Improving portability. It is no longer assumed that pointers

can be converted to a type long and vice versa. The configure
script tries to detect the appropriate type to use. Also,
some run-time checking is performed in the streams module
to detect which conversions specifier strings to use for
printing numbers.

19 files changed, 377 insertions, 104 deletions

diff --git a/ChangeLog b/ChangeLog
index e9319f9d..bdbda5c6 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,73 @@
+2009-11-23  Kaz Kylheku  <kkylheku@gmail.com>
+
+	Improving portability. It is no longer assumed that pointers
+	can be converted to a type long and vice versa. The configure
+	script tries to detect the appropriate type to use. Also,
+	some run-time checking is performed in the streams module
+	to detect which conversions specifier strings to use for
+	printing numbers.
+
+	* Makefile (conftest, conftest.o, conftest.syms): New targets.
+	Used by the configure script. 
+
+	* configure (intptr, nm): New configuration variables.
+	Generating config.make is no longer the last step; compiler tests are
+	performed after config.make is set up, so that rules in the Makefile
+	can be used for doing the compiling. (This is the cleanest way to do
+	it, since the paths to the tools may contain Make variable expansion
+	syntax).  New steps are added to try to detect whether the compiler has
+	a wider integer type than the c89 long, and which of the available
+	types (including, potentially, the extra wide type) is suitable for
+	holding a pointer. Results are generated into a header config.h.
+
+	* dep.mk: Regenerated.
+
+	* lib.h (NUM_MAX, NUM_MIN): Now derived from INT_PTR_MAX and
+	INT_PTR_MIN macros, which come from config.h.
+	(cnum): New typedef name.
+	(cobj ops, tag, auto_str, static_str, litptr, lit_noex):
+	Changed long to cnum.
+	(num, c_num): Declaration updated.
+
+	* lib.c (equal, length, num, c_num, plus, minus, neg, search_str,
+	cat_str, vector, vec_set_fill, obj_print, obj_pprint): Changed
+	long to cnum.
+
+	* gc.c (mark_obj): Changed long to cnum.
+
+	* hash.c (stuct hash, ll_hash, hash_mark, hash_grow,
+	hash_process_weak): Changed long to cnum.
+
+	* match.c (complex_open, do_output_line, do_output, match_files):
+	Changed long to cnum.
+
+	* parser.h (lineno): Declaration updated.
+
+	* parser.l (lineno): Redefined as cnum.
+	(grammar): Changed long to cnum.
+
+	* parser.y (%union/yystype): num member changed to cnum.
+	of config.h added.
+
+	* regex.c (nfa_run, nfa_machine_match_span, search_regex):
+	Changed long to cnum.
+
+	* regex.h (struct nfa_machine): Members last_accept_pos and count
+	changed to cnum.
+	(nfa_run, nfa_machine_match_span): Declarations updated.
+
+	* stream.c (struct fmt): New type.
+	(fmt_tab): New static array.
+	(num_fmt): New static pointer.
+	(detect_format_string): New function.
+	(vformat): Changed long to cnum. Formatting of numbers uses
+	num_fmt.
+	(stream_init): Call detect_format_string.
+
+	* txr.c, unwind.c, utf8.c: include config.h.
+
+	* unwind.h (internal_error): Local declaration of num updated.
+
 2009-11-21  Kaz Kylheku  <kkylheku@gmail.com>
 
 	Introducing symbol packages. Internal symbols are now in
diff --git a/Makefile b/Makefile
index dddce7d2..809e7eda 100644
--- a/Makefile
+++ b/Makefile
@@ -59,7 +59,7 @@ clean:
 
 .PHONY: distclean
 distclean: clean
-	rm -f config.make config.log
+	rm -f config.h config.make config.log
 
