From cbcfcabc8f9b7f0ca14c6dce2ee5644e4e80e5e4 Mon Sep 17 00:00:00 2001
From: John Marshall <jm18@sanger.ac.uk>
Date: Tue, 9 Jun 2015 03:29:54 +0100
Subject: [PATCH] Add kgetline() to kstring.c/.h

Similar to BSD's getline() but omits the \n terminator and manages the
memory as a kstring.  Call with "(kgets_func *) fgets" to read from stdio,
or implement an fgets()-style function to read from other streams, e.g.,
a wrapper around gzgets() that reorders its parameters as per fgets().
---
 kstring.c           | 20 ++++++++++++++++++
 kstring.h           |  7 +++++++
 test/kstring_test.c | 58 ++++++++++++++++++++++++++++++++++++++++++++++++++++-
 3 files changed, 84 insertions(+), 1 deletion(-)

diff --git a/kstring.c b/kstring.c
index f029317..253d281 100644
--- a/kstring.c
+++ b/kstring.c
@@ -105,6 +105,26 @@ int ksplit_core(char *s, int delimiter, int *_max, int **_offsets)
 	return n;
 }
 
+int kgetline(kstring_t *s, kgets_func *fgets_fn, void *fp)
+{
+	size_t l0 = s->l;
+
+	while (s->l == l0 || s->s[s->l-1] != '\n') {
+		if (s->m - s->l < 200) ks_resize(s, s->m + 200);
+		if (fgets_fn(s->s + s->l, s->m - s->l, fp) == NULL) break;
+		s->l += strlen(s->s + s->l);
+	}
+
+	if (s->l == l0) return EOF;
+
+	if (s->l > l0 && s->s[s->l-1] == '\n') {
+		s->l--;
+		if (s->l > l0 && s->s[s->l-1] == '\r') s->l--;
+	}
+	s->s[s->l] = '\0';
+	return 0;
+}
+
 /**********************
  * Boyer-Moore search *
  **********************/
diff --git a/kstring.h b/kstring.h
index 2567efc..f13fcd9 100644
--- a/kstring.h
+++ b/kstring.h
@@ -82,6 +82,13 @@ extern "C" {
 	 * if sep is not changed. */
 	char *kstrtok(const char *str, const char *sep, ks_tokaux_t *aux);
 
+	/* kgetline() uses the supplied fgets()-like function to read a "\n"-
+	 * or "\r\n"-terminated line from fp.  The line read is appended to the
+	 * kstring without its terminator and 0 is returned; EOF is returned at
+	 * EOF or on error (determined by querying fp, as per fgets()). */
+	typedef char *kgets_func(char *, int, void *);
+	int kgetline(kstring_t *s, kgets_func *fgets, void *fp);
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/test/kstring_test.c b/test/kstring_test.c
index 76f9532..9f9b6e6 100644
--- a/test/kstring_test.c
+++ b/test/kstring_test.c
@@ -1,4 +1,5 @@
 #include <limits.h>
+#include <stdarg.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
@@ -37,7 +38,40 @@ void test_kputl(kstring_t *ks, long n)
 	check("kputl()", ks, buf);
 }
 
-int main()
+static char *mem_gets(char *buf, int buflen, void *vtextp)
+{
+	const char **textp = (const char **) vtextp;
+
+	const char *nl = strchr(*textp, '\n');
+	size_t n = nl? nl - *textp + 1 : strlen(*textp);
+
+	if (n == 0) return NULL;
+
+	if (n > buflen-1) n = buflen-1;
+	memcpy(buf, *textp, n);
+	buf[n] = '\0';
+	*textp += n;
+	return buf;
+}
+
+void test_kgetline(kstring_t *ks, const char *text, ...)
+{
+	const char *exp;
+	va_list arg;
+
+	va_start(arg, text);
+	while ((exp = va_arg(arg, const char *)) != NULL) {
+		ks->l = 0;
+		if (kgetline(ks, mem_gets, &text) != 0) kputs("EOF", ks);
+		check("kgetline()", ks, exp);
+	}
+	va_end(arg);
+
+	ks->l = 0;
+	if (kgetline(ks, mem_gets, &text) == 0) check("kgetline()", ks, "EOF");
+}
+
+int main(int argc, char **argv)
 {
 	kstring_t ks;
 
@@ -65,6 +99,28 @@ int main()
 	test_kputl(&ks, -LONG_MAX);
 	test_kputl(&ks, LONG_MIN);
 
+	test_kgetline(&ks, "", NULL);
+	test_kgetline(&ks, "apple", "apple", NULL);
+	test_kgetline(&ks, "banana\n", "banana", NULL);
+	test_kgetline(&ks, "carrot\r\n", "carrot", NULL);
+	test_kgetline(&ks, "\n", "", NULL);
+	test_kgetline(&ks, "\n\n", "", "", NULL);
+	test_kgetline(&ks, "foo\nbar", "foo", "bar", NULL);
+	test_kgetline(&ks, "foo\nbar\n", "foo", "bar", NULL);
+	test_kgetline(&ks,
+		"abcdefghijklmnopqrstuvwxyz0123456789\nABCDEFGHIJKLMNOPQRSTUVWXYZ\n",
+		"abcdefghijklmnopqrstuvwxyz0123456789",
+		"ABCDEFGHIJKLMNOPQRSTUVWXYZ", NULL);
+
+	if (argc > 1) {
+		FILE *f = fopen(argv[1], "r");
+		if (f) {
+			for (ks.l = 0; kgetline(&ks, (kgets_func *)fgets, f) == 0; ks.l = 0)
+				puts(ks.s);
+			fclose(f);
+		}
+	}
+
 	free(ks.s);
 
 	if (nfail > 0) {