blob: 9494ac2deafcf0722ff1a7c11d9f862d0824ebfc
1 | #!/bin/sh |
2 | |
3 | # SUSv3 compliant sed tests. |
4 | # Copyright 2005 by Rob Landley <rob@landley.net> |
5 | # Licensed under GPLv2, see file LICENSE in this source tree. |
6 | |
7 | . ./testing.sh |
8 | |
9 | # testing "description" "commands" "result" "infile" "stdin" |
10 | |
11 | # Corner cases |
12 | testing "sed no files (stdin)" 'sed ""' "hello\n" "" "hello\n" |
13 | testing "sed explicit stdin" 'sed "" -' "hello\n" "" "hello\n" |
14 | testing "sed handles empty lines" "sed -e 's/\$/@/'" "@\n" "" "\n" |
15 | testing "sed stdin twice" 'sed "" - -' "hello" "" "hello" |
16 | |
17 | # Trailing EOF. |
18 | # Match $, at end of each file or all files? |
19 | |
20 | # -e corner cases |
21 | # without -e |
22 | # multiple -e |
23 | # interact with a |
24 | # -eee arg1 arg2 arg3 |
25 | # -f corner cases |
26 | # -e -f -e |
27 | # -n corner cases |
28 | # no newline at EOF? |
29 | # -r corner cases |
30 | # Just make sure it works. |
31 | # -i corner cases: |
32 | # sed -i - |
33 | # permissions |
34 | # -i on a symlink |
35 | # on a directory |
36 | # With $ last-line test |
37 | # Continue with \ |
38 | # End of script with trailing \ |
39 | |
40 | # command list |
41 | testing "sed accepts blanks before command" "sed -e '1 d'" "" "" "" |
42 | testing "sed accepts newlines in -e" "sed -e 'i\ |
43 | 1 |
44 | a\ |
45 | 3'" "1\n2\n3\n" "" "2\n" |
46 | testing "sed accepts multiple -e" "sed -e 'i\' -e '1' -e 'a\' -e '3'" \ |
47 | "1\n2\n3\n" "" "2\n" |
48 | |
49 | # substitutions |
50 | testing "sed -n" "sed -n -e s/foo/bar/ -e s/bar/baz/" "" "" "foo\n" |
51 | testing "sed with empty match" "sed 's/z*//g'" "string\n" "" "string\n" |
52 | testing "sed s//p" "sed -e s/foo/bar/p -e s/bar/baz/p" "bar\nbaz\nbaz\n" \ |
53 | "" "foo\n" |
54 | testing "sed -n s//p" "sed -ne s/abc/def/p" "def\n" "" "abc\n" |
55 | testing "sed s//g (exhaustive)" "sed -e 's/[[:space:]]*/,/g'" ",1,2,3,4,5,\n" \ |
56 | "" "12345\n" |
57 | testing "sed s arbitrary delimiter" "sed -e 's woo boing '" "boing\n" "" "woo\n" |
58 | testing "sed s chains" "sed -e s/foo/bar/ -e s/bar/baz/" "baz\n" "" "foo\n" |
59 | testing "sed s chains2" "sed -e s/foo/bar/ -e s/baz/nee/" "bar\n" "" "foo\n" |
60 | testing "sed s [delimiter]" "sed -e 's@[@]@@'" "onetwo" "" "one@two" |
61 | testing "sed s with \\t (GNU ext)" "sed 's/\t/ /'" "one two" "" "one\ttwo" |
62 | |
63 | # branch |
64 | testing "sed b (branch)" "sed -e 'b one;p;: one'" "foo\n" "" "foo\n" |
65 | testing "sed b (branch with no label jumps to end)" "sed -e 'b;p'" \ |
66 | "foo\n" "" "foo\n" |
67 | |
68 | # test and branch |
69 | testing "sed t (test/branch)" "sed -e 's/a/1/;t one;p;: one;p'" \ |
70 | "1\n1\nb\nb\nb\nc\nc\nc\n" "" "a\nb\nc\n" |
71 | testing "sed t (test/branch clears test bit)" "sed -e 's/a/b/;:loop;t loop'" \ |
72 | "b\nb\nc\n" "" "a\nb\nc\n" |
73 | testing "sed T (!test/branch)" "sed -e 's/a/1/;T notone;p;: notone;p'" \ |
74 | "1\n1\n1\nb\nb\nc\nc\n" "" "a\nb\nc\n" |
75 | |
76 | test x"$SKIP_KNOWN_BUGS" = x"" && { |
77 | # Normal sed end-of-script doesn't print "c" because n flushed the pattern |
78 | # space. If n hits EOF, pattern space is empty when script ends. |
79 | # Query: how does this interact with no newline at EOF? |
80 | testing "sed n (flushes pattern space, terminates early)" "sed -e 'n;p'" \ |
81 | "a\nb\nb\nc\n" "" "a\nb\nc\n" |
