String buffers, part 2c: abstract out string.format.

Sponsored by fmad.io.

src/Makefile.dep

@@ -197,7 +197,8 @@ lj_state.o: lj_state.c lj_obj.h lua.h luaconf.h lj_def.h lj_arch.h \
 lj_str.o: lj_str.c lj_obj.h lua.h luaconf.h lj_def.h lj_arch.h lj_gc.h \
  lj_err.h lj_errmsg.h lj_str.h lj_char.h lj_prng.h
 lj_strfmt.o: lj_strfmt.c lj_obj.h lua.h luaconf.h lj_def.h lj_arch.h \
- lj_buf.h lj_gc.h lj_str.h lj_state.h lj_char.h lj_strfmt.h
+ lj_err.h lj_errmsg.h lj_buf.h lj_gc.h lj_str.h lj_meta.h lj_state.h \
+ lj_char.h lj_strfmt.h lj_lib.h
 lj_strfmt_num.o: lj_strfmt_num.c lj_obj.h lua.h luaconf.h lj_def.h \
  lj_arch.h lj_buf.h lj_gc.h lj_str.h lj_strfmt.h
 lj_strscan.o: lj_strscan.c lj_obj.h lua.h luaconf.h lj_def.h lj_arch.h \

src/lib_string.c

@@ -640,89 +640,14 @@ LJLIB_CF(string_gsub)
 
 /* ------------------------------------------------------------------------ */
 
-/* Emulate tostring() inline. */
-static GCstr *string_fmt_tostring(lua_State *L, int arg, int retry)
-{
-  TValue *o = L->base+arg-1;
-  cTValue *mo;
-  lj_assertL(o < L->top, "bad usage");  /* Caller already checks for existence. */
-  if (LJ_LIKELY(tvisstr(o)))
-    return strV(o);
-  if (retry != 2 && !tvisnil(mo = lj_meta_lookup(L, o, MM_tostring))) {
-    copyTV(L, L->top++, mo);
-    copyTV(L, L->top++, o);
-    lua_call(L, 1, 1);
-    copyTV(L, L->base+arg-1, --L->top);
-    return NULL;  /* Buffer may be overwritten, retry. */
-  }
-  return lj_strfmt_obj(L, o);
-}
-
 LJLIB_CF(string_format)		LJLIB_REC(.)
 {
-  int arg, top = (int)(L->top - L->base);
-  GCstr *fmt;
-  SBuf *sb;
-  FormatState fs;
-  SFormat sf;
   int retry = 0;
-again:
+  SBuf *sb;
-  arg = 1;
+  do {
     sb = lj_buf_tmp_(L);
-  fmt = lj_lib_checkstr(L, arg);
+    retry = lj_strfmt_putarg(L, sb, 1, -retry);
-  lj_strfmt_init(&fs, strdata(fmt), fmt->len);
+  } while (retry > 0);
-  while ((sf = lj_strfmt_parse(&fs)) != STRFMT_EOF) {
-    if (sf == STRFMT_LIT) {
-      lj_buf_putmem(sb, fs.str, fs.len);
-    } else if (sf == STRFMT_ERR) {
-      lj_err_callerv(L, LJ_ERR_STRFMT, strdata(lj_str_new(L, fs.str, fs.len)));
-    } else {
-      if (++arg > top)
-	luaL_argerror(L, arg, lj_obj_typename[0]);
-      switch (STRFMT_TYPE(sf)) {
-      case STRFMT_INT:
-	if (tvisint(L->base+arg-1)) {
-	  int32_t k = intV(L->base+arg-1);
-	  if (sf == STRFMT_INT)
-	    lj_strfmt_putint(sb, k);  /* Shortcut for plain %d. */
-	  else
-	    lj_strfmt_putfxint(sb, sf, k);
-	} else {
-	  lj_strfmt_putfnum_int(sb, sf, lj_lib_checknum(L, arg));
-	}
-	break;
-      case STRFMT_UINT:
-	if (tvisint(L->base+arg-1))
-	  lj_strfmt_putfxint(sb, sf, intV(L->base+arg-1));
-	else
-	  lj_strfmt_putfnum_uint(sb, sf, lj_lib_checknum(L, arg));
-	break;
-      case STRFMT_NUM:
-	lj_strfmt_putfnum(sb, sf, lj_lib_checknum(L, arg));
-	break;
-      case STRFMT_STR: {
-	GCstr *str = string_fmt_tostring(L, arg, retry);
-	if (str == NULL)
-	  retry = 1;
-	else if ((sf & STRFMT_T_QUOTED))
-	  lj_strfmt_putquoted(sb, str);  /* No formatting. */
-	else
-	  lj_strfmt_putfstr(sb, sf, str);
-	break;
-	}
-      case STRFMT_CHAR:
-	lj_strfmt_putfchar(sb, sf, lj_lib_checkint(L, arg));
-	break;
-      case STRFMT_PTR:  /* No formatting. */
-	lj_strfmt_putptr(sb, lj_obj_ptr(G(L), L->base+arg-1));
-	break;
-      default:
-	lj_assertL(0, "bad string format type");
-	break;
-      }
-    }
-  }
-  if (retry++ == 1) goto again;
   setstrV(L, L->top-1, lj_buf_str(L, sb));
   lj_gc_check(L);
   return 1;

