Hello,
There are various program, especially the one written with GNU grep in
mind, that expects various flags that grep in base doesn't have. While
some of the flags (like --color) can be easily worked out (i.e. by
patching/customising these programs) one thing that it isn't easily
workable is --null, because it alters the way grep outputs its data.
--null makes grep output an ASCII NUL byte after the file name, so a
program can parse the output of grep unambiguously even when the file
names contains "funny" characters, such as a newline.
GNU grep isn't the only one with a --null flag, also FreeBSD and NetBSD
grep do (at least by looking at their manpages), so it's somewhat
widespread.
I searched the archives on marc.info but I haven't seen a previous
discussion about this.
The following patch was tried against GNU grep (installed from packages)
and seem to behave consistently.
I used the same text of the FreeBSD/NetBSD manpage for the description
of the --null option, but I really dislike it: it feels way to verbose
for what it's trying to say, but I wasn't able to come up with something
better.
I'm not familiar at all with the grep codebase, so I hope I'm not
missing something. If this has some chances of being accepted, I guess
I should also add some regress test; I'll try to work on them soon, but
in the meantime I'm sending this.
Thanks,
Omar Polo
diff e992327fc31d0277a6f8518613a7db1b9face78b /home/op/w/openbsd-src
blob - 5cc228df222c54a0553f289b5da8bbbe6afd171e
file + usr.bin/grep/grep.1
--- usr.bin/grep/grep.1
+++ usr.bin/grep/grep.1
@@ -49,6 +49,7 @@
.Op Fl -context Ns Op = Ns Ar num
.Op Fl -label Ns = Ns Ar name
.Op Fl -line-buffered
+.Op Fl -null
.Op Ar pattern
.Op Ar
.Ek
@@ -297,6 +298,25 @@ instead of the filename before lines.
Force output to be line buffered.
By default, output is line buffered when standard output is a terminal
and block buffered otherwise.
+.It Fl -null
+Output a zero byte (the ASCII NUL character) instead of the character
+that normally follows a file name.
+For example,
+.Nm Fl l Fl -null
+outputs a zero byte after each file name instead of the usual newline.
+This option makes the output unambiguous, even in the presence of file
+names containing unusual characters like newlines.
+This option can be used with commands like
+.Xr find 1
+.Fl print0 Ns ,
+.Xr perl 1
+.Fl 0 Ns ,
+.Xr sort 1
+.Fl z Ns , and
+.Xr args 1
+.Fl 0
+to process arbitrary file names, even those that contain newline
+characters.
.El
.Sh EXIT STATUS
The
blob - f41b5e20ca68c9e9a36d2f7dd3c44329c621f29b
file + usr.bin/grep/grep.c
--- usr.bin/grep/grep.c
+++ usr.bin/grep/grep.c
@@ -80,6 +80,7 @@ int vflag; /* -v: only show non-matching lines */
int wflag; /* -w: pattern must start and end on word boundaries */
int xflag; /* -x: pattern must match entire line */
int lbflag; /* --line-buffered */
+int nullflag; /* --null */
const char *labelname; /* --label=name */
int binbehave = BIN_FILE_BIN;
@@ -89,6 +90,7 @@ enum {
HELP_OPT,
MMAP_OPT,
LINEBUF_OPT,
+ NULL_OPT,
LABEL_OPT,
};
@@ -134,6 +136,7 @@ static const struct option long_options[] =
{"mmap", no_argument, NULL, MMAP_OPT},
{"label", required_argument, NULL, LABEL_OPT},
{"line-buffered", no_argument, NULL, LINEBUF_OPT},
+ {"null", no_argument, NULL, NULL_OPT},
{"after-context", required_argument, NULL, 'A'},
{"before-context", required_argument, NULL, 'B'},
{"context", optional_argument, NULL, 'C'},
@@ -436,6 +439,9 @@ main(int argc, char *argv[])
case LINEBUF_OPT:
lbflag = 1;
break;
+ case NULL_OPT:
+ nullflag = 1;
+ break;
case HELP_OPT:
default:
usage();
blob - b3d24ae662beb72c5632190c5c819bcc92f0389a
file + usr.bin/grep/grep.h
--- usr.bin/grep/grep.h
+++ usr.bin/grep/grep.h
@@ -68,7 +68,7 @@ extern int cflags, eflags;
extern int Aflag, Bflag, Eflag, Fflag, Hflag, Lflag,
Rflag, Zflag,
bflag, cflag, hflag, iflag, lflag, mflag, nflag, oflag, qflag,
- sflag, vflag, wflag, xflag;
+ sflag, vflag, wflag, xflag, nullflag;
extern int binbehave;
extern const char *labelname;
blob - e16d08e7d859609c2ccbc0ac4bba670188f81abf
file + usr.bin/grep/util.c
--- usr.bin/grep/util.c
+++ usr.bin/grep/util.c
@@ -172,13 +172,13 @@ procfile(char *fn)
if (cflag) {
if (!hflag)
- printf("%s:", ln.file);
+ printf("%s%c", ln.file, nullflag ? '\0' : ':');
printf("%llu%s\n", c, overflow ? "+" : "");
}
if (lflag && c != 0)
- printf("%s\n", fn);
+ printf("%s%c", fn, nullflag ? '\0' : '\n');
if (Lflag && c == 0)
- printf("%s\n", fn);
+ printf("%s%c", fn, nullflag ? '\0' : '\n');
if (c && !cflag && !lflag && !Lflag &&
binbehave == BIN_FILE_BIN && nottext && !qflag)
printf("Binary file %s matches\n", fn);
@@ -266,9 +266,9 @@ print:
if (Bflag > 0)
printqueue();
linesqueued = 0;
- printline(l, ':', oflag ? &pmatch : NULL);
+ printline(l, nullflag ? '\0' : ':', oflag ? &pmatch :
NULL);
} else {
- printline(l, '-', oflag ? &pmatch : NULL);
+ printline(l, nullflag ? '\0' : '-', oflag ? &pmatch :
NULL);
tail--;
}
}
@@ -660,7 +660,7 @@ printline(str_t *line, int sep, regmatch_t *pmatch)
}
if (nflag) {
if (n)
- putchar(sep);
+ putchar(nullflag ? '\0' : sep);
printf("%lld", line->line_no);
++n;
}
