BusyBox
/
busybox
mirror of https://github.com/mirror/busybox.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

sed: fix 2 bugs (one testsuite entry + one newly found) 

but more importantly make code more understandable
1_5_stable
Denis Vlasenko 2007-01-29 14:31:47 +00:00
parent 826c85f382
commit fe7a9f1277
2 changed files with 63 additions and 59 deletions
editors
testsuite

118
editors/sed.c
View File

@ -72,7 +72,7 @@ typedef struct sed_cmd_s {
int beg_line; /* 'sed 1p' 0 == apply commands to all lines */ int beg_line; /* 'sed 1p' 0 == apply commands to all lines */
int end_line; /* 'sed 1,3p' 0 == one line only. -1 = last line ($) */ int end_line; /* 'sed 1,3p' 0 == one line only. -1 = last line ($) */
FILE *file; /* File (sw) command writes to, -1 for none. */ FILE *sw_file; /* File (sw) command writes to, -1 for none. */
char *string; /* Data string for (saicytb) commands. */ char *string; /* Data string for (saicytb) commands. */
unsigned short which_match; /* (s) Which match to replace (0 for all) */ unsigned short which_match; /* (s) Which match to replace (0 for all) */
@ -82,7 +82,7 @@ typedef struct sed_cmd_s {
unsigned int in_match:1; /* Next line also included in match? */ unsigned int in_match:1; /* Next line also included in match? */
unsigned int sub_p:1; /* (s) print option */ unsigned int sub_p:1; /* (s) print option */
// int sw_last_char; /* Last line written by (sw) had no '\n' */ char sw_last_char; /* Last line written by (sw) had no '\n' */
/* GENERAL FIELDS */ /* GENERAL FIELDS */
char cmd; /* The command char: abcdDgGhHilnNpPqrstwxy:={} */ char cmd; /* The command char: abcdDgGhHilnNpPqrstwxy:={} */
@ -130,8 +130,8 @@ static void sed_free_and_close_stuff(void)
while (sed_cmd) { while (sed_cmd) {
sed_cmd_t *sed_cmd_next = sed_cmd->next; sed_cmd_t *sed_cmd_next = sed_cmd->next;
if (sed_cmd->file) if (sed_cmd->sw_file)
xprint_and_close_file(sed_cmd->file); xprint_and_close_file(sed_cmd->sw_file);
if (sed_cmd->beg_match) { if (sed_cmd->beg_match) {
regfree(sed_cmd->beg_match); regfree(sed_cmd->beg_match);
@ -423,7 +423,7 @@ static char *parse_cmd_args(sed_cmd_t *sed_cmd, char *cmdstr)
bb_error_msg_and_die("command only uses one address"); bb_error_msg_and_die("command only uses one address");
cmdstr += parse_file_cmd(sed_cmd, cmdstr, &sed_cmd->string); cmdstr += parse_file_cmd(sed_cmd, cmdstr, &sed_cmd->string);
if (sed_cmd->cmd == 'w') if (sed_cmd->cmd == 'w')
sed_cmd->file = xfopen(sed_cmd->string, "w"); sed_cmd->sw_file = xfopen(sed_cmd->string, "w");
/* handle branch commands */ /* handle branch commands */
} else if (strchr(":btT", sed_cmd->cmd)) { } else if (strchr(":btT", sed_cmd->cmd)) {
int length; int length;
@ -717,21 +717,22 @@ static void add_input_file(FILE *file)
bbg.input_file_list[bbg.input_file_count++] = file; bbg.input_file_list[bbg.input_file_count++] = file;
} }
enum {
MASK_NO_EOL_CHAR = 0x100,
//MASK_FIRST_LINE = 0x200,
};
/* Get next line of input from bbg.input_file_list, flushing append buffer and /* Get next line of input from bbg.input_file_list, flushing append buffer and
* noting if we ran out of files without a newline on the last line we read. * noting if we ran out of files without a newline on the last line we read.
*/ */
static char *get_next_line(int *last_char) enum {
NO_EOL_CHAR = 1,
};
static char *get_next_line(char *gets_char)
{ {
char *temp = NULL; char *temp = NULL;
int len, lc; int len, gc;
lc = 0;
flush_append(); flush_append();
/* will be returned if last line in the file
* doesn't end with either '\n' or '\0' */
gc = NO_EOL_CHAR;
while (bbg.current_input_file < bbg.input_file_count) { while (bbg.current_input_file < bbg.input_file_count) {
/* Read line up to a newline or NUL byte, inclusive, /* Read line up to a newline or NUL byte, inclusive,
* return malloc'ed char[]. length of the chunk read * return malloc'ed char[]. length of the chunk read
@ -743,26 +744,29 @@ static char *get_next_line(int *last_char)
char c = temp[len-1]; char c = temp[len-1];
if (c == '\n' || c == '\0') { if (c == '\n' || c == '\0') {
temp[len-1] = '\0'; temp[len-1] = '\0';
lc |= (unsigned char)c; gc = (unsigned char)c;
break; break;
} }
/* will be returned if last line in the file /* NB: I had the idea of peeking next file(s) and returning
