Improve some unicode functions

Author: jart

ape/ape.S

@@ -1476,7 +1476,22 @@ kernel:	movabs	$ape_stack_vaddr,%rsp
 	test	%rax,%rax
 	jz	1f
 	movb	$METAL,(%rax)
-1:	xor	%eax,%eax
+1:	push	$0
+	mov	%rsp,%rbp
+	mov	.Lenv0(%rip),%rax
+	mov	%rax,(%rbp)			# envp[0][0]
+	push	$0				# argv[0][0]
+	push	$0				# auxv[1][1]
+	push	$0				# auxv[1][0]
+	push	%rbp				# auxv[0][1]
+	push	$31				# auxv[0][0] AT_EXECFN
+	push	$0				# envp[1]
+	push	$.Lenv0				# envp[0]
+	push	$0				# argv[1]
+	push	%rbp				# argv[0]
+	push	$1				# argc
+	xor	%ebp,%ebp
+	xor	%eax,%eax
 	xor	%ecx,%ecx
 	xor	%edx,%edx
 	xor	%edi,%edi
@@ -1485,20 +1500,10 @@ kernel:	movabs	$ape_stack_vaddr,%rsp
 	xor	%r9d,%r9d
 	xor	%r10d,%r10d
 	xor	%r11d,%r11d
-	push	$0				# auxv[1][1]
-	push	$0				# auxv[1][0]
-	push	$.Larg0				# auxv[0][1]
-	push	$31				# auxv[0][0] AT_EXECFN
-	push	$0				# envp[1]
-	push	$.Lenv0				# envp[0]
-	push	$0				# argv[1]
-	push	$.Larg0				# argv[0]
-	push	$1				# argc
 	jmp	_start
 	.endfn	kernel
 
 	.rodata
-.Larg0:	.asciz	"ape.com"
 .Lenv0:	.asciz	"METAL=1"
 	.previous
 

libc/runtime/getdosargv.c

@@ -89,14 +89,16 @@ textwindows noasan int GetDosArgv(const char16_t *cmdline, char *buf,
   argc = 0;
   st.wc = DecodeDosArgv(&st.s);
   while (st.wc) {
-    while (st.wc && isspace(st.wc)) st.wc = DecodeDosArgv(&st.s);
+    while (st.wc && (st.wc == ' ' || st.wc == '\t')) {
+      st.wc = DecodeDosArgv(&st.s);
+    }
     if (!st.wc) break;
     if (++argc < max) {
       argv[argc - 1] = st.p < st.pe ? st.p : NULL;
     }
     inquote = false;
     while (st.wc) {
-      if (!inquote && isspace(st.wc)) break;
+      if (!inquote && (st.wc == ' ' || st.wc == '\t')) break;
       if (st.wc == '"' || st.wc == '\\') {
         slashes = 0;
         quotes = 0;

libc/str/isspace.c

@@ -19,7 +19,8 @@
 #include "libc/str/str.h"
 
 /**
- * Returns true if c is space, \t, \r, \n, \f, or \v.
+ * Returns nonzero if c is space, \t, \r, \n, \f, or \v.
+ * @see isblank()
  */
 int isspace(int c) {
   return c == ' ' || c == '\t' || c == '\r' || c == '\n' || c == '\f' ||

libc/str/iswalnum.c

@@ -18,6 +18,9 @@
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
 
-int iswalnum(wint_t wc) {
-  return isalnum(wc);
+/**
+ * Returns nonzero if c is lower, alpha, or digit.
+ */
+int iswalnum(wint_t c) {
+  return iswdigit(c) || iswalpha(c);
 }

libc/str/iswalpha.c

@@ -18,6 +18,9 @@
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
 
-int iswalpha(wint_t wc) {
-  return isalpha(wc);
+/**
+ * Returns nonzero if c is alphabetical.
+ */
+int iswalpha(wint_t c) {
+  return iswupper(c) || iswlower(c);
 }

libc/str/iswblank.c

@@ -18,6 +18,9 @@
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
 
-int iswblank(wint_t wc) {
-  return isblank(wc);
+/**
+ * Returns nonzero if c is space or tab.
+ */
+int iswblank(wint_t c) {
+  return c == ' ' || c == '\t';
 }

libc/str/iswcntrl.c

@@ -19,8 +19,8 @@
 #include "libc/str/str.h"
 
