 2019-11-08 14:40:18 +0200 | [diff] [blame] | 78 | November 8, 2019: |
| 79 | Fix from Ori Bernstein to get UTF-8 characters instead of |
| 80 | bytes when FS = "". This is currently the only bit of |
| 81 | the One True Awk that understands multibyte characters. |
| 82 | From Arnold Robbins, apply some cleanups in the test suite. |
| 83 | |
Arnold D. Robbins | 2a8f175 | 2019-10-25 11:03:02 -0400 | [diff] [blame] | 84 | October 25, 2019: |
Arnold D. Robbins | b73bfab | 2019-11-08 14:31:05 +0200 | [diff] [blame] | 85 | More fixes and cleanups from NetBSD, courtesy of Christos |
Arnold D. Robbins | 2a8f175 | 2019-10-25 11:03:02 -0400 | [diff] [blame] | 86 | Zoulas. Merges PRs 54 and 55. |
| 87 | |
Arnold D. Robbins | 961eec1 | 2019-10-24 09:42:51 -0400 | [diff] [blame] | 88 | October 24, 2019: |
| 89 | Import second round of code cleanups from NetBSD. Much thanks |
Arnold D. Robbins | 78c79c0 | 2020-01-31 08:40:11 +0200 | [diff] [blame] | 90 | to Christos Zoulas (GitHub user zoulasc). Merges PR 53. |
Arnold D. Robbins | 1d6ddfd | 2019-10-24 10:06:10 -0400 | [diff] [blame] | 91 | Add an optimization for string concatenation, also from |
| 92 | Christos. |
Arnold D. Robbins | 961eec1 | 2019-10-24 09:42:51 -0400 | [diff] [blame] | 93 | |
Arnold D. Robbins | 1633ba1 | 2019-10-17 13:06:21 -0400 | [diff] [blame] | 94 | October 17, 2019: |
| 95 | Import code cleanups from NetBSD. Much thanks to Christos |
Arnold D. Robbins | 78c79c0 | 2020-01-31 08:40:11 +0200 | [diff] [blame] | 96 | Zoulas (GitHub user zoulasc). Merges PR 51. |
Arnold D. Robbins | 1633ba1 | 2019-10-17 13:06:21 -0400 | [diff] [blame] | 97 | |
Arnold D. Robbins | 7cae39d | 2019-10-06 22:34:20 +0300 | [diff] [blame] | 98 | October 6, 2019: |
| 99 | Import code from NetBSD awk that implements RS as a regular |
| 100 | expression. |
| 101 | |
Arnold D. Robbins | 34a6f41 | 2019-09-10 09:57:07 +0300 | [diff] [blame] | 102 | September 10, 2019: |
| 103 | Fixes for various array / memory overruns found via gcc's |
Arnold D. Robbins | 78c79c0 | 2020-01-31 08:40:11 +0200 | [diff] [blame] | 104 | -fsanitize=unknown. Thanks to Alexander Richardson (GitHub |
Arnold D. Robbins | 34a6f41 | 2019-09-10 09:57:07 +0300 | [diff] [blame] | 105 | user arichardson). Merges PRs 47 and 48. |
| 106 | |
Arnold D. Robbins | c95b960 | 2019-07-28 20:09:24 +0300 | [diff] [blame] | 107 | July 28, 2019: |
| 108 | Import grammar optimization from NetBSD: Two string constants |
| 109 | concatenated together get turned into a single string. |
| 110 | |
Martijn Dekker | 5b602ca | 2019-07-26 10:46:58 +0200 | [diff] [blame] | 111 | July 26, 2019: |
| 112 | Support POSIX-specified C-style escape sequences "\a" (alarm) |
| 113 | and "\v" (vertical tab) in command line arguments and regular |
| 114 | expressions, further to the support for them in strings added on |
| 115 | Apr 9, 1989. These now no longer match as literal "a" and "v" |
| 116 | characters (as they don't on other awk implementations). |
Arnold D. Robbins | c95b960 | 2019-07-28 20:09:24 +0300 | [diff] [blame] | 117 | Thanks to Martijn Dekker. |
Martijn Dekker | 5b602ca | 2019-07-26 10:46:58 +0200 | [diff] [blame] | 118 | |
Arnold D. Robbins | 4e34346 | 2019-07-17 21:14:52 +0300 | [diff] [blame] | 119 | July 17, 2019: |
| 120 | Pull in a number of code cleanups and minor fixes from |
| 121 | Warner Losh's bsd-ota branch. The only user visible change |
| 122 | is the use of random(3) as the random number generator. |
| 123 | Thanks to Warner Losh for collecting all these fixes in |
| 124 | one easy place to get them from. |
| 125 | |
Arnold D. Robbins | b8e0827 | 2019-07-16 20:54:44 +0300 | [diff] [blame] | 126 | July 16, 2019: |
| 127 | Fix field splitting to use FS value as of the time a record |
| 128 | was read or assigned to. Thanks to GitHub user Cody Mello (melloc) |
| 129 | for the fix. (Merged from his branch, via PR #42.) Updated |
| 130 | testdir/T.split per said PR as well. |
| 131 | |
Arnold D. Robbins | 7e368c2 | 2019-06-24 01:13:10 -0600 | [diff] [blame] | 132 | June 24, 2019: |
| 133 | Extract awktest.tar into testdir directory. Add some very |
| 134 | simple mechanics to the makefile for running the tests and |
| 135 | for cleaning up. No changes to awk itself. |
| 136 | |
Cody Mello | ae99b75 | 2019-06-17 10:08:54 -0900 | [diff] [blame] | 137 | June 17, 2019: |
| 138 | Disallow deleting SYMTAB and its elements, which creates |
| 139 | use-after-free bugs. Thanks to GitHub user Cody Mello (melloc) |
| 140 | for the fix. (Merged from PR #43.) |
| 141 | |
Arnold D. Robbins | 28dacbd | 2019-06-04 23:53:31 -0600 | [diff] [blame] | 142 | June 5, 2019: |
| 143 | Allow unmatched right parenthesis in a regular expression to |
| 144 | be treated literally. Fixes Issue #40. Thanks to GitHub user |
| 145 | Warner Losh (bsdimp) for the report. Thanks to Arnold Robbins |
| 146 | for the fix. |
| 147 | |
Arnold D. Robbins | 4189ef5 | 2019-05-29 21:04:18 +0300 | [diff] [blame] | 148 | May 29,2019: |
| 149 | Fix check for command line arguments to no longer require that |
| 150 | first character after '=' not be another '='. Reverts change of |
| 151 | August 11, 1989. Thanks to GitHub user Jamie Landeg Jones for |
| 152 | pointing out the issue; from Issue #38. |
| 153 | |
Arnold D. Robbins | 89354cc | 2019-04-07 20:39:15 +0300 | [diff] [blame] | 154 | Apr 7, 2019: |
| 155 | Update awktest.tar(p.50) to use modern options to sort. Needed |
| 156 | for Android development. Thanks to GitHub user mohd-akram (Mohamed |
Arnold D. Robbins | 28dacbd | 2019-06-04 23:53:31 -0600 | [diff] [blame] | 157 | Akram). From Issue #33. |
Arnold D. Robbins | 89354cc | 2019-04-07 20:39:15 +0300 | [diff] [blame] | 158 | |
Arnold D. Robbins | 55edb1b | 2019-03-12 21:54:57 +0200 | [diff] [blame] | 159 | Mar 12, 2019: |
| 160 | Added very simplistic support for cross-compiling in the |
| 161 | makefile. We are NOT going to go in the direction of the |
| 162 | autotools, though. Thanks to GitHub user nee-san for |
| 163 | the basic change. (Merged from PR #34.) |
| 164 | |
Martijn Dekker | e6ecf52 | 2019-03-05 03:45:40 +0100 | [diff] [blame] | 165 | Mar 5, 2019: |
Martijn Dekker | 8a22222 | 2019-01-23 09:12:27 +0000 | [diff] [blame] | 166 | Added support for POSIX-standard interval expressions (a.k.a. |
| 167 | bounds, a.k.a. repetition expressions) in regular expressions, |
| 168 | backported (via NetBSD) from Apple awk-24 (20070501). |
Arnold D. Robbins | 33ead6d | 2019-03-05 21:39:02 +0200 | [diff] [blame] | 169 | Thanks to Martijn Dekker <martijn@inlv.org> for the port. |
| 170 | (Merged from PR #30.) |
| 171 | |
| 172 | Mar 3, 2019: |
| 173 | Merge PRs as follows: |
| 174 | #12: Avoid undefined behaviour when using ctype(3) functions in |
| 175 | relex(). Thanks to GitHub user iamleot. |
| 176 | #31: Make getline handle numeric strings, and update FIXES. Thanks |
Arnold D. Robbins | 28dacbd | 2019-06-04 23:53:31 -0600 | [diff] [blame] | 177 | to GitHub user arnoldrobbins. |
Arnold D. Robbins | 33ead6d | 2019-03-05 21:39:02 +0200 | [diff] [blame] | 178 | #32: maketab: support build systems with read-only source. Thanks |