src/lj_strfmt.c

@@ -9,11 +9,14 @@
 #define LUA_CORE
 
 #include "lj_obj.h"
+#include "lj_err.h"
 #include "lj_buf.h"
 #include "lj_str.h"
+#include "lj_meta.h"
 #include "lj_state.h"
 #include "lj_char.h"
 #include "lj_strfmt.h"
+#include "lj_lib.h"
 
 /* -- Format parser ------------------------------------------------------- */
 
@@ -196,10 +199,8 @@ SBuf * LJ_FASTCALL lj_strfmt_putptr(SBuf *sb, const void *v)
 }
 
 /* Add quoted string to buffer. */
-SBuf * LJ_FASTCALL lj_strfmt_putquoted(SBuf *sb, GCstr *str)
+static SBuf *strfmt_putquotedlen(SBuf *sb, const char *s, MSize len)
 {
-  const char *s = strdata(str);
-  MSize len = str->len;
   lj_buf_putb(sb, '"');
   while (len--) {
     uint32_t c = (uint32_t)(uint8_t)*s++;
@@ -225,6 +226,13 @@ SBuf * LJ_FASTCALL lj_strfmt_putquoted(SBuf *sb, GCstr *str)
   return sb;
 }
 
+#if LJ_HASJIT
+SBuf * LJ_FASTCALL lj_strfmt_putquoted(SBuf *sb, GCstr *str)
+{
+  return strfmt_putquotedlen(sb, strdata(str), str->len);
+}
+#endif
+
 /* -- Formatted conversions to buffer ------------------------------------- */
 
 /* Add formatted char to buffer. */
@@ -240,18 +248,26 @@ SBuf *lj_strfmt_putfchar(SBuf *sb, SFormat sf, int32_t c)
 }
 
 /* Add formatted string to buffer. */
-SBuf *lj_strfmt_putfstr(SBuf *sb, SFormat sf, GCstr *str)
+static SBuf *strfmt_putfstrlen(SBuf *sb, SFormat sf, const char *s, MSize len)
 {
-  MSize len = str->len <= STRFMT_PREC(sf) ? str->len : STRFMT_PREC(sf);
   MSize width = STRFMT_WIDTH(sf);
-  char *w = lj_buf_more(sb, width > len ? width : len);
+  char *w;
-  if ((sf & STRFMT_F_LEFT)) w = lj_buf_wmem(w, strdata(str), len);
+  if (len > STRFMT_PREC(sf)) len = STRFMT_PREC(sf);
+  w = lj_buf_more(sb, width > len ? width : len);
+  if ((sf & STRFMT_F_LEFT)) w = lj_buf_wmem(w, s, len);
   while (width-- > len) *w++ = ' ';
-  if (!(sf & STRFMT_F_LEFT)) w = lj_buf_wmem(w, strdata(str), len);
+  if (!(sf & STRFMT_F_LEFT)) w = lj_buf_wmem(w, s, len);
   sb->w = w;
   return sb;
 }
 
+#if LJ_HASJIT
+SBuf *lj_strfmt_putfstr(SBuf *sb, SFormat sf, GCstr *str)
+{
+  return strfmt_putfstrlen(sb, sf, strdata(str), str->len);
+}
+#endif
+
 /* Add formatted signed/unsigned integer to buffer. */
 SBuf *lj_strfmt_putfxint(SBuf *sb, SFormat sf, uint64_t k)
 {
@@ -346,6 +362,91 @@ SBuf *lj_strfmt_putfnum_uint(SBuf *sb, SFormat sf, lua_Number n)
   return lj_strfmt_putfxint(sb, sf, (uint64_t)k);
 }
 