* doesn't end with either '\n' or '\0' */ * NO_EOL_CHAR only if it is the *last* non-empty
lc |= MASK_NO_EOL_CHAR; * input file. But there is a case where this won't work:
* file1: "a woo\nb woo"
* file2: "c no\nd no"
* sed -ne 's/woo/bang/p' input1 input2 => "a bang\nb bang"
* (note: *no* newline after "b bang"!) */
break; break;
} }
/* Close this file and advance to next one */ /* Close this file and advance to next one */
fclose(bbg.input_file_list[bbg.current_input_file++]); fclose(bbg.input_file_list[bbg.current_input_file++]);
/* "this is the first line from new input file" */
//lc |= MASK_FIRST_LINE;
} }
*last_char = lc; *gets_char = gc;
return temp; return temp;
} }
/* Output line of text. */ /* Output line of text. */
/* Note: /* Note:
* The tricks with MASK_FIRST_LINE and last_puts_char are there to emulate gnu sed. * The tricks with NO_EOL_CHAR and last_puts_char are there to emulate gnu sed.
* Without them, we had this: * Without them, we had this:
* echo -n thingy >z1 * echo -n thingy >z1
* echo -n again >z2 * echo -n again >z2
@ -774,37 +778,32 @@ static char *get_next_line(int *last_char)
* bbox: * bbox:
* 00000000 74 68 7a 6e 67 79 61 67 61 7a 6e |thzngyagazn| * 00000000 74 68 7a 6e 67 79 61 67 61 7a 6e |thzngyagazn|
*/ */
static void puts_maybe_newline(char *s, FILE *file, int last_char) static void puts_maybe_newline(char *s, FILE *file, char *last_puts_char, char last_gets_char)
{ {
static char last_puts_char = '\n'; char lpc = *last_puts_char;
/* Is this a first line from new file /* Is this a first line from new file
* and old file didn't end with '\n' or '\0'? */ * and old file didn't end with '\n' or '\0'? */
// if ((last_char & MASK_FIRST_LINE) && last_puts_char != '\n') { if (lpc != '\n' && lpc != '\0') {
if (last_puts_char != '\n' && last_puts_char != '\0') {
fputc('\n', file); fputc('\n', file);
last_puts_char = '\n'; lpc = '\n';
} }
fputs(s, file); fputs(s, file);
/* why 'x'? - just something which is not '\n' */ /* 'x' - just something which is not '\n', '\0' or NO_EOL_CHAR */
if (s[0]) if (s[0])
last_puts_char = 'x'; lpc = 'x';
if (!(last_char & MASK_NO_EOL_CHAR)) { /* had trailing '\n' or '\0'? */ if (last_gets_char != NO_EOL_CHAR) { /* had trailing '\n' or '\0'? */
last_char &= 0xff; fputc(last_gets_char, file);
fputc(last_char, file); lpc = last_gets_char;
last_puts_char = last_char;
} }
if (ferror(file)) { if (ferror(file)) {
xfunc_error_retval = 4; /* It's what gnu sed exits with... */ xfunc_error_retval = 4; /* It's what gnu sed exits with... */
bb_error_msg_and_die(bb_msg_write_error); bb_error_msg_and_die(bb_msg_write_error);
} }
*last_puts_char = lpc;
/* Seems to be unused */
/*return last_char;*/
} }
#define sed_puts(s, n) (puts_maybe_newline(s, bbg.nonstdout, n)) #define sed_puts(s, n) (puts_maybe_newline(s, bbg.nonstdout, &last_puts_char, n))
/* Process all the lines in all the files */ /* Process all the lines in all the files */
@ -812,12 +811,13 @@ static void process_files(void)
{ {
char *pattern_space, *next_line; char *pattern_space, *next_line;
int linenum = 0; int linenum = 0;
int last_char, next_last_char = 0; char last_puts_char = '\n';
char last_gets_char, next_gets_char;
sed_cmd_t *sed_cmd; sed_cmd_t *sed_cmd;
int substituted; int substituted;
/* Prime the pump */ /* Prime the pump */
next_line = get_next_line(&next_last_char); next_line = get_next_line(&next_gets_char);
/* go through every line in each file */ /* go through every line in each file */
again: again:
@ -826,11 +826,11 @@ again:
/* Advance to next line. Stop if out of lines. */ /* Advance to next line. Stop if out of lines. */
pattern_space = next_line; pattern_space = next_line;
if (!pattern_space) return; if (!pattern_space) return;
last_char = next_last_char; last_gets_char = next_gets_char;
/* Read one line in advance so we can act on the last line, /* Read one line in advance so we can act on the last line,
* the '$' address */ * the '$' address */
next_line = get_next_line(&next_last_char); next_line = get_next_line(&next_gets_char);
linenum++; linenum++;
restart: restart:
/* for every line, go through all the commands */ /* for every line, go through all the commands */
@ -924,7 +924,7 @@ restart:
* (of current file) is printed. Even if * (of current file) is printed. Even if