| 179 | to GitHub user enh. |
Martijn Dekker | 8a22222 | 2019-01-23 09:12:27 +0000 | [diff] [blame] | 180 | |
Arnold D. Robbins | 9206c64 | 2019-01-25 13:02:19 +0200 | [diff] [blame] | 181 | Jan 25, 2019: |
| 182 | Make getline handle numeric strings properly in all cases. |
| 183 | (Thanks, Arnold.) |
| 184 | |
| 185 | Jan 21, 2019: |
| 186 | Merged a number of small fixes from GitHub pull requests. |
| 187 | Thanks to GitHub users Arnold Robbins (arnoldrobbins), |
| 188 | Cody Mello (melloc) and Christoph Junghans (junghans). |
| 189 | PR numbers: 13-21, 23, 24, 27. |
| 190 | |
Brian Kernighan | e8c2800 | 2018-10-25 13:28:54 -0400 | [diff] [blame] | 191 | Oct 25, 2018: |
| 192 | Added test in maketab.c to prevent generating a proctab entry |
| 193 | for YYSTYPE_IS_DEFINED. It was harmless but some gcc settings |
| 194 | generated a warning message. Thanks to Nan Xiao for report. |
| 195 | |
Brian Kernighan | 22aff9e | 2018-08-27 08:52:34 -0400 | [diff] [blame] | 196 | Aug 27, 2018: |
| 197 | Disallow '$' in printf formats; arguments evaluated in order |
| 198 | and printed in order. |
| 199 | |
| 200 | Added some casts to silence warnings on debugging printfs. |
| 201 | (Thanks, Arnold.) |
| 202 | |
Brian Kernighan | 0f4e1ba | 2018-08-24 09:09:59 -0400 | [diff] [blame] | 203 | Aug 23, 2018: |
| 204 | A long list of fixes courtesy of Arnold Robbins, |
| 205 | to whom profound thanks. |
| 206 | |
| 207 | 1. ofs-rebuild: OFS value used to rebuild the record was incorrect. |
| 208 | Fixed August 19, 2014. Revised fix August 2018. |
| 209 | |
| 210 | 2. system-status: Instead of a floating-point division by 256, use |
| 211 | the wait(2) macros to create a reasonable exit status. |
| 212 | Fixed March 12, 2016. |
| 213 | |
| 214 | 3. space: Use provided xisblank() function instead of ispace() for |
| 215 | matching [[:blank:]]. |
| 216 | |
| 217 | 4. a-format: Add POSIX standard %a and %A to supported formats. Check |
| 218 | at runtime that this format is available. |
| 219 | |
| 220 | 5. decr-NF: Decrementing NF did not change $0. This is a decades-old |
| 221 | bug. There are interactions with the old and new value of OFS as well. |
| 222 | Most of the fix came from the NetBSD awk. |
| 223 | |
| 224 | 6. string-conv: String conversions of scalars were sticky. Once a |
| 225 | conversion to string happened, even with OFMT, that value was used until |
| 226 | a new numeric value was assigned, even if OFMT differed from CONVFMT, |
| 227 | and also if CONVFMT changed. |
| 228 | |
| 229 | 7. unary-plus: Unary plus on a string constant returned the string. |
| 230 | Instead, it should convert the value to numeric and give that value. |
| 231 | |
| 232 | Also added Arnold's tests for these to awktest.tar as T.arnold. |
| 233 | |
Brian Kernighan | ba7569c | 2018-08-16 09:41:13 -0400 | [diff] [blame] | 234 | Aug 15, 2018: |
| 235 | fixed mangled awktest.tar (thanks, Arnold), posted all |
| 236 | current (very minor) fixes to github / onetrueawk |
| 237 | |
Brian Kernighan | 3ed9e24 | 2018-08-15 10:45:03 -0400 | [diff] [blame] | 238 | Jun 7, 2018: |
| 239 | (yes, a long layoff) |
| 240 | Updated some broken tests (beebe.tar, T.lilly) |
| 241 | [thanks to Arnold Robbins] |
| 242 | |
| 243 | Mar 26, 2015: |
| 244 | buffer overflow in error reporting; thanks to tobias ulmer |
| 245 | and john-mark gurney for spotting it and the fix. |
| 246 | |
| 247 | Feb 4, 2013: |
| 248 | cleaned up a handful of tests that didn't seem to actually |
| 249 | test for correct behavior: T.latin1, T.gawk. |
| 250 | |
| 251 | Jan 5, 2013: |
| 252 | added ,NULL initializer to static Cells in run.c; not really |
| 253 | needed but cleaner. Thanks to Michael Bombardieri. |
| 254 | |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 255 | Dec 20, 2012: |
| 256 | fiddled makefile to get correct yacc and bison flags. pick yacc |
| 257 | (linux) or bison (mac) as necessary. |
| 258 | |
| 259 | added __attribute__((__noreturn__)) to a couple of lines in |
| 260 | proto.h, to silence someone's enthusiastic checker. |
| 261 | |
| 262 | fixed obscure call by value bug in split(a[1],a) reported on |
| 263 | 9fans. the management of temporary values is just a mess; i |
| 264 | took a shortcut by making an extra string copy. thanks |
| 265 | to paul patience and arnold robbins for passing it on and for |
| 266 | proposed patches. |
| 267 | |
| 268 | tiny fiddle in setfval to eliminate -0 results in T.expr, which |
| 269 | has irritated me for 20+ years. |
| 270 | |
| 271 | Aug 10, 2011: |
| 272 | another fix to avoid core dump with delete(ARGV); again, many thanks |
| 273 | to ruslan ermilov. |
| 274 | |
| 275 | Aug 7, 2011: |
| 276 | split(s, a, //) now behaves the same as split(s, a, "") |
| 277 | |
| 278 | Jun 12, 2011: |
| 279 | /pat/, \n /pat/ {...} is now legal, though bad style to use. |
| 280 | |
| 281 | added checks to new -v code that permits -vnospace; thanks to |
Arnold D. Robbins | 795a06b | 2019-07-28 05:51:52 -0600 | [diff] [blame] | 282 | ruslan ermilov for spotting this and providing the patch. |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 283 | |
| 284 | removed fixed limit on number of open files; thanks to aleksey |
Arnold D. Robbins | 795a06b | 2019-07-28 05:51:52 -0600 | [diff] [blame] | 285 | cheusov and christos zoulos. |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 286 | |
| 287 | fixed day 1 bug that resurrected deleted elements of ARGV when |
| 288 | used as filenames (in lib.c). |
| 289 | |
| 290 | minor type fiddles to make gcc -Wall -pedantic happier (but not |
| 291 | totally so); turned on -fno-strict-aliasing in makefile. |
| 292 | |
| 293 | May 6, 2011: |
| 294 | added #ifdef for isblank. |
| 295 | now allows -ffoo as well as -f foo arguments. |
| 296 | (thanks, ruslan) |
| 297 | |
| 298 | May 1, 2011: |
| 299 | after advice from todd miller, kevin lo, ruslan ermilov, |
| 300 | and arnold robbins, changed srand() to return the previous |
| 301 | seed (which is 1 on the first call of srand). the seed is |
| 302 | an Awkfloat internally though converted to unsigned int to |
Arnold D. Robbins | 795a06b | 2019-07-28 05:51:52 -0600 | [diff] [blame] | 303 | pass to the library srand(). thanks, everyone. |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 304 | |
| 305 | fixed a subtle (and i hope low-probability) overflow error |
Arnold D. Robbins | 795a06b | 2019-07-28 05:51:52 -0600 | [diff] [blame] | 306 | in fldbld, by adding space for one extra \0. thanks to |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 307 | robert bassett for spotting this one and providing a fix. |
| 308 | |
| 309 | removed the files related to compilation on windows. i no |
| 310 | longer have anything like a current windows environment, so |
| 311 | i can't test any of it. |
| 312 | |
| 313 | May 23, 2010: |
| 314 | fixed long-standing overflow bug in run.c; many thanks to |
| 315 | nelson beebe for spotting it and providing the fix. |
| 316 | |
| 317 | fixed bug that didn't parse -vd=1 properly; thanks to santiago |
| 318 | vila for spotting it. |
| 319 | |
| 320 | Feb 8, 2010: |