 /**
- * Returns nonzero if wc is C0 or C1 control code.
+ * Returns nonzero if c is C0 or C1 control code.
  */
-int iswcntrl(wint_t wc) {
-  return (0x00 <= wc && wc <= 0x1F) || (0x7F <= wc && wc <= 0x9F);
+int iswcntrl(wint_t c) {
+  return (0x00 <= c && c <= 0x1F) || (0x7F <= c && c <= 0x9F);
 }

libc/str/iswctype.c

@@ -19,71 +19,30 @@
 #include "libc/macros.internal.h"
 #include "libc/str/str.h"
 
-#define ALNUM  1
-#define ALPHA  2
-#define BLANK  3
-#define CNTRL  4
-#define DIGIT  5
-#define GRAPH  6
-#define LOWER  7
-#define PRINT  8
-#define PUNCT  9
-#define SPACE  10
-#define UPPER  11
-#define XDIGIT 12
-
-static const struct {
-  char name[7];
-  char type;
-} kWcTypes[] = {
-    {"alnum", ALNUM}, {"alpha", ALPHA}, {"blank", BLANK}, {"cntrl", CNTRL},
-    {"digit", DIGIT}, {"graph", GRAPH}, {"lower", LOWER}, {"print", PRINT},
-    {"punct", PUNCT}, {"space", SPACE}, {"upper", UPPER}, {"xdigit", XDIGIT},
+static const int (*const kWcTypeFuncs[])(wint_t) = {
+    iswalnum,   //
+    iswalpha,   //
+    iswblank,   //
+    iswcntrl,   //
+    iswdigit,   //
+    iswgraph,   //
+    iswlower,   //
+    iswprint,   //
+    iswpunct,   //
+    iswspace,   //
+    iswupper,   //
+    iswxdigit,  //
 };
 
-static int CompareStrings(const char *l, const char *r) {
-  size_t i = 0;
-  while (l[i] == r[i] && r[i]) ++i;
-  return (l[i] & 0xff) - (r[i] & 0xff);
-}
-
-wctype_t wctype(const char *name) {
-  unsigned i;
-  for (i = 0; i < ARRAYLEN(kWcTypes); ++i) {
-    if (CompareStrings(name, kWcTypes[i].name) == 0) {
-      return kWcTypes[i].type;
-    }
-  }
-  return 0;
-}
-
-int iswctype(wint_t wc, wctype_t type) {
-  switch (type) {
-    case ALNUM:
-      return iswalnum(wc);
-    case ALPHA:
-      return iswalpha(wc);
-    case BLANK:
-      return iswblank(wc);
-    case CNTRL:
-      return iswcntrl(wc);
-    case DIGIT:
-      return iswdigit(wc);
-    case GRAPH:
-      return iswgraph(wc);
-    case LOWER:
-      return iswlower(wc);
-    case PRINT:
-      return iswprint(wc);
-    case PUNCT:
-      return iswpunct(wc);
-    case SPACE:
-      return iswspace(wc);
-    case UPPER:
-      return iswupper(wc);
-    case XDIGIT:
-      return iswxdigit(wc);
-    default:
-      return 0;
+/**
+ * Returns nonzero if c has property.
+ *
+ * @param t is number returned by wctype
+ */
+int iswctype(wint_t c, wctype_t t) {
+  if (1 <= t && t <= ARRAYLEN(kWcTypeFuncs)) {
+    return kWcTypeFuncs[t - 1](c);
+  } else {
+    return 0;
   }
 }

libc/str/iswdigit.c

@@ -18,6 +18,9 @@
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
 
-int iswdigit(wint_t wc) {
-  return isdigit(wc);
+/**
+ * Returns nonzero if c is decimal digit.
+ */
+int iswdigit(wint_t c) {
+  return '0' <= c && c <= '9';
 }

libc/str/iswgraph.c

@@ -18,6 +18,9 @@
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "libc/str/str.h"
 
-int iswgraph(wint_t wc) {
-  return isgraph(wc);
+/**
+ * Returns nonzero if c is printable and not a space.
+ */
+int iswgraph(wint_t c) {
+  return iswprint(c) && !iswspace(c);
 }