+/* Format stack arguments to buffer. */
+int lj_strfmt_putarg(lua_State *L, SBuf *sb, int arg, int retry)
+{
+  int narg = (int)(L->top - L->base);
+  GCstr *fmt = lj_lib_checkstr(L, arg);
+  FormatState fs;
+  SFormat sf;
+  lj_strfmt_init(&fs, strdata(fmt), fmt->len);
+  while ((sf = lj_strfmt_parse(&fs)) != STRFMT_EOF) {
+    if (sf == STRFMT_LIT) {
+      lj_buf_putmem(sb, fs.str, fs.len);
+    } else if (sf == STRFMT_ERR) {
+      lj_err_callerv(L, LJ_ERR_STRFMT,
+		     strdata(lj_str_new(L, fs.str, fs.len)));
+    } else {
+      TValue *o = &L->base[arg++];
+      if (arg > narg)
+	lj_err_arg(L, arg, LJ_ERR_NOVAL);
+      switch (STRFMT_TYPE(sf)) {
+      case STRFMT_INT:
+	if (tvisint(o)) {
+	  int32_t k = intV(o);
+	  if (sf == STRFMT_INT)
+	    lj_strfmt_putint(sb, k);  /* Shortcut for plain %d. */
+	  else
+	    lj_strfmt_putfxint(sb, sf, k);
+	} else {
+	  lj_strfmt_putfnum_int(sb, sf, lj_lib_checknum(L, arg));
+	}
+	break;
+      case STRFMT_UINT:
+	if (tvisint(o))
+	  lj_strfmt_putfxint(sb, sf, intV(o));
+	else
+	  lj_strfmt_putfnum_uint(sb, sf, lj_lib_checknum(L, arg));
+	break;
+      case STRFMT_NUM:
+	lj_strfmt_putfnum(sb, sf, lj_lib_checknum(L, arg));
+	break;
+      case STRFMT_STR: {
+	MSize len;
+	const char *s;
+	cTValue *mo;
+	if (LJ_UNLIKELY(!tvisstr(o)) && retry >= 0 &&
+	    !tvisnil(mo = lj_meta_lookup(L, o, MM_tostring))) {
+	  /* Call __tostring metamethod once. */
+	  copyTV(L, L->top++, mo);
+	  copyTV(L, L->top++, o);
+	  lua_call(L, 1, 1);
+	  o = &L->base[arg-1];  /* Stack may have been reallocated. */
+	  copyTV(L, o, --L->top);  /* Replace inline for retry. */
+	  if (retry < 2) {  /* Global buffer may have been overwritten. */
+	    retry = 1;
+	    break;
+	  }
+	}
+	if (LJ_LIKELY(tvisstr(o))) {
+	  len = strV(o)->len;
+	  s = strVdata(o);
+	} else {
+	  GCstr *str = lj_strfmt_obj(L, o);
+	  len = str->len;
+	  s = strdata(str);
+	}
+	if ((sf & STRFMT_T_QUOTED))
+	  strfmt_putquotedlen(sb, s, len);  /* No formatting. */
+	else
+	  strfmt_putfstrlen(sb, sf, s, len);
+	break;
+	}
+      case STRFMT_CHAR:
+	lj_strfmt_putfchar(sb, sf, lj_lib_checkint(L, arg));
+	break;
+      case STRFMT_PTR:  /* No formatting. */
+	lj_strfmt_putptr(sb, lj_obj_ptr(G(L), o));
+	break;
+      default:
+	lj_assertL(0, "bad string format type");
+	break;
+      }
+    }
+  }
+  return retry;
+}
+
 /* -- Conversions to strings ---------------------------------------------- */
 
 /* Convert integer to string. */

src/lj_strfmt.h

@@ -95,7 +95,9 @@ LJ_FUNC SBuf * LJ_FASTCALL lj_strfmt_putint(SBuf *sb, int32_t k);
 LJ_FUNC SBuf * LJ_FASTCALL lj_strfmt_putnum(SBuf *sb, cTValue *o);
 #endif
 LJ_FUNC SBuf * LJ_FASTCALL lj_strfmt_putptr(SBuf *sb, const void *v);
+#if LJ_HASJIT
 LJ_FUNC SBuf * LJ_FASTCALL lj_strfmt_putquoted(SBuf *sb, GCstr *str);
+#endif
 
 /* Formatted conversions to buffer. */
 LJ_FUNC SBuf *lj_strfmt_putfxint(SBuf *sb, SFormat sf, uint64_t k);
@@ -103,7 +105,10 @@ LJ_FUNC SBuf *lj_strfmt_putfnum_int(SBuf *sb, SFormat sf, lua_Number n);
 LJ_FUNC SBuf *lj_strfmt_putfnum_uint(SBuf *sb, SFormat sf, lua_Number n);
 LJ_FUNC SBuf *lj_strfmt_putfnum(SBuf *sb, SFormat, lua_Number n);
 LJ_FUNC SBuf *lj_strfmt_putfchar(SBuf *sb, SFormat, int32_t c);
+#if LJ_HASJIT
 LJ_FUNC SBuf *lj_strfmt_putfstr(SBuf *sb, SFormat, GCstr *str);
+#endif
+LJ_FUNC int lj_strfmt_putarg(lua_State *L, SBuf *sb, int arg, int retry);
 
 /* Conversions to strings. */
 LJ_FUNC GCstr * LJ_FASTCALL lj_strfmt_int(lua_State *L, int32_t k);