bytestring: Split bytestrings by a string delimiter

This introduces the functions bytestring_splitstr_first() and
bytestring_splitstr_next() which can be used to iterate through substrings
of a bytestring separated by a specified substring.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

diff --git a/ccan/bytestring/bytestring.c b/ccan/bytestring/bytestring.c
index 81cca4edbe66354779cb977058d659cde5d7ad08..09ab006ad510c3d9a4d4cf795cd64279756d2957 100644 (file)
--- a/ccan/bytestring/bytestring.c
+++ b/ccan/bytestring/bytestring.c
@@ -104,3 +104,45 @@ struct bytestring bytestring_splitchrs_next(struct bytestring whole,
 
        return _splitchrs(whole, delim, (prev.ptr - whole.ptr) + prev.len + 1);
 }
+
+static struct bytestring _splitstr(struct bytestring whole,
+                                  struct bytestring delim, size_t start)
+{
+       struct bytestring remainder, nextdelim;
+
+       assert(start <= whole.len);
+
+       remainder = bytestring_slice(whole, start, whole.len);
+       nextdelim = bytestring_bytestring(remainder, delim);
+       if (nextdelim.ptr)
+               return bytestring_slice(whole, start,
+                                       nextdelim.ptr - whole.ptr);
+       else
+               return remainder;
+}
+
+struct bytestring bytestring_splitstr_first(struct bytestring whole,
+                                            struct bytestring delim)
+{
+       if (whole.len == 0)
+               return bytestring_NULL;
+
+       return _splitstr(whole, delim, 0);
+}
+
+struct bytestring bytestring_splitstr_next(struct bytestring whole,
+                                          struct bytestring delim,
+                                          struct bytestring prev)
+{
+       if (!prev.ptr)
+               return bytestring_NULL;
+
+       /* prev has to be a substring of whole */
+       assert(prev.ptr >= whole.ptr);
+
+       if ((prev.ptr + prev.len) == (whole.ptr + whole.len))
+               return bytestring_NULL;
+
+       return _splitstr(whole, delim,
+                        (prev.ptr - whole.ptr) + prev.len + delim.len);
+}

diff --git a/ccan/bytestring/bytestring.h b/ccan/bytestring/bytestring.h
index 5a8d3cd896fb003704bada0f535d4e11c5354171..0d1c929e599564d92984c836b9f434b5195dbd1e 100644 (file)
--- a/ccan/bytestring/bytestring.h
+++ b/ccan/bytestring/bytestring.h
@@ -276,4 +276,32 @@ struct bytestring bytestring_splitchrs_next(struct bytestring whole,
             (_s).ptr;                                         \
             (_s) = bytestring_splitchrs_next((_w), (_delim), (_s)))
 
+/**
+ * bytestring_splitstr_first - split a bytestring on a delimiter string
+ * @whole: a bytestring
+ * @delim: delimiter substring
+ *
+ * Returns the first substring of @whole delimited by the substring in
+ * @delim.
+ */
+struct bytestring bytestring_splitstr_first(struct bytestring whole,
+                                            struct bytestring delim);
+
+/**
+ * bytestring_splitstr_next - split a bytestring on a delimiter string
+ * @whole: a bytestring
+ * @delim: delimiter string
+ * @prev: last substring
+ *
+ * Returns the next @delim delimited substring of @whole after @prev.
+ */
+struct bytestring bytestring_splitstr_next(struct bytestring whole,
+                                          struct bytestring delim,
+                                          struct bytestring prev);
+
+#define bytestring_foreach_splitstr(_s, _w, _delim) \
+       for ((_s) = bytestring_splitstr_first((_w), (_delim)); \
+            (_s).ptr;                                         \
+            (_s) = bytestring_splitstr_next((_w), (_delim), (_s)))
+
 #endif /* CCAN_BYTESTRING_H_ */

diff --git a/ccan/bytestring/test/run.c b/ccan/bytestring/test/run.c
index f1f68d6620b96f2b7be01ba03ebd10e72a7c16ac..8b8e7ecdb8dfec0b92d262460640d7f997d8b830 100644 (file)
--- a/ccan/bytestring/test/run.c
+++ b/ccan/bytestring/test/run.c
@@ -17,7 +17,7 @@ int main(void)
        int n;
 
        /* This is how many tests you plan to run */
-       plan_tests(109);
+       plan_tests(126);
 
        bs = bytestring(str1, sizeof(str1) - 1);
        ok1(bs.ptr == str1);
@@ -198,6 +198,39 @@ int main(void)
        }
        ok1(n == 6);
 
+       bs7 = bytestring_splitstr_first(bs, BYTESTRING("t "));
+       ok1(bs7.ptr == bs.ptr);
+       ok1(bytestring_eq(bs7, BYTESTRING("tes")));
+       bs7 = bytestring_splitstr_next(bs, BYTESTRING("t "), bs7);
+       ok1(bs7.ptr == bs.ptr + 5);
+       ok1(bytestring_eq(bs7, BYTESTRING("string")));
+       bs7 = bytestring_splitstr_next(bs, BYTESTRING("t "), bs7);
+       ok1(!bs7.ptr);
+       bs7 = bytestring_splitstr_next(bs, BYTESTRING("t "), bs7);
+       ok1(!bs7.ptr);
+
+       bs7 = bytestring_splitstr_first(bs2, BYTESTRING("abc"));
+       ok1(bs7.ptr == bs2.ptr);
+       ok1(bs7.len == 0);
+       bs7 = bytestring_splitstr_next(bs2, BYTESTRING("abc"), bs7);
+       ok1(bs7.ptr == bs2.ptr + 3);
+       ok1(bytestring_eq(bs7, BYTESTRING("\0def")));
+       bs7 = bytestring_splitstr_next(bs2, BYTESTRING("t "), bs7);
+       ok1(!bs7.ptr);
+       bs7 = bytestring_splitstr_next(bs2, BYTESTRING("t "), bs7);
+       ok1(!bs7.ptr);
+
+       bs7 = bytestring_splitstr_first(BYTESTRING(""), BYTESTRING(""));
+       ok1(!bs7.ptr);
+
+       n = 0;
+       bytestring_foreach_splitstr(bs7, BYTESTRING("foofoo"),
+                                    BYTESTRING("foo")) {
+               n++;
+               ok1(bs7.ptr && !bs7.len);
+       }
+       ok1(n == 3);
+       
        /* This exits depending on whether all tests passed */
        return exit_status();
 }