82 | } |
83 | # non-GNU sed: N does _not_ flush pattern space, therefore c is eaten @ script end |
84 | # GNU sed: N flushes pattern space, therefore c is printed too @ script end |
85 | testing "sed N (flushes pattern space (GNU behavior))" "sed -e 'N;p'" \ |
86 | "a\nb\na\nb\nc\n" "" "a\nb\nc\n" |
87 | |
88 | testing "sed N test2" "sed ':a;N;s/\n/ /;ta'" \ |
89 | "a b c\n" "" "a\nb\nc\n" |
90 | |
91 | testing "sed N test3" "sed 'N;s/\n/ /'" \ |
92 | "a b\nc\n" "" "a\nb\nc\n" |
93 | |
94 | testing "sed address match newline" 'sed "/b/N;/b\\nc/i woo"' \ |
95 | "a\nwoo\nb\nc\nd\n" "" "a\nb\nc\nd\n" |
96 | |
97 | # Multiple lines in pattern space |
98 | testing "sed N (stops at end of input) and P (prints to first newline only)" \ |
99 | "sed -n 'N;P;p'" "a\na\nb\n" "" "a\nb\nc\n" |
100 | |
101 | # Hold space |
102 | testing "sed G (append hold space to pattern space)" 'sed G' "a\n\nb\n\nc\n\n" \ |
103 | "" "a\nb\nc\n" |
104 | #testing "sed g/G (swap/append hold and patter space)" |
105 | #testing "sed g (swap hold/pattern space)" |
106 | |
107 | testing "sed d ends script iteration" \ |
108 | "sed -e '/ook/d;s/ook/ping/p;i woot'" "" "" "ook\n" |
109 | testing "sed d ends script iteration (2)" \ |
110 | "sed -e '/ook/d;a\' -e 'bang'" "woot\nbang\n" "" "ook\nwoot\n" |
111 | |
112 | # Multiple files, with varying newlines and NUL bytes |
113 | test x"$SKIP_KNOWN_BUGS" = x"" && { |
114 | testing "sed embedded NUL" "sed -e 's/woo/bang/'" "\0bang\0woo\0" "" \ |
115 | "\0woo\0woo\0" |
116 | } |
117 | testing "sed embedded NUL g" "sed -e 's/woo/bang/g'" "bang\0bang\0" "" \ |
118 | "woo\0woo\0" |
119 | test x"$SKIP_KNOWN_BUGS" = x"" && { |
120 | $ECHO -e "/woo/a he\0llo" > sed.commands |
121 | testing "sed NUL in command" "sed -f sed.commands" "woo\nhe\0llo\n" "" "woo" |
122 | rm sed.commands |
123 | } |
124 | |
125 | # sed has funky behavior with newlines at the end of file. Test lots of |
126 | # corner cases with the optional newline appending behavior. |
127 | |
128 | testing "sed normal newlines" "sed -e 's/woo/bang/' input -" "bang\nbang\n" \ |
129 | "woo\n" "woo\n" |
130 | testing "sed leave off trailing newline" "sed -e 's/woo/bang/' input -" \ |
131 | "bang\nbang" "woo\n" "woo" |
132 | testing "sed autoinsert newline" "sed -e 's/woo/bang/' input -" "bang\nbang" \ |
133 | "woo" "woo" |
134 | testing "sed empty file plus cat" "sed -e 's/nohit//' input -" "one\ntwo" \ |
135 | "" "one\ntwo" |
136 | testing "sed cat plus empty file" "sed -e 's/nohit//' input -" "one\ntwo" \ |
137 | "one\ntwo" "" |
138 | test x"$SKIP_KNOWN_BUGS" = x"" && { |
139 | testing "sed append autoinserts newline" "sed -e '/woot/a woo' -" \ |
140 | "woot\nwoo\n" "" "woot" |
141 | } |
142 | testing "sed insert doesn't autoinsert newline" "sed -e '/woot/i woo' -" \ |
143 | "woo\nwoot" "" "woot" |
144 | testing "sed print autoinsert newlines" "sed -e 'p' -" "one\none" "" "one" |
145 | testing "sed print autoinsert newlines two files" "sed -e 'p' input -" \ |
146 | "one\none\ntwo\ntwo" "one" "two" |
147 | testing "sed noprint, no match, no newline" "sed -ne 's/woo/bang/' input" \ |
148 | "" "no\n" "" |
149 | testing "sed selective matches with one nl" "sed -ne 's/woo/bang/p' input -" \ |
150 | "a bang\nc bang\n" "a woo\nb no" "c woo\nd no" |
151 | testing "sed selective matches insert newline" \ |
152 | "sed -ne 's/woo/bang/p' input -" "a bang\nb bang\nd bang" \ |
