ericonr
/
util-linux
mirror of https://github.com/ericonr/util-linux.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

col: make input to tolerate invalid wide characters 

The getwchar(3) will choke and exit if invalid character is encountered.
This change will make col(1) to print broken multibyte characters as
\x{hex} string.

Reported-by: Vitaly Lipatov <lav@etersoft.ru>
Addresses: https://github.com/karelzak/util-linux/issues/1198
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
This commit is contained in:
Sami Kerola 2020-11-26 22:01:39 +00:00
parent 77d44c1e51
commit 14a36f96ee
No known key found for this signature in database
GPG Key ID: 0D46FEF7E61DBB46
5 changed files with 77 additions and 46 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
tests/expected/col/multibyte
View File

@ -1 +0,0 @@
col: failed on line 1: EILSEQ

1
tests/expected/col/multibyte-invalid Normal file
View File

@ -0,0 +1 @@
abc\x80\x80jkl

1
tests/expected/col/multibyte-valid Normal file
View File

@ -0,0 +1 @@
Dateiname der Versandh\xfclle

11
tests/ts/col/multibyte
View File

@ -24,8 +24,17 @@ ts_init "$*"
ts_check_test_command "$TS_CMD_COL"
ts_check_test_command "$TS_HELPER_STRERROR"
ts_init_subtest "valid"
cat $TS_SELF/multibyte.data |
LC_ALL=C.UTF-8 ts_run $TS_CMD_COL 2>&1 > /dev/null |
LC_ALL=C ts_run $TS_CMD_COL 2>&1 |
sed -e "s@$($TS_HELPER_STRERROR EILSEQ)@EILSEQ@" > $TS_OUTPUT
ts_finalize_subtest
ts_init_subtest "invalid"
printf '%s\n' $'abc\200\200jkl' |
LC_ALL=C ts_run $TS_CMD_COL 2>&1 |
sed -e "s@$($TS_HELPER_STRERROR EILSEQ)@EILSEQ@" > $TS_OUTPUT
ts_finalize_subtest
ts_finalize

109
text-utils/col.c
View File

@ -599,6 +599,49 @@ static void free_line_allocations(struct col_alloc *root)
}
#endif
static void process_char(struct col_ctl *ctl, struct col_lines *lns)
{
/* Deal printable characters */
if (!iswgraph(lns->ch) && handle_not_graphic(ctl, lns))
return;
/* Must stuff ch in a line - are we at the right one? */
if ((size_t)lns->cur_line != lns->this_line - lns->adjust)
update_cur_line(ctl, lns);
/* Does line buffer need to grow? */
if (ctl->l->l_lsize <= ctl->l->l_line_len + 1) {
size_t need;
need = ctl->l->l_lsize ? ctl->l->l_lsize * 2 : NALLOC;
ctl->l->l_line = xrealloc(ctl->l->l_line, need * sizeof(struct col_char));
ctl->l->l_lsize = need;
}
/* Store character */
lns->c = &ctl->l->l_line[ctl->l->l_line_len++];
lns->c->c_char = lns->ch;
lns->c->c_set = lns->cur_set;
if (0 < lns->cur_col)
lns->c->c_column = lns->cur_col;
else
lns->c->c_column = 0;
lns->c->c_width = wcwidth(lns->ch);
/*
* If things are put in out of order, they will need sorting
* when it is flushed.
*/
if (lns->cur_col < ctl->l->l_max_col)
ctl->l->l_needs_sort = 1;
else
ctl->l->l_max_col = lns->cur_col;
if (0 < lns->c->c_width)
lns->cur_col += lns->c->c_width;
}
int main(int argc, char **argv)
{
struct col_ctl ctl = {
@ -623,52 +666,30 @@ int main(int argc, char **argv)
while (feof(stdin) == 0) {
errno = 0;
/* Get character */
if ((lns.ch = getwchar()) == WEOF) {
lns.ch = getwchar();
if (lns.ch == WEOF) {
if (errno == EILSEQ) {
warn(_("failed on line %lu"), lns.max_line + 1);
ret = EXIT_FAILURE;
}
break;
}
/* Illegal multibyte sequence */
int c;
char buf[5];
size_t len, i;
/* Deal printable characters */
if (!iswgraph(lns.ch) && handle_not_graphic(&ctl, &lns))
continue;
/* Must stuff ch in a line - are we at the right one? */
if ((size_t)lns.cur_line != lns.this_line - lns.adjust)
update_cur_line(&ctl, &lns);
/* Does line buffer need to grow? */
if (ctl.l->l_lsize <= ctl.l->l_line_len + 1) {
size_t need;
need = ctl.l->l_lsize ? ctl.l->l_lsize * 2 : NALLOC;
ctl.l->l_line = xrealloc(ctl.l->l_line, need * sizeof(struct col_char));
ctl.l->l_lsize = need;
}
/* Store character */
lns.c = &ctl.l->l_line[ctl.l->l_line_len++];
lns.c->c_char = lns.ch;
lns.c->c_set = lns.cur_set;
if (0 < lns.cur_col)
lns.c->c_column = lns.cur_col;
else
lns.c->c_column = 0;
lns.c->c_width = wcwidth(lns.ch);
/*
* If things are put in out of order, they will need sorting
* when it is flushed.
*/
if (lns.cur_col < ctl.l->l_max_col)
ctl.l->l_needs_sort = 1;
else
ctl.l->l_max_col = lns.cur_col;
if (0 < lns.c->c_width)
lns.cur_col += lns.c->c_width;
c = getchar();
if (c == EOF)
break;
sprintf(buf, "\\x%02x", (unsigned char) c);
len = strlen(buf);
for (i = 0; i < len; i++) {
lns.ch = buf[i];
process_char(&ctl, &lns);
}
} else
/* end of file */
break;
} else
/* the common case */
process_char(&ctl, &lns);
}
/* goto the last line that had a character on it */