| 321 | i give up. replaced isblank with isspace in b.c; there are |
| 322 | no consistent header files. |
| 323 | |
| 324 | Nov 26, 2009: |
| 325 | fixed a long-standing issue with when FS takes effect. a |
| 326 | change to FS is now noticed immediately for subsequent splits. |
| 327 | |
| 328 | changed the name getline() to awkgetline() to avoid yet another |
| 329 | name conflict somewhere. |
| 330 | |
| 331 | Feb 11, 2009: |
| 332 | temporarily for now defined HAS_ISBLANK, since that seems to |
| 333 | be the best way through the thicket. isblank arrived in C99, |
| 334 | but seems to be arriving at different systems at different |
| 335 | times. |
| 336 | |
| 337 | Oct 8, 2008: |
| 338 | fixed typo in b.c that set tmpvec wrongly. no one had ever |
| 339 | run into the problem, apparently. thanks to alistair crooks. |
| 340 | |
| 341 | Oct 23, 2007: |
| 342 | minor fix in lib.c: increase inputFS to 100, change malloc |
Arnold D. Robbins | 795a06b | 2019-07-28 05:51:52 -0600 | [diff] [blame] | 343 | for fields to n+1. |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 344 | |
| 345 | fixed memory fault caused by out of order test in setsval. |
| 346 | |
| 347 | thanks to david o'brien, freebsd, for both fixes. |
| 348 | |
| 349 | May 1, 2007: |
| 350 | fiddle in makefile to fix for BSD make; thanks to igor sobrado. |
| 351 | |
| 352 | Mar 31, 2007: |
| 353 | fixed some null pointer refs calling adjbuf. |
| 354 | |
| 355 | Feb 21, 2007: |
| 356 | fixed a bug in matching the null RE in sub and gsub. thanks to al aho |
| 357 | who actually did the fix (in b.c), and to wolfgang seeberg for finding |
| 358 | it and providing a very compact test case. |
| 359 | |
| 360 | fixed quotation in b.c; thanks to Hal Pratt and the Princeton Dante |
| 361 | Project. |
| 362 | |
| 363 | removed some no-effect asserts in run.c. |
| 364 | |
| 365 | fiddled maketab.c to not complain about bison-generated values. |
| 366 | |
| 367 | removed the obsolete -V argument; fixed --version to print the |
| 368 | version and exit. |
| 369 | |
| 370 | fixed wording and an outright error in the usage message; thanks to igor |
| 371 | sobrado and jason mcintyre. |
| 372 | |
| 373 | fixed a bug in -d that caused core dump if no program followed. |
| 374 | |
| 375 | Jan 1, 2007: |
| 376 | dropped mac.code from makefile; there are few non-MacOSX |
| 377 | mac's these days. |
| 378 | |
| 379 | Jan 17, 2006: |
| 380 | system() not flagged as unsafe in the unadvertised -safe option. |
| 381 | found it while enhancing tests before shipping the ;login: article. |
| 382 | practice what you preach. |
| 383 | |
| 384 | removed the 9-years-obsolete -mr and -mf flags. |
| 385 | |
| 386 | added -version and --version options. |
| 387 | |
| 388 | core dump on linux with BEGIN {nextfile}, now fixed. |
| 389 | |
Arnold D. Robbins | 795a06b | 2019-07-28 05:51:52 -0600 | [diff] [blame] | 390 | removed some #ifdef's in run.c and lex.c that appear to no |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 391 | longer be necessary. |
| 392 | |
| 393 | Apr 24, 2005: |
| 394 | modified lib.c so that values of $0 et al are preserved in the END |
| 395 | block, apparently as required by posix. thanks to havard eidnes |
| 396 | for the report and code. |
| 397 | |
| 398 | Jan 14, 2005: |
| 399 | fixed infinite loop in parsing, originally found by brian tsang. |
| 400 | thanks to arnold robbins for a suggestion that started me |
| 401 | rethinking it. |
| 402 | |
| 403 | Dec 31, 2004: |
Arnold D. Robbins | 795a06b | 2019-07-28 05:51:52 -0600 | [diff] [blame] | 404 | prevent overflow of -f array in main, head off potential error in |
| 405 | call of SYNTAX(), test malloc return in lib.c, all with thanks to |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 406 | todd miller. |
| 407 | |
| 408 | Dec 22, 2004: |
| 409 | cranked up size of NCHARS; coverity thinks it can be overrun with |
| 410 | smaller size, and i think that's right. added some assertions to b.c |
| 411 | to catch places where it might overrun. the RE code is still fragile. |
| 412 | |
| 413 | Dec 5, 2004: |
| 414 | fixed a couple of overflow problems with ridiculous field numbers: |
| 415 | e.g., print $(2^32-1). thanks to ruslan ermilov, giorgos keramidas |
| 416 | and david o'brien at freebsd.org for patches. this really should |
| 417 | be re-done from scratch. |
| 418 | |
| 419 | Nov 21, 2004: |
| 420 | fixed another 25-year-old RE bug, in split. it's another failure |
| 421 | to (re-)initialize. thanks to steve fisher for spotting this and |
| 422 | providing a good test case. |
| 423 | |
| 424 | Nov 22, 2003: |
| 425 | fixed a bug in regular expressions that dates (so help me) from 1977; |
| 426 | it's been there from the beginning. an anchored longest match that |
| 427 | was longer than the number of states triggered a failure to initialize |
| 428 | the machine properly. many thanks to moinak ghosh for not only finding |
| 429 | this one but for providing a fix, in some of the most mysterious |
| 430 | code known to man. |
| 431 | |
| 432 | fixed a storage leak in call() that appears to have been there since |
Arnold D. Robbins | 795a06b | 2019-07-28 05:51:52 -0600 | [diff] [blame] | 433 | 1983 or so -- a function without an explicit return that assigns a |
| 434 | string to a parameter leaked a Cell. thanks to moinak ghosh for |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 435 | spotting this very subtle one. |
| 436 | |
| 437 | Jul 31, 2003: |
| 438 | fixed, thanks to andrey chernov and ruslan ermilov, a bug in lex.c |
| 439 | that mis-handled the character 255 in input. (it was being compared |
| 440 | to EOF with a signed comparison.) |
| 441 | |
| 442 | Jul 29, 2003: |
| 443 | fixed (i think) the long-standing botch that included the beginning of |
| 444 | line state ^ for RE's in the set of valid characters; this led to a |
| 445 | variety of odd problems, including failure to properly match certain |
| 446 | regular expressions in non-US locales. thanks to ruslan for keeping |
| 447 | at this one. |
| 448 | |
| 449 | Jul 28, 2003: |
| 450 | n-th try at getting internationalization right, with thanks to volker |
| 451 | kiefel, arnold robbins and ruslan ermilov for advice, though they |
| 452 | should not be blamed for the outcome. according to posix, "." is the |
| 453 | radix character in programs and command line arguments regardless of |
| 454 | the locale; otherwise, the locale should prevail for input and output |
| 455 | of numbers. so it's intended to work that way. |
Arnold D. Robbins | 795a06b | 2019-07-28 05:51:52 -0600 | [diff] [blame] | 456 | |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 457 | i have rescinded the attempt to use strcoll in expanding shorthands in |
| 458 | regular expressions (cclenter). its properties are much too |
| 459 | surprising; for example [a-c] matches aAbBc in locale en_US but abBcC |
| 460 | in locale fr_CA. i can see how this might arise by implementation |
| 461 | but i cannot explain it to a human user. (this behavior can be seen |
| 462 | in gawk as well; we're leaning on the same library.) |
