shithub: sl

Info  •  Files  •  Log  •  Branches

string-split: add optional :trim T/NIL to filter out empty strings

--- a/src/sl.c

+++ b/src/sl.c

@@ -15,7 +15,7 @@

 sl_v sl_commadot, sl_trycatch, sl_backquote;

 sl_v sl_conssym, sl_symbolsym, sl_fixnumsym, sl_vectorsym, sl_builtinsym, sl_vu8sym;

 sl_v sl_defsym, sl_defmacrosym, sl_forsym, sl_setqsym;

-sl_v sl_booleansym, sl_nullsym, sl_evalsym, sl_fnsym;

+sl_v sl_booleansym, sl_nullsym, sl_evalsym, sl_fnsym, sl_trimsym;

 sl_v sl_nulsym, sl_alarmsym, sl_backspacesym, sl_tabsym, sl_linefeedsym, sl_newlinesym;

 sl_v sl_vtabsym, sl_pagesym, sl_returnsym, sl_escsym, sl_spacesym, sl_deletesym;

 sl_v sl_errio, sl_errparse, sl_errtype, sl_errarg, sl_errmem;

@@ -1272,6 +1272,7 @@

 	sl_evalsym = csymbol("eval");

 	sl_vu8sym = csymbol("vu8");

 	sl_fnsym = csymbol("fn");

+	sl_trimsym = csymbol(":trim");

 	sl_nulsym = csymbol("nul");

 	sl_alarmsym = csymbol("alarm");

 	sl_backspacesym = csymbol("backspace");

--- a/src/sl.h

+++ b/src/sl.h

@@ -415,7 +415,7 @@

 extern sl_v sl_commadot, sl_trycatch, sl_backquote;

 extern sl_v sl_conssym, sl_symbolsym, sl_fixnumsym, sl_vectorsym, sl_builtinsym, sl_vu8sym;

 extern sl_v sl_defsym, sl_defmacrosym, sl_forsym, sl_setqsym;

-extern sl_v sl_booleansym, sl_nullsym, sl_evalsym, sl_fnsym;

+extern sl_v sl_booleansym, sl_nullsym, sl_evalsym, sl_fnsym, sl_trimsym;

 extern sl_v sl_nulsym, sl_alarmsym, sl_backspacesym, sl_tabsym, sl_linefeedsym, sl_newlinesym;

 extern sl_v sl_vtabsym, sl_pagesym, sl_returnsym, sl_escsym, sl_spacesym, sl_deletesym;

 extern sl_v sl_errio, sl_errparse, sl_errtype, sl_errarg, sl_errmem;

--- a/src/string.c

+++ b/src/string.c

@@ -141,11 +141,30 @@

 BUILTIN("string-split", string_split)

 {

-	argcount(nargs, 2);

+	if(nargs < 1)

+		argcount(nargs, 1);

 	char *s = tostring(args[0]);

-	char *delim = tostring(args[1]);

 	usize len = cv_len(ptr(args[0]));

-	usize dlen = cv_len(ptr(args[1]));

+	// split on whitespace by default

+	const char *delim0 = " \t\n\r\v", *delim = delim0;

+	usize dlen = 5;

+	int n = 1;

+	// second is either a :trim or a separator

+	if(nargs > n && args[n] != sl_trimsym){

+		delim = tostring(args[n]);

+		dlen = cv_len(ptr(args[n]));

+		n++;

+	}

+	bool trim = false;

+	// it can only be a :trim X now

+	if(nargs > n){

+		if(args[n] != sl_trimsym)

+			lerrorf(sl_errarg, "invalid argument at position %d", n);

+		n++;

+		if(nargs <= n)

+			argcount(nargs, n+1);

+		trim = args[n] != sl_nil;

+	}

 	usize ssz, tokend, tokstart, i = 0;

 	sl_v first = sl_nil, c = sl_nil, last;

 	usize junk;

@@ -155,15 +174,20 @@

 	do{

 		// find and allocate next token

 		tokstart = tokend = i;

-		while(i < len && !u8_memchr(delim, u8_nextmemchar(s, &i), dlen, &junk))

+		while(i < len && !u8_memchr((char*)delim, u8_nextmemchar(s, &i), dlen, &junk))

 			tokend = i;

 		ssz = tokend - tokstart;

+		if(ssz == 0 && trim)

+			continue;

+

 		last = c; // save previous cons cell

 		c = mk_cons(cvalue_string(ssz), sl_nil);

-		// we've done allocation; reload movable pointers

-		s = cvalue_data(args[0]);

-		delim = cvalue_data(args[1]);

+		if(delim != delim0){

+			// we've done allocation; reload movable pointers

+			s = cvalue_data(args[0]);

+			delim = cvalue_data(args[1]);

+		}

 		if(ssz)

 			memmove(cvalue_data(car_(c)), &s[tokstart], ssz);

@@ -177,7 +201,7 @@

 		// note this tricky condition: if the string ends with a

 		// delimiter, we need to go around one more time to add an

 		// empty string. this happens when (i == len && tokend < i)

-	}while(i < len || (i == len && (tokend != i)));

+	}while(i < len || (i == len && tokend != i));

 	sl_free_gc_handles(2);

 	return first;

 }

--- a/test/unittest.lsp

+++ b/test/unittest.lsp

@@ -507,6 +507,7 @@

   (io-close b))

 (def s "привет\0пока")

+(def s2 "hello       \t   \n world\n ")

 (assert (equal? s (string-encode (string-decode s))))

 (assert (equal? (string s "\0") (string-encode (string-decode s t))))

@@ -551,6 +552,11 @@

 (assert (equal? '("привет" "пока") (string-split s "\0")))

 (assert (equal? '("пр" "вет" "пок" "") (string-split s "аи\0")))

 (assert (equal? '("" "") (string-split "1" "1")))

+

+(assert (equal? '("hello" "world") (string-split s2 :trim T)))

+(assert (equal? '("hello" "\t" "\n" "world\n") (string-split s2 " " :trim T)))

+(assert (equal? (list s2) (string-split s2 "X" :trim T)))

+(assert (equal? (list s2) (string-split s2 "X")))

 (assert (equal? #\а (string-char s 10)))

 (assert (equal? #\nul (string-char s 6)))