* that line is nonterminated, we print * that line is nonterminated, we print
* '\n' here (gnu sed does the same) */ * '\n' here (gnu sed does the same) */
sed_puts(pattern_space, (last_char & 0x200) | '\n'); sed_puts(pattern_space, '\n');
break; break;
/* Delete up through first newline */ /* Delete up through first newline */
case 'D': case 'D':
@ -950,12 +950,12 @@ restart:
/* handle p option */ /* handle p option */
if (sed_cmd->sub_p) if (sed_cmd->sub_p)
sed_puts(pattern_space, last_char); sed_puts(pattern_space, last_gets_char);
/* handle w option */ /* handle w option */
if (sed_cmd->file) if (sed_cmd->sw_file)
/*sed_cmd->sw_last_char =*/ puts_maybe_newline( puts_maybe_newline(
pattern_space, sed_cmd->file, pattern_space, sed_cmd->sw_file,
last_char); &sed_cmd->sw_last_char, last_gets_char);
break; break;
/* Append line to linked list to be printed later */ /* Append line to linked list to be printed later */
@ -972,7 +972,7 @@ restart:
case 'c': case 'c':
/* Only triggers on last line of a matching range. */ /* Only triggers on last line of a matching range. */
if (!sed_cmd->in_match) if (!sed_cmd->in_match)
sed_puts(sed_cmd->string, MASK_NO_EOL_CHAR); sed_puts(sed_cmd->string, NO_EOL_CHAR);
goto discard_line; goto discard_line;
/* Read file, append contents to output */ /* Read file, append contents to output */
@ -995,20 +995,20 @@ restart:
/* Write pattern space to file. */ /* Write pattern space to file. */
case 'w': case 'w':
/*sed_cmd->sw_last_char =*/ puts_maybe_newline( puts_maybe_newline(
pattern_space, sed_cmd->file, pattern_space, sed_cmd->sw_file,
last_char); &sed_cmd->sw_last_char, last_gets_char);
break; break;
/* Read next line from input */ /* Read next line from input */
case 'n': case 'n':
if (!bbg.be_quiet) if (!bbg.be_quiet)
sed_puts(pattern_space, last_char); sed_puts(pattern_space, last_gets_char);
if (next_line) { if (next_line) {
free(pattern_space); free(pattern_space);
pattern_space = next_line; pattern_space = next_line;
last_char = next_last_char; last_gets_char = next_gets_char;
next_line = get_next_line(&next_last_char); next_line = get_next_line(&next_gets_char);
linenum++; linenum++;
break; break;
} }
@ -1037,8 +1037,8 @@ restart:
pattern_space = realloc(pattern_space, len + strlen(next_line) + 2); pattern_space = realloc(pattern_space, len + strlen(next_line) + 2);
pattern_space[len] = '\n'; pattern_space[len] = '\n';
strcpy(pattern_space + len+1, next_line); strcpy(pattern_space + len+1, next_line);
last_char = next_last_char; last_gets_char = next_gets_char;
next_line = get_next_line(&next_last_char); next_line = get_next_line(&next_gets_char);
linenum++; linenum++;
break; break;
} }
@ -1093,7 +1093,7 @@ restart:
strcat(pattern_space, "\n"); strcat(pattern_space, "\n");
if (bbg.hold_space) if (bbg.hold_space)
strcat(pattern_space, bbg.hold_space); strcat(pattern_space, bbg.hold_space);
last_char = '\n'; last_gets_char = '\n';
break; break;
} }
@ -1125,7 +1125,7 @@ restart:
{ {
char *tmp = pattern_space; char *tmp = pattern_space;
pattern_space = bbg.hold_space ? : xzalloc(1); pattern_space = bbg.hold_space ? : xzalloc(1);
last_char = '\n'; last_gets_char = '\n';
bbg.hold_space = tmp; bbg.hold_space = tmp;
break; break;
} }
@ -1140,7 +1140,7 @@ restart:
/* we will print the line unless we were told to be quiet ('-n') /* we will print the line unless we were told to be quiet ('-n')
or if the line was suppressed (ala 'd'elete) */ or if the line was suppressed (ala 'd'elete) */
if (!bbg.be_quiet) if (!bbg.be_quiet)
sed_puts(pattern_space, last_char); sed_puts(pattern_space, last_gets_char);
/* Delete and such jump here. */ /* Delete and such jump here. */
discard_line: discard_line:

4
testsuite/sed.tests
View File

@ -139,6 +139,10 @@ testing "sed selective matches noinsert newline" \
testing "sed clusternewline" \ testing "sed clusternewline" \
"sed -e '/one/a 111' -e '/two/i 222' -e p input -" \ "sed -e '/one/a 111' -e '/two/i 222' -e p input -" \
"one\none\n111\n222\ntwo\ntwo" "one" "two" "one\none\n111\n222\ntwo\ntwo" "one" "two"
testing "sed subst+write" \
"sed -e 's/i/z/' -e 'woutputw' input -; echo -n X; cat outputw" \
"thzngy\nagaznXthzngy\nagazn" "thingy" "again"
rm outputw
# Test end-of-file matching behavior # Test end-of-file matching behavior