| 463 | |
| 464 | the issue appears to be that strcoll is meant for sorting, where |
| 465 | merging upper and lower case may make sense (though note that unix |
| 466 | sort does not do this by default either). it is not appropriate |
| 467 | for regular expressions, where the goal is to match specific |
| 468 | patterns of characters. in any case, the notations [:lower:], etc., |
| 469 | are available in awk, and they are more likely to work correctly in |
| 470 | most locales. |
| 471 | |
| 472 | a moratorium is hereby declared on internationalization changes. |
| 473 | i apologize to friends and colleagues in other parts of the world. |
| 474 | i would truly like to get this "right", but i don't know what |
| 475 | that is, and i do not want to keep making changes until it's clear. |
| 476 | |
| 477 | Jul 4, 2003: |
| 478 | fixed bug that permitted non-terminated RE, as in "awk /x". |
| 479 | |
| 480 | Jun 1, 2003: |
| 481 | subtle change to split: if source is empty, number of elems |
| 482 | is always 0 and the array is not set. |
| 483 | |
| 484 | Mar 21, 2003: |
| 485 | added some parens to isblank, in another attempt to make things |
| 486 | internationally portable. |
| 487 | |
| 488 | Mar 14, 2003: |
| 489 | the internationalization changes, somewhat modified, are now |
| 490 | reinstated. in theory awk will now do character comparisons |
| 491 | and case conversions in national language, but "." will always |
| 492 | be the decimal point separator on input and output regardless |
| 493 | of national language. isblank(){} has an #ifndef. |
| 494 | |
| 495 | this no longer compiles on windows: LC_MESSAGES isn't defined |
| 496 | in vc6++. |
| 497 | |
| 498 | fixed subtle behavior in field and record splitting: if FS is |
| 499 | a single character and RS is not empty, \n is NOT a separator. |
| 500 | this tortuous reading is found in the awk book; behavior now |
| 501 | matches gawk and mawk. |
| 502 | |
| 503 | Dec 13, 2002: |
| 504 | for the moment, the internationalization changes of nov 29 are |
| 505 | rolled back -- programs like x = 1.2 don't work in some locales, |
| 506 | because the parser is expecting x = 1,2. until i understand this |
| 507 | better, this will have to wait. |
| 508 | |
| 509 | Nov 29, 2002: |
| 510 | modified b.c (with tiny changes in main and run) to support |
| 511 | locales, using strcoll and iswhatever tests for posix character |
| 512 | classes. thanks to ruslan ermilov (ru@freebsd.org) for code. |
| 513 | the function isblank doesn't seem to have propagated to any |
| 514 | header file near me, so it's there explicitly. not properly |
| 515 | tested on non-ascii character sets by me. |
| 516 | |
| 517 | Jun 28, 2002: |
| 518 | modified run/format() and tran/getsval() to do a slightly better |
| 519 | job on using OFMT for output from print and CONVFMT for other |
Arnold D. Robbins | 795a06b | 2019-07-28 05:51:52 -0600 | [diff] [blame] | 520 | number->string conversions, as promised by posix and done by |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 521 | gawk and mawk. there are still places where it doesn't work |
| 522 | right if CONVFMT is changed; by then the STR attribute of the |
| 523 | variable has been irrevocably set. thanks to arnold robbins for |
| 524 | code and examples. |
| 525 | |
| 526 | fixed subtle bug in format that could get core dump. thanks to |
| 527 | Jaromir Dolecek <jdolecek@NetBSD.org> for finding and fixing. |
| 528 | minor cleanup in run.c / format() at the same time. |
| 529 | |
| 530 | added some tests for null pointers to debugging printf's, which |
| 531 | were never intended for external consumption. thanks to dave |
| 532 | kerns (dkerns@lucent.com) for pointing this out. |
| 533 | |
| 534 | GNU compatibility: an empty regexp matches anything (thanks to |
| 535 | dag-erling smorgrav, des@ofug.org). subject to reversion if |
| 536 | this does more harm than good. |
| 537 | |
| 538 | pervasive small changes to make things more const-correct, as |
| 539 | reported by gcc's -Wwrite-strings. as it says in the gcc manual, |
| 540 | this may be more nuisance than useful. provoked by a suggestion |
| 541 | and code from arnaud desitter, arnaud@nimbus.geog.ox.ac.uk |
| 542 | |
| 543 | minor documentation changes to note that this now compiles out |
| 544 | of the box on Mac OS X. |
| 545 | |
| 546 | Feb 10, 2002: |
| 547 | changed types in posix chars structure to quiet solaris cc. |
| 548 | |
| 549 | Jan 1, 2002: |
| 550 | fflush() or fflush("") flushes all files and pipes. |
| 551 | |
Arnold D. Robbins | 795a06b | 2019-07-28 05:51:52 -0600 | [diff] [blame] | 552 | length(arrayname) returns number of elements; thanks to |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 553 | arnold robbins for suggestion. |
| 554 | |
| 555 | added a makefile.win to make it easier to build on windows. |
| 556 | based on dan allen's buildwin.bat. |
| 557 | |
| 558 | Nov 16, 2001: |
| 559 | added support for posix character class names like [:digit:], |
| 560 | which are not exactly shorter than [0-9] and perhaps no more |
| 561 | portable. thanks to dag-erling smorgrav for code. |
| 562 | |
| 563 | Feb 16, 2001: |
| 564 | removed -m option; no longer needed, and it was actually |
| 565 | broken (noted thanks to volker kiefel). |
| 566 | |
| 567 | Feb 10, 2001: |
| 568 | fixed an appalling bug in gettok: any sequence of digits, +,-, E, e, |
| 569 | and period was accepted as a valid number if it started with a period. |
| 570 | this would never have happened with the lex version. |
| 571 | |
| 572 | other 1-character botches, now fixed, include a bare $ and a |
| 573 | bare " at the end of the input. |
| 574 | |
| 575 | Feb 7, 2001: |
| 576 | more (const char *) casts in b.c and tran.c to silence warnings. |
| 577 | |
| 578 | Nov 15, 2000: |
| 579 | fixed a bug introduced in august 1997 that caused expressions |
| 580 | like $f[1] to be syntax errors. thanks to arnold robbins for |
| 581 | noticing this and providing a fix. |
| 582 | |
| 583 | Oct 30, 2000: |
| 584 | fixed some nextfile bugs: not handling all cases. thanks to |
| 585 | arnold robbins for pointing this out. new regressions added. |
| 586 | |
| 587 | close() is now a function. it returns whatever the library |
| 588 | fclose returns, and -1 for closing a file or pipe that wasn't |
| 589 | opened. |
| 590 | |
| 591 | Sep 24, 2000: |
| 592 | permit \n explicitly in character classes; won't work right |
| 593 | if comes in as "[\n]" but ok as /[\n]/, because of multiple |
| 594 | processing of \'s. thanks to arnold robbins. |
| 595 | |
| 596 | July 5, 2000: |
| 597 | minor fiddles in tran.c to keep compilers happy about uschar. |
| 598 | thanks to norman wilson. |
| 599 | |
| 600 | May 25, 2000: |
| 601 | yet another attempt at making 8-bit input work, with another |
Arnold D. Robbins | 795a06b | 2019-07-28 05:51:52 -0600 | [diff] [blame] | 602 | band-aid in b.c (member()), and some (uschar) casts to head |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 603 | off potential errors in subscripts (like isdigit). also |