153 | "a woo\nb woo" "c no\nd woo" |
154 | testing "sed selective matches noinsert newline" \ |
155 | "sed -ne 's/woo/bang/p' input -" "a bang\nb bang" "a woo\nb woo" \ |
156 | "c no\nd no" |
157 | testing "sed clusternewline" \ |
158 | "sed -e '/one/a 111' -e '/two/i 222' -e p input -" \ |
159 | "one\none\n111\n222\ntwo\ntwo" "one" "two" |
160 | testing "sed subst+write" \ |
161 | "sed -e 's/i/z/' -e 'woutputw' input -; $ECHO -n X; cat outputw" \ |
162 | "thzngy\nagaznXthzngy\nagazn" "thingy" "again" |
163 | rm outputw |
164 | testing "sed trailing NUL" \ |
165 | "sed 's/i/z/' input -" \ |
166 | "a\0b\0\nc" "a\0b\0" "c" |
167 | testing "sed escaped newline in command" \ |
168 | "sed 's/a/z\\ |
169 | z/' input" \ |
170 | "z\nz" "a" "" |
171 | |
172 | # Test end-of-file matching behavior |
173 | |
174 | testing "sed match EOF" "sed -e '"'$p'"'" "hello\nthere\nthere" "" \ |
175 | "hello\nthere" |
176 | testing "sed match EOF two files" "sed -e '"'$p'"' input -" \ |
177 | "one\ntwo\nthree\nfour\nfour" "one\ntwo" "three\nfour" |
178 | # sed match EOF inline: gnu sed 4.1.5 outputs this: |
179 | #00000000 6f 6e 65 0a 6f 6f 6b 0a 6f 6f 6b 0a 74 77 6f 0a |one.ook.ook.two.| |
180 | #00000010 0a 74 68 72 65 65 0a 6f 6f 6b 0a 6f 6f 6b 0a 66 |.three.ook.ook.f| |
181 | #00000020 6f 75 72 |our| |
182 | # which looks buggy to me. |
183 | $ECHO -ne "three\nfour" > input2 |
184 | testing "sed match EOF inline" \ |
185 | "sed -e '"'$i ook'"' -i input input2 && cat input input2" \ |
186 | "one\nook\ntwothree\nook\nfour" "one\ntwo" "" |
187 | rm input2 |
188 | |
189 | # Test lie-to-autoconf |
190 | |
191 | testing "sed lie-to-autoconf" "sed --version | grep -o 'GNU sed version '" \ |
192 | "GNU sed version \n" "" "" |
193 | |
194 | # Jump to nonexistent label |
195 | test x"$SKIP_KNOWN_BUGS" = x"" && { |
196 | # Incompatibility: illegal jump is not detected if input is "" |
197 | # (that is, no lines at all). GNU sed 4.1.5 complains even in this case |
198 | testing "sed nonexistent label" "sed -e 'b walrus' 2>/dev/null || echo yes" \ |
199 | "yes\n" "" "" |
200 | } |
201 | |
202 | testing "sed backref from empty s uses range regex" \ |
203 | "sed -e '/woot/s//eep \0 eep/'" "eep woot eep" "" "woot" |
204 | |
205 | testing "sed backref from empty s uses range regex with newline" \ |
206 | "sed -e '/woot/s//eep \0 eep/'" "eep woot eep\n" "" "woot\n" |
207 | |
208 | # -i with no filename |
209 | |
210 | touch ./- # Detect gnu failure mode here. |
211 | testing "sed -i with no arg [GNUFAIL]" "sed -e '' -i 2> /dev/null || echo yes" \ |
212 | "yes\n" "" "" |
213 | rm ./- # Clean up |
214 | |
215 | testing "sed s/xxx/[/" "sed -e 's/xxx/[/'" "[\n" "" "xxx\n" |
216 | |
217 | # Ponder this a bit more, why "woo not found" from gnu version? |
218 | #testing "sed doesn't substitute in deleted line" \ |
219 | # "sed -e '/ook/d;s/ook//;t woo;a bang;'" "bang" "" "ook\n" |
220 | |
221 | # This makes both seds very unhappy. Why? |
222 | #testing "sed -g (exhaustive)" "sed -e 's/[[:space:]]*/,/g'" ",1,2,3,4,5," \ |
223 | # "" "12345" |
224 | |
225 | # testing "description" "commands" "result" "infile" "stdin" |
226 | |
227 | testing "sed n command must reset 'substituted' bit" \ |
228 | "sed 's/1/x/;T;n;: next;s/3/y/;t quit;n;b next;: quit;q'" \ |
229 | "0\nx\n2\ny\n" "" "0\n1\n2\n3\n" |
230 | |
231 | testing "sed d does not break n,m matching" \ |
232 | "sed -n '1d;1,3p'" \ |
