From: Martin Mares <mj@ucw.cz>
Date: Mon, 23 Jul 2012 18:27:08 +0000 (+0200)
Subject: Added --trim
X-Git-Tag: v1.0~46
X-Git-Url: http://mj.ucw.cz/gitweb/?a=commitdiff_plain;h=eaea4ae3e62b3cc92a720b75f4ed416c4584877f;p=xsv.git

Added --trim
---

diff --git a/xsv.c b/xsv.c
index 38f1fde..87c8dff 100644
--- a/xsv.c
+++ b/xsv.c
@@ -128,6 +128,11 @@ restart:
 	}
 }
 
+static int is_ws(int c)
+{
+	return (c == ' ' || c == '\t' || c == '\f');
+}
+
 static void csv_write(void)
 {
 	unsigned char *line = line_nth(&in_line, 0);
@@ -171,7 +176,7 @@ static int ws_read(void)
 			continue;
 		if (c == '\n')
 			return 1;
-		if (c == ' ' || c == '\t' || c == '\f') {
+		if (is_ws(c)) {
 			ensure_field();
 			if (!ws)
 				new_field();
@@ -185,6 +190,20 @@ static int ws_read(void)
 	}
 }
 
+/*** Transforms ***/
+
+static void trim_fields(void)
+{
+	unsigned char *line = line_nth(&in_line, 0);
+	for (int i = 0; i < fields_count(&in_fields); i++) {
+		struct field *f = fields_nth(&in_fields, i);
+		while (f->len && is_ws(line[f->start_pos]))
+			f->start_pos++, f->len--;
+		while (f->len && is_ws(line[f->start_pos + f->len - 1]))
+			f->len--;
+	}
+}
+
 /*** Field selection ***/
 
 struct selector {
@@ -250,10 +269,10 @@ Formats:\n\
 -w, --ws		Values separated by arbitrary whitespace\n\
 \n\
 Format parameters:\n\
--d, --fs=<char>	Delimiter of fields\n\
+-d, --fs=<char>		Delimiter of fields\n\
 \n\
 Other options:\n\
-(so far none)\n\
+    --trim		Trim leading and trailing whitespaces in fields\n\
 ");
 	exit(0);
 }
@@ -270,11 +289,13 @@ static const char short_options[] = "cd:tw";
 
 enum long_options {
 	OPT_HELP = 256,
+	OPT_TRIM = 257,
 };
 
 static const struct option long_options[] = {
 	{ "csv",		0,	NULL,	'c' },
 	{ "fs",			1,	NULL,	'd' },
+	{ "trim",		0,	NULL,	OPT_TRIM },
 	{ "tsv",		0,	NULL,	't' },
 	{ "ws",			0,	NULL,	'w' },
 	{ "help",		0,	NULL,	OPT_HELP },
@@ -329,6 +350,7 @@ static struct format *current_format(void)
 int main(int argc, char **argv)
 {
 	int opt;
+	int want_trim = 0;
 
 	while ((opt = getopt_long(argc, argv, short_options, long_options, NULL)) >= 0)
 		switch (opt) {
@@ -349,6 +371,9 @@ int main(int argc, char **argv)
 				break;
 			case OPT_HELP:
 				usage();
+			case OPT_TRIM:
+				want_trim = 1;
+				break;
 			default:
 				bad_args(NULL);
 		}
@@ -375,6 +400,9 @@ int main(int argc, char **argv)
 		if (!in_format->read_line())
 			break;
 
+		if (want_trim)
+			trim_fields();
+
 		fields_reset(&out_fields);
 		select_fields();