| 604 | changed HAT to NCHARS-2. thanks again to santiago vila. |
| 605 | |
| 606 | changed maketab.c to ignore apparently out of range definitions |
| 607 | instead of halting; new freeBSD generates one. thanks to |
| 608 | jon snader <jsnader@ix.netcom.com> for pointing out the problem. |
| 609 | |
| 610 | May 2, 2000: |
| 611 | fixed an 8-bit problem in b.c by making several char*'s into |
| 612 | unsigned char*'s. not clear i have them all yet. thanks to |
| 613 | Santiago Vila <sanvila@unex.es> for the bug report. |
| 614 | |
| 615 | Apr 21, 2000: |
| 616 | finally found and fixed a memory leak in function call; it's |
| 617 | been there since functions were added ~1983. thanks to |
| 618 | jon bentley for the test case that found it. |
| 619 | |
| 620 | added test in envinit to catch environment "variables" with |
| 621 | names beginning with '='; thanks to Berend Hasselman. |
| 622 | |
| 623 | Jul 28, 1999: |
| 624 | added test in defn() to catch function foo(foo), which |
| 625 | otherwise recurses until core dump. thanks to arnold |
| 626 | robbins for noticing this. |
| 627 | |
| 628 | Jun 20, 1999: |
| 629 | added *bp in gettok in lex.c; appears possible to exit function |
| 630 | without terminating the string. thanks to russ cox. |
| 631 | |
| 632 | Jun 2, 1999: |
| 633 | added function stdinit() to run to initialize files[] array, |
| 634 | in case stdin, etc., are not constants; some compilers care. |
| 635 | |
| 636 | May 10, 1999: |
| 637 | replaced the ERROR ... FATAL, etc., macros with functions |
| 638 | based on vprintf, to avoid problems caused by overrunning |
| 639 | fixed-size errbuf array. thanks to ralph corderoy for the |
| 640 | impetus, and for pointing out a string termination bug in |
| 641 | qstring as well. |
| 642 | |
| 643 | Apr 21, 1999: |
| 644 | fixed bug that caused occasional core dumps with commandline |
| 645 | variable with value ending in \. (thanks to nelson beebe for |
| 646 | the test case.) |
| 647 | |
| 648 | Apr 16, 1999: |
Arnold D. Robbins | 795a06b | 2019-07-28 05:51:52 -0600 | [diff] [blame] | 649 | with code kindly provided by Bruce Lilly, awk now parses |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 650 | /=/ and similar constructs more sensibly in more places. |
| 651 | Bruce also provided some helpful test cases. |
| 652 | |
| 653 | Apr 5, 1999: |
| 654 | changed true/false to True/False in run.c to make it |
| 655 | easier to compile with C++. Added some casts on malloc |
| 656 | and realloc to be honest about casts; ditto. changed |
| 657 | ltype int to long in struct rrow to reduce some 64-bit |
| 658 | complaints; other changes scattered throughout for the |
| 659 | same purpose. thanks to Nelson Beebe for these portability |
| 660 | improvements. |
| 661 | |
| 662 | removed some horrible pointer-int casting in b.c and elsewhere |
| 663 | by adding ptoi and itonp to localize the casts, which are |
| 664 | all benign. fixed one incipient bug that showed up on sgi |
| 665 | in 64-bit mode. |
| 666 | |
| 667 | reset lineno for new source file; include filename in error |
| 668 | message. also fixed line number error in continuation lines. |
| 669 | (thanks to Nelson Beebe for both of these.) |
| 670 | |
| 671 | Mar 24, 1999: |
| 672 | Nelson Beebe notes that irix 5.3 yacc dies with a bogus |
| 673 | error; use a newer version or switch to bison, since sgi |
| 674 | is unlikely to fix it. |
| 675 | |
| 676 | Mar 5, 1999: |
| 677 | changed isnumber to is_number to avoid the problem caused by |
| 678 | versions of ctype.h that include the name isnumber. |
| 679 | |
| 680 | distribution now includes a script for building on a Mac, |
| 681 | thanks to Dan Allen. |
| 682 | |
| 683 | Feb 20, 1999: |
| 684 | fixed memory leaks in run.c (call) and tran.c (setfval). |
| 685 | thanks to Stephen Nutt for finding these and providing the fixes. |
| 686 | |
| 687 | Jan 13, 1999: |
| 688 | replaced srand argument by (unsigned int) in run.c; |
| 689 | avoids problem on Mac and potentially on Unix & Windows. |
| 690 | thanks to Dan Allen. |
| 691 | |
| 692 | added a few (int) casts to silence useless compiler warnings. |
| 693 | e.g., errorflag= in run.c jump(). |
| 694 | |
| 695 | added proctab.c to the bundle outout; one less thing |
| 696 | to have to compile out of the box. |
| 697 | |
| 698 | added calls to _popen and _pclose to the win95 stub for |
| 699 | pipes (thanks to Steve Adams for this helpful suggestion). |
| 700 | seems to work, though properties are not well understood |
| 701 | by me, and it appears that under some circumstances the |
| 702 | pipe output is truncated. Be careful. |
| 703 | |
| 704 | Oct 19, 1998: |
| 705 | fixed a couple of bugs in getrec: could fail to update $0 |
Arnold D. Robbins | 795a06b | 2019-07-28 05:51:52 -0600 | [diff] [blame] | 706 | after a getline var; because inputFS wasn't initialized, |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 707 | could split $0 on every character, a misleading diversion. |
| 708 | |
| 709 | fixed caching bug in makedfa: LRU was actually removing |
| 710 | least often used. |
| 711 | |
| 712 | thanks to ross ridge for finding these, and for providing |
| 713 | great bug reports. |
| 714 | |
| 715 | May 12, 1998: |
| 716 | fixed potential bug in readrec: might fail to update record |
| 717 | pointer after growing. thanks to dan levy for spotting this |
| 718 | and suggesting the fix. |
| 719 | |
| 720 | Mar 12, 1998: |
| 721 | added -V to print version number and die. |
| 722 | |
Arnold D. Robbins | 32093f5 | 2018-08-22 20:40:26 +0300 | [diff] [blame] | 723 | [notify dave kerns, dkerns@dacsoup.ih.lucent.com] |
| 724 | |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 725 | Feb 11, 1998: |
| 726 | subtle silent bug in lex.c: if the program ended with a number |
| 727 | longer than 1 digit, part of the input would be pushed back and |
| 728 | parsed again because token buffer wasn't terminated right. |
| 729 | example: awk 'length($0) > 10'. blush. at least i found it |
| 730 | myself. |
| 731 | |
| 732 | Aug 31, 1997: |
| 733 | s/adelete/awkdelete/: SGI uses this in malloc.h. |
| 734 | thanks to nelson beebe for pointing this one out. |
| 735 | |
| 736 | Aug 21, 1997: |
| 737 | fixed some bugs in sub and gsub when replacement includes \\. |
| 738 | this is a dark, horrible corner, but at least now i believe that |
| 739 | the behavior is the same as gawk and the intended posix standard. |
| 740 | thanks to arnold robbins for advice here. |
| 741 | |
| 742 | Aug 9, 1997: |
| 743 | somewhat regretfully, replaced the ancient lex-based lexical |
| 744 | analyzer with one written in C. it's longer, generates less code, |
| 745 | and more portable; the old one depended too much on mysterious |
| 746 | properties of lex that were not preserved in other environments. |
| 747 | in theory these recognize the same language. |
| 748 | |
| 749 | now using strtod to test whether a string is a number, instead of |
| 750 | the convoluted original function. should be more portable and |
| 751 | reliable if strtod is implemented right. |
| 752 | |