233 | "second\nthird\n" "" "first\nsecond\nthird\nfourth\n" |
234 | |
235 | testing "sed d does not break n,regex matching" \ |
236 | "sed -n '1d;1,/hir/p'" \ |
237 | "second\nthird\n" "" "first\nsecond\nthird\nfourth\n" |
238 | |
239 | testing "sed d does not break n,regex matching #2" \ |
240 | "sed -n '1,5d;1,/hir/p'" \ |
241 | "second2\nthird2\n" "" \ |
242 | "first\nsecond\nthird\nfourth\n""first2\nsecond2\nthird2\nfourth2\n" |
243 | |
244 | testing "sed 2d;2,1p (gnu compat)" \ |
245 | "sed -n '2d;2,1p'" \ |
246 | "third\n" "" \ |
247 | "first\nsecond\nthird\nfourth\n" |
248 | |
249 | # Regex means: "match / at BOL or nothing, then one or more not-slashes". |
250 | # The bug was that second slash in /usr/lib was treated as "at BOL" too. |
251 | testing "sed beginning (^) matches only once" \ |
252 | "sed 's,\(^/\|\)[^/][^/]*,>\0<,g'" \ |
253 | ">/usr</>lib<\n" "" \ |
254 | "/usr/lib\n" |
255 | |
256 | testing "sed c" \ |
257 | "sed 'crepl'" \ |
258 | "repl\nrepl\n" "" \ |
259 | "first\nsecond\n" |
260 | |
261 | testing "sed nested {}s" \ |
262 | "sed '/asd/ { p; /s/ { s/s/c/ }; p; q }'" \ |
263 | "qwe\nasd\nacd\nacd\n" "" \ |
264 | "qwe\nasd\nzxc\n" |
265 | |
266 | testing "sed a cmd ended by double backslash" \ |
267 | "sed -e '/| one /a \\ |
268 | | three \\\\' -e '/| one-/a \\ |
269 | | three-* \\\\'" \ |
270 | ' | one \\ |
271 | | three \\ |
272 | | two \\ |
273 | ' '' \ |
274 | ' | one \\ |
275 | | two \\ |
276 | ' |
277 | |
278 | # first three lines are deleted; 4th line is matched and printed by "2,3" and by "4" ranges |
279 | testing "sed with N skipping lines past ranges on next cmds" \ |
280 | "sed -n '1{N;N;d};1p;2,3p;3p;4p'" \ |
281 | "4\n4\n" "" "1\n2\n3\n4\n" |
282 | |
283 | testing "sed -i with address modifies all files, not only first" \ |
284 | "cp input input2; sed -i -e '1s/foo/bar/' input input2 && cat input input2; rm input2" \ |
285 | "bar\nbar\n" "foo\n" "" |
286 | |
287 | testing "sed understands \r" \ |
288 | "sed 's/r/\r/'" \ |
289 | "\rrr\n" "" "rrr\n" |
290 | |
291 | testing "sed -i finishes ranges correctly" \ |
292 | "sed '1,2d' -i input; echo \$?; cat input" \ |
293 | "0\n3\n4\n" "1\n2\n3\n4\n" "" |
294 | |
295 | testing "sed zero chars match/replace advances correctly 1" \ |
296 | "sed 's/l*/@/g'" \ |
297 | "@h@e@o@\n" "" "helllo\n" |
298 | |
299 | testing "sed zero chars match/replace advances correctly 2" \ |
300 | "sed 's [^ .]* x g'" \ |
301 | "x x.x\n" "" " a.b\n" |
302 | |
303 | testing "sed zero chars match/replace logic must not falsely trigger here 1" \ |
304 | "sed 's/a/A/g'" \ |
305 | "_AAA1AA\n" "" "_aaa1aa\n" |
306 | |
307 | testing "sed zero chars match/replace logic must not falsely trigger here 2" \ |
308 | "sed 's/ *$/_/g'" \ |
309 | "qwerty_\n" "" "qwerty\n" |
310 | |
311 | testing "sed /\$_in_regex/ should not match newlines, only end-of-line" \ |
312 | "sed ': testcont; /\\\\$/{ =; N; b testcont }'" \ |
313 | "\ |
314 | this is a regular line |
315 | 2 |
316 | line with \\ |
317 | continuation |
318 | more regular lines |
319 | 5 |
320 | line with \\ |
321 | continuation |
322 | " \ |
323 | "" "\ |
324 | this is a regular line |
325 | line with \\ |
326 | continuation |
327 | more regular lines |
328 | line with \\ |
329 | continuation |
330 | " |
331 | |
332 | # testing "description" "commands" "result" "infile" "stdin" |
333 | |
334 | exit $FAILCOUNT |
335 |