| 753 | removed now-pointless optimization in makefile that tries to avoid |
| 754 | recompilation when awkgram.y is changed but symbols are not. |
| 755 | |
| 756 | removed most fixed-size arrays, though a handful remain, some |
| 757 | of which are unchecked. you have been warned. |
| 758 | |
| 759 | Aug 4, 1997: |
| 760 | with some trepidation, replaced the ancient code that managed |
| 761 | fields and $0 in fixed-size arrays with arrays that grow on |
| 762 | demand. there is still some tension between trying to make this |
| 763 | run fast and making it clean; not sure it's right yet. |
| 764 | |
| 765 | the ill-conceived -mr and -mf arguments are now useful only |
| 766 | for debugging. previous dynamic string code removed. |
| 767 | |
| 768 | numerous other minor cleanups along the way. |
| 769 | |
| 770 | Jul 30, 1997: |
| 771 | using code provided by dan levy (to whom profuse thanks), replaced |
| 772 | fixed-size arrays and awkward kludges by a fairly uniform mechanism |
| 773 | to grow arrays as needed for printf, sub, gsub, etc. |
| 774 | |
| 775 | Jul 23, 1997: |
| 776 | falling off the end of a function returns "" and 0, not 0. |
| 777 | thanks to arnold robbins. |
| 778 | |
| 779 | Jun 17, 1997: |
| 780 | replaced several fixed-size arrays by dynamically-created ones |
| 781 | in run.c; added overflow tests to some previously unchecked cases. |
| 782 | getline, toupper, tolower. |
| 783 | |
| 784 | getline code is still broken in that recursive calls may wind |
| 785 | up using the same space. [fixed later] |
| 786 | |
| 787 | increased RECSIZE to 8192 to push problems further over the horizon. |
| 788 | |
| 789 | added \r to \n as input line separator for programs, not data. |
| 790 | damn CRLFs. |
| 791 | |
| 792 | modified format() to permit explicit printf("%c", 0) to include |
| 793 | a null byte in output. thanks to ken stailey for the fix. |
| 794 | |
| 795 | added a "-safe" argument that disables file output (print >, |
| 796 | print >>), process creation (cmd|getline, print |, system), and |
| 797 | access to the environment (ENVIRON). this is a first approximation |
| 798 | to a "safe" version of awk, but don't rely on it too much. thanks |
| 799 | to joan feigenbaum and matt blaze for the inspiration long ago. |
| 800 | |
| 801 | Jul 8, 1996: |
| 802 | fixed long-standing bug in sub, gsub(/a/, "\\\\&"); thanks to |
| 803 | ralph corderoy. |
| 804 | |
| 805 | Jun 29, 1996: |
| 806 | fixed awful bug in new field splitting; didn't get all the places |
| 807 | where input was done. |
| 808 | |
| 809 | Jun 28, 1996: |
| 810 | changed field-splitting to conform to posix definition: fields are |
| 811 | split using the value of FS at the time of input; it used to be |
| 812 | the value when the field or NF was first referred to, a much less |
| 813 | predictable definition. thanks to arnold robbins for encouragement |
| 814 | to do the right thing. |
| 815 | |
| 816 | May 28, 1996: |
| 817 | fixed appalling but apparently unimportant bug in parsing octal |
| 818 | numbers in reg exprs. |
| 819 | |
| 820 | explicit hex in reg exprs now limited to 2 chars: \xa, \xaa. |
| 821 | |
| 822 | May 27, 1996: |
| 823 | cleaned up some declarations so gcc -Wall is now almost silent. |
| 824 | |
| 825 | makefile now includes backup copies of ytab.c and lexyy.c in case |
| 826 | one makes before looking; it also avoids recreating lexyy.c unless |
| 827 | really needed. |
| 828 | |
| 829 | s/aprintf/awkprint, s/asprintf/awksprintf/ to avoid some name clashes |
| 830 | with unwisely-written header files. |
| 831 | |
| 832 | thanks to jeffrey friedl for several of these. |
| 833 | |
| 834 | May 26, 1996: |
| 835 | an attempt to rationalize the (unsigned) char issue. almost all |
| 836 | instances of unsigned char have been removed; the handful of places |
| 837 | in b.c where chars are used as table indices have been hand-crafted. |
| 838 | added some latin-1 tests to the regression, but i'm not confident; |
| 839 | none of my compilers seem to care much. thanks to nelson beebe for |
| 840 | pointing out some others that do care. |
| 841 | |
| 842 | May 2, 1996: |
| 843 | removed all register declarations. |
| 844 | |
| 845 | enhanced split(), as in gawk, etc: split(s, a, "") splits s into |
| 846 | a[1]...a[length(s)] with each character a single element. |
| 847 | |
| 848 | made the same changes for field-splitting if FS is "". |
| 849 | |
| 850 | added nextfile, as in gawk: causes immediate advance to next |
| 851 | input file. (thanks to arnold robbins for inspiration and code). |
| 852 | |
| 853 | small fixes to regexpr code: can now handle []], [[], and |
Arnold D. Robbins | 795a06b | 2019-07-28 05:51:52 -0600 | [diff] [blame] | 854 | variants; [] is now a syntax error, rather than matching |
Brian Kernighan | e37f882 | 2012-12-22 10:46:35 -0500 | [diff] [blame] | 855 | everything; [z-a] is now empty, not z. far from complete |
| 856 | or correct, however. (thanks to jeffrey friedl for pointing out |
| 857 | some awful behaviors.) |
| 858 | |
| 859 | Apr 29, 1996: |
| 860 | replaced uchar by uschar everywhere; apparently some compilers |
| 861 | usurp this name and this causes conflicts. |
| 862 | |
| 863 | fixed call to time in run.c (bltin); arg is time_t *. |
| 864 | |
| 865 | replaced horrible pointer/long punning in b.c by a legitimate |
| 866 | union. should be safer on 64-bit machines and cleaner everywhere. |
| 867 | (thanks to nelson beebe for pointing out some of these problems.) |
| 868 | |
| 869 | replaced nested comments by #if 0...#endif in run.c, lib.c. |
| 870 | |
| 871 | removed getsval, setsval, execute macros from run.c and lib.c. |
| 872 | machines are 100x faster than they were when these macros were |
| 873 | first used. |
| 874 | |
| 875 | revised filenames: awk.g.y => awkgram.y, awk.lx.l => awklex.l, |
| 876 | y.tab.[ch] => ytab.[ch], lex.yy.c => lexyy.c, all in the aid of |
| 877 | portability to nameless systems. |
| 878 | |
| 879 | "make bundle" now includes yacc and lex output files for recipients |
| 880 | who don't have yacc or lex. |
| 881 | |
| 882 | Aug 15, 1995: |
| 883 | initialized Cells in setsymtab more carefully; some fields |
| 884 | were not set. (thanks to purify, all of whose complaints i |
| 885 | think i now understand.) |
| 886 | |
| 887 | fixed at least one error in gsub that looked at -1-th element |
| 888 | of an array when substituting for a null match (e.g., $). |
| 889 | |
| 890 | delete arrayname is now legal; it clears the elements but leaves |
| 891 | the array, which may not be the right behavior. |
| 892 | |
| 893 | modified makefile: my current make can't cope with the test used |
| 894 | to avoid unnecessary yacc invocations. |
| 895 | |
| 896 | Jul 17, 1995: |
| 897 | added dynamically growing strings to awk.lx.l and b.c |
| 898 | to permit regular expressions to be much bigger. |
| 899 | the state arrays can still overflow. |
| 900 | |
| 901 | Aug 24, 1994: |
| 902 | detect duplicate arguments in function definitions (mdm). |
| 903 | |
| 904 | May 11, 1994: |
| 905 | trivial fix to printf to limit string size in sub(). |
| 906 | |
| 907 | Apr 22, 1994: |
| 908 | fixed yet another subtle self-assignment problem: |
| 909 | $1 = $2; $1 = $1 clobbered $1. |
| 910 | |
| 911 | Regression tests now use private echo, to avoid quoting problems. |
| 912 | |
| 913 | Feb 2, 1994: |
| 914 | changed error() to print line number as %d, not %g. |
| 915 | |
| 916 | Jul 23, 1993: |
| 917 | cosmetic changes: increased sizes of some arrays, |
| 918 | reworded some error messages. |
| 919 | |
| 920 | added CONVFMT as in posix (just replaced OFMT in getsval) |
| 921 | |
| 922 | FILENAME is now "" until the first thing that causes a file |
| 923 | to be opened. |
| 924 | |
| 925 | Nov 28, 1992: |
| 926 | deleted yyunput and yyoutput from proto.h; |
| 927 | different versions of lex give these different declarations. |
| 928 | |
| 929 | May 31, 1992: |
| 930 | added -mr N and -mf N options: more record and fields. |
| 931 | these really ought to adjust automatically. |
| 932 | |
| 933 | cleaned up some error messages; "out of space" now means |
| 934 | malloc returned NULL in all cases. |
| 935 | |
| 936 | changed rehash so that if it runs out, it just returns; |
| 937 | things will continue to run slow, but maybe a bit longer. |
| 938 | |
| 939 | Apr 24, 1992: |
| 940 | remove redundant close of stdin when using -f -. |
| 941 | |
| 942 | got rid of core dump with -d; awk -d just prints date. |
| 943 | |
| 944 | Apr 12, 1992: |
| 945 | added explicit check for /dev/std(in,out,err) in redirection. |
| 946 | unlike gawk, no /dev/fd/n yet. |
| 947 | |
| 948 | added (file/pipe) builtin. hard to test satisfactorily. |
| 949 | not posix. |
| 950 | |
| 951 | Feb 20, 1992: |
| 952 | recompile after abortive changes; should be unchanged. |
| 953 | |
| 954 | Dec 2, 1991: |
| 955 | die-casting time: converted to ansi C, installed that. |
| 956 | |
| 957 | Nov 30, 1991: |
| 958 | fixed storage leak in freefa, failing to recover [N]CCL. |
| 959 | thanks to Bill Jones (jones@cs.usask.ca) |
| 960 | |
| 961 | Nov 19, 1991: |
| 962 | use RAND_MAX instead of literal in builtin(). |
| 963 | |
| 964 | Nov 12, 1991: |
| 965 | cranked up some fixed-size arrays in b.c, and added a test for |
| 966 | overflow in penter. thanks to mark larsen. |
| 967 | |
| 968 | Sep 24, 1991: |
| 969 | increased buffer in gsub. a very crude fix to a general problem. |
| 970 | and again on Sep 26. |
| 971 | |
| 972 | Aug 18, 1991: |
| 973 | enforce variable name syntax for commandline variables: has to |
| 974 | start with letter or _. |
| 975 | |
| 976 | Jul 27, 1991: |
| 977 | allow newline after ; in for statements. |
| 978 | |
| 979 | Jul 21, 1991: |
| 980 | fixed so that in self-assignment like $1=$1, side effects |
| 981 | like recomputing $0 take place. (this is getting subtle.) |
| 982 | |
| 983 | Jun 30, 1991: |
| 984 | better test for detecting too-long output record. |
| 985 | |
| 986 | Jun 2, 1991: |
| 987 | better defense against very long printf strings. |
| 988 | made break and continue illegal outside of loops. |
| 989 | |
| 990 | May 13, 1991: |
| 991 | removed extra arg on gettemp, tempfree. minor error message rewording. |
| 992 | |
| 993 | May 6, 1991: |
| 994 | fixed silly bug in hex parsing in hexstr(). |
| 995 | removed an apparently unnecessary test in isnumber(). |
| 996 | warn about weird printf conversions. |
| 997 | fixed unchecked array overwrite in relex(). |
| 998 | |
| 999 | changed for (i in array) to access elements in sorted order. |
| 1000 | then unchanged it -- it really does run slower in too many cases. |
| 1001 | left the code in place, commented out. |
| 1002 | |
| 1003 | Feb 10, 1991: |
| 1004 | check error status on all writes, to avoid banging on full disks. |
| 1005 | |
| 1006 | Jan 28, 1991: |
| 1007 | awk -f - reads the program from stdin. |
| 1008 | |
| 1009 | Jan 11, 1991: |
| 1010 | failed to set numeric state on $0 in cmd|getline context in run.c. |
| 1011 | |
| 1012 | Nov 2, 1990: |
| 1013 | fixed sleazy test for integrality in getsval; use modf. |
| 1014 | |
| 1015 | Oct 29, 1990: |
| 1016 | fixed sleazy buggy code in lib.c that looked (incorrectly) for |
| 1017 | too long input lines. |
| 1018 | |
| 1019 | Oct 14, 1990: |
| 1020 | fixed the bug on p. 198 in which it couldn't deduce that an |
| 1021 | argument was an array in some contexts. replaced the error |
| 1022 | message in intest() by code that damn well makes it an array. |
| 1023 | |
| 1024 | Oct 8, 1990: |
| 1025 | fixed horrible bug: types and values were not preserved in |
| 1026 | some kinds of self-assignment. (in assign().) |
| 1027 | |
| 1028 | Aug 24, 1990: |
| 1029 | changed NCHARS to 256 to handle 8-bit characters in strings |
| 1030 | presented to match(), etc. |
| 1031 | |
| 1032 | Jun 26, 1990: |
| 1033 | changed struct rrow (awk.h) to use long instead of int for lval, |
| 1034 | since cfoll() stores a pointer in it. now works better when int's |
| 1035 | are smaller than pointers! |
| 1036 | |
| 1037 | May 6, 1990: |
| 1038 | AVA fixed the grammar so that ! is uniformly of the same precedence as |
| 1039 | unary + and -. This renders illegal some constructs like !x=y, which |
| 1040 | now has to be parenthesized as !(x=y), and makes others work properly: |
| 1041 | !x+y is (!x)+y, and x!y is x !y, not two pattern-action statements. |
| 1042 | (These problems were pointed out by Bob Lenk of Posix.) |
| 1043 | |
| 1044 | Added \x to regular expressions (already in strings). |
| 1045 | Limited octal to octal digits; \8 and \9 are not octal. |
| 1046 | Centralized the code for parsing escapes in regular expressions. |
| 1047 | Added a bunch of tests to T.re and T.sub to verify some of this. |
| 1048 | |
| 1049 | Feb 9, 1990: |
| 1050 | fixed null pointer dereference bug in main.c: -F[nothing]. sigh. |
| 1051 | |
| 1052 | restored srand behavior: it returns the current seed. |
| 1053 | |
| 1054 | Jan 18, 1990: |
| 1055 | srand now returns previous seed value (0 to start). |
| 1056 | |
| 1057 | Jan 5, 1990: |
| 1058 | fix potential problem in tran.c -- something was freed, |
| 1059 | then used in freesymtab. |
| 1060 | |
| 1061 | Oct 18, 1989: |
| 1062 | another try to get the max number of open files set with |
| 1063 | relatively machine-independent code. |
| 1064 | |
| 1065 | small fix to input() in case of multiple reads after EOF. |
| 1066 | |
| 1067 | Oct 11, 1989: |
| 1068 | FILENAME is now defined in the BEGIN block -- too many old |
| 1069 | programs broke. |
| 1070 | |
| 1071 | "-" means stdin in getline as well as on the commandline. |
| 1072 | |
| 1073 | added a bunch of casts to the code to tell the truth about |
| 1074 | char * vs. unsigned char *, a right royal pain. added a |
| 1075 | setlocale call to the front of main, though probably no one |
| 1076 | has it usefully implemented yet. |
| 1077 | |
| 1078 | Aug 24, 1989: |
| 1079 | removed redundant relational tests against nullnode if parse |
| 1080 | tree already had a relational at that point. |
| 1081 | |
| 1082 | Aug 11, 1989: |
| 1083 | fixed bug: commandline variable assignment has to look like |
| 1084 | var=something. (consider the man page for =, in file =.1) |
| 1085 | |
| 1086 | changed number of arguments to functions to static arrays |
| 1087 | to avoid repeated malloc calls. |
| 1088 | |
| 1089 | Aug 2, 1989: |
| 1090 | restored -F (space) separator |
| 1091 | |
| 1092 | Jul 30, 1989: |
| 1093 | added -v x=1 y=2 ... for immediate commandline variable assignment; |
| 1094 | done before the BEGIN block for sure. they have to precede the |
| 1095 | program if the program is on the commandline. |
| 1096 | Modified Aug 2 to require a separate -v for each assignment. |
| 1097 | |
| 1098 | Jul 10, 1989: |
| 1099 | fixed ref-thru-zero bug in environment code in tran.c |
| 1100 | |
| 1101 | Jun 23, 1989: |
| 1102 | add newline to usage message. |
| 1103 | |
| 1104 | Jun 14, 1989: |
| 1105 | added some missing ansi printf conversion letters: %i %X %E %G. |
| 1106 | no sensible meaning for h or L, so they may not do what one expects. |
| 1107 | |
| 1108 | made %* conversions work. |
| 1109 | |
| 1110 | changed x^y so that if n is a positive integer, it's done |
| 1111 | by explicit multiplication, thus achieving maximum accuracy. |
| 1112 | (this should be done by pow() but it seems not to be locally.) |
| 1113 | done to x ^= y as well. |
| 1114 | |
| 1115 | Jun 4, 1989: |
| 1116 | ENVIRON array contains environment: if shell variable V=thing, |
| 1117 | ENVIRON["V"] is "thing" |
| 1118 | |
| 1119 | multiple -f arguments permitted. error reporting is naive. |
| 1120 | (they were permitted before, but only the last was used.) |
| 1121 | |
| 1122 | fixed a really stupid botch in the debugging macro dprintf |
| 1123 | |
| 1124 | fixed order of evaluation of commandline assignments to match |
| 1125 | what the book claims: an argument of the form x=e is evaluated |
| 1126 | at the time it would have been opened if it were a filename (p 63). |
| 1127 | this invalidates the suggested answer to ex 4-1 (p 195). |
| 1128 | |
| 1129 | removed some code that permitted -F (space) fieldseparator, |
| 1130 | since it didn't quite work right anyway. (restored aug 2) |
| 1131 | |
| 1132 | Apr 27, 1989: |
| 1133 | Line number now accumulated correctly for comment lines. |
| 1134 | |
| 1135 | Apr 26, 1989: |
| 1136 | Debugging output now includes a version date, |
| 1137 | if one compiles it into the source each time. |
| 1138 | |
| 1139 | Apr 9, 1989: |
| 1140 | Changed grammar to prohibit constants as 3rd arg of sub and gsub; |
| 1141 | prevents class of overwriting-a-constant errors. (Last one?) |
| 1142 | This invalidates the "banana" example on page 43 of the book. |
| 1143 | |
| 1144 | Added \a ("alert"), \v (vertical tab), \xhhh (hexadecimal), |
| 1145 | as in ANSI, for strings. Rescinded the sloppiness that permitted |
| 1146 | non-octal digits in \ooo. Warning: not all compilers and libraries |
| 1147 | will be able to deal with \x correctly. |
| 1148 | |
| 1149 | Jan 9, 1989: |
| 1150 | Fixed bug that caused tempcell list to contain a duplicate. |
| 1151 | The fix is kludgy. |
| 1152 | |
| 1153 | Dec 17, 1988: |
| 1154 | Catches some more commandline errors in main. |
| 1155 | Removed redundant decl of modf in run.c (confuses some compilers). |
| 1156 | Warning: there's no single declaration of malloc, etc., in awk.h |
| 1157 | that seems to satisfy all compilers. |
| 1158 | |
| 1159 | Dec 7, 1988: |
| 1160 | Added a bit of code to error printing to avoid printing nulls. |
| 1161 | (Not clear that it actually would.) |
| 1162 | |
| 1163 | Nov 27, 1988: |
| 1164 | With fear and trembling, modified the grammar to permit |
| 1165 | multiple pattern-action statements on one line without |
| 1166 | an explicit separator. By definition, this capitulation |
| 1167 | to the ghost of ancient implementations remains undefined |
| 1168 | and thus subject to change without notice or apology. |
| 1169 | DO NOT COUNT ON IT. |
| 1170 | |
| 1171 | Oct 30, 1988: |
| 1172 | Fixed bug in call() that failed to recover storage. |
| 1173 | |
| 1174 | A warning is now generated if there are more arguments |
| 1175 | in the call than in the definition (in lieu of fixing |
| 1176 | another storage leak). |
| 1177 | |
| 1178 | Oct 20, 1988: |
| 1179 | Fixed %c: if expr is numeric, use numeric value; |
| 1180 | otherwise print 1st char of string value. still |
| 1181 | doesn't work if the value is 0 -- won't print \0. |
| 1182 | |
| 1183 | Added a few more checks for running out of malloc. |
| 1184 | |
| 1185 | Oct 12, 1988: |
| 1186 | Fixed bug in call() that freed local arrays twice. |
| 1187 | |
| 1188 | Fixed to handle deletion of non-existent array right; |
| 1189 | complains about attempt to delete non-array element. |
| 1190 | |
| 1191 | Sep 30, 1988: |
| 1192 | Now guarantees to evaluate all arguments of built-in |
| 1193 | functions, as in C; the appearance is that arguments |
| 1194 | are evaluated before the function is called. Places |
| 1195 | affected are sub (gsub was ok), substr, printf, and |
| 1196 | all the built-in arithmetic functions in bltin(). |
| 1197 | A warning is generated if a bltin() is called with |
| 1198 | the wrong number of arguments. |
| 1199 | |
| 1200 | This requires changing makeprof on p167 of the book. |
| 1201 | |
| 1202 | Aug 23, 1988: |
| 1203 | setting FILENAME in BEGIN caused core dump, apparently |
| 1204 | because it was freeing space not allocated by malloc. |
| 1205 | |
| 1206 | July 24, 1988: |
| 1207 | fixed egregious error in toupper/tolower functions. |
| 1208 | still subject to rescinding, however. |
| 1209 | |
| 1210 | July 2, 1988: |
| 1211 | flush stdout before opening file or pipe |
| 1212 | |
| 1213 | July 2, 1988: |
| 1214 | performance bug in b.c/cgoto(): not freeing some sets of states. |
| 1215 | partial fix only right now, and the number of states increased |
| 1216 | to make it less obvious. |
| 1217 | |
| 1218 | June 1, 1988: |
| 1219 | check error status on close |
| 1220 | |
| 1221 | May 28, 1988: |
| 1222 | srand returns seed value it's using. |
| 1223 | see 1/18/90 |
| 1224 | |
| 1225 | May 22, 1988: |
| 1226 | Removed limit on depth of function calls. |
| 1227 | |
| 1228 | May 10, 1988: |
| 1229 | Fixed lib.c to permit _ in commandline variable names. |
| 1230 | |
| 1231 | Mar 25, 1988: |
| 1232 | main.c fixed to recognize -- as terminator of command- |
| 1233 | line options. Illegal options flagged. |
| 1234 | Error reporting slightly cleaned up. |
| 1235 | |
| 1236 | Dec 2, 1987: |
| 1237 | Newer C compilers apply a strict scope rule to extern |
| 1238 | declarations within functions. Two extern declarations in |
| 1239 | lib.c and tran.c have been moved to obviate this problem. |
| 1240 | |
| 1241 | Oct xx, 1987: |
| 1242 | Reluctantly added toupper and tolower functions. |
| 1243 | Subject to rescinding without notice. |
| 1244 | |
| 1245 | Sep 17, 1987: |
| 1246 | Error-message printer had printf(s) instead of |
| 1247 | printf("%s",s); got core dumps when the message |
| 1248 | included a %. |
| 1249 | |
| 1250 | Sep 12, 1987: |
| 1251 | Very long printf strings caused core dump; |
| 1252 | fixed aprintf, asprintf, format to catch them. |
| 1253 | Can still get a core dump in printf itself. |
| 1254 | |
| 1255 | |