 .PHONY: depend
 depend:
@@ -101,3 +101,17 @@ install: $(PROG)
 config.make:
 	@echo "config.make missing: you didn't run ./configure"
 	@exit 1
+
+#
+# Special targets used by ./configure
+#
+
+conftest: conftest.c
+	$(CC) -o $@ $^
+
+conftest.o: CFLAGS := $(LANG_FLAGS) 
+
+conftest.o: conftest.c
+
+conftest.syms: conftest.o
+	$(NM) -t o -P $^ > $@
diff --git a/configure b/configure
index fdd97854..356338d4 100755
--- a/configure
+++ b/configure
@@ -104,10 +104,12 @@ mandir=${mandir-'$(prefix)/share/man'}
 cross=${cross-}
 compiler_prefix=${compiler_prefix-}
 cc=${cc-'$(cross)$(compiler_prefix)gcc'}
+intptr=
 tool_prefix=${tool_prefix-}
 lex=${lex-'$(cross)$(tool_prefix)flex'}
 lexlib=${lexlib--lfl}
 yacc=${yacc-'$(cross)$(tool_prefix)yacc'}
+nm=${nm-'$(cross)$(tool_prefix)nm'}
 opt_flags=${opt_flags--O2}
 lang_flags=${lang_flags--ansi -std=c89 -D_POSIX_C_SOURCE=2}
 diag_flags=${diag_flags--Wall}
@@ -208,6 +210,12 @@ cc [$cc]
   compiling C sources to object files, and for linking object files to
   executables. This becomes the CC variable in the Makefile.
 
+intptr [$intptr]
+
+  Specifies the name of the C integer type wide enough such that a pointer
+  value can be converted to it. If this is blank, the configure script
+  will try to auto detect it.
+
 tool_prefix [$tool_prefix]
 
   Specifies a prefix to be added to tool commands other than the
@@ -226,6 +234,10 @@ yacc [$yacc]
 
   Specifies the program to use for compiling yacc scanners to C.
 
+nm [$nm]
+
+  Specifies the nm program for dumping symbols from an object file.
+
 opt_flags [$opt_flags]
 
   Specifies optimization flags to use for compiling and linking
@@ -403,20 +415,9 @@ else
 fi
 
 #
-# Save configuration in config.log
-#
-cat > config.log <<!
-
-Configured on $(date) using
-
-   $cmdline
-
-!
-
-#
 # Finally, we generate config.make
 #
-printf "generating config.make\n"
+printf "generating config.make ...\n"
 
 cat > config.make <<!
 # absolute path to source code directory
@@ -454,6 +455,7 @@ CC := $cc
 LEX := $lex
 LEXLIB := $lexlib
 YACC := $yacc
+NM := $nm
 
 OPT_FLAGS := $opt_flags
 LANG_FLAGS := $lang_flags
@@ -465,9 +467,137 @@ TXR_DBG_OPTS := $txr_dbg_opts
 !
 
 #
+# Start config.h header
+#
+> config.h
+
+#
+# Check C compiler sanity
+#
+printf "Checking whether your C compiler can make a simple executable ... "
+
+cat > conftest.c <<!
+#include <stdio.h>
+int main(void)
+{
+  printf("Hello, world!\n");
+  return 0;
+}
+!
+
+if ! make conftest > /dev/null 2>&1 || ! [ -x conftest ] ; then
+  printf "failed\n"
+  exit 1
+fi
+
+rm -f conftest
+printf "okay\n"
+
+#
+# Check what kind of C type we have for integers wider than long,
+# if any.
+#
+printf "Checking what C type we have for integers wider than \"long\" ... "
+
+for try_type in int64 __int64 "long long" ; do
+  cat > conftest.c <<!
+$try_type value;
+!
+  rm -f conftest.o
+  if make conftest.o > /dev/null 2>&1 ; then
+    longlong=$try_type
+    break
+  fi
+done
+
+if [ -n "$longlong" ] ; then
+  printf '"%s"\n' "$longlong"
+  printf "#define HAVE_LONGLONG_T 1\n" >> config.h
+  printf "typedef $longlong longlong_t;\n" >> config.h
+else
+  printf "none\n"
+fi
+
+printf "Checking what C integer type can hold a pointer ... "
+
+if [ -z "$intptr" ] ; then
+  cat > conftest.c <<!
+#include "config.h"
+char sizeof_ptr[sizeof (char *)];
+char sizeof_short[sizeof (short)];
+char sizeof_int[sizeof (int)];
+char sizeof_long[sizeof (long)];
+#ifdef HAVE_LONGLONG_T
+char sizeof_longlong_t[sizeof (longlong_t)];
+#endif
+!
+  rm -f conftest.o conftest.syms
+
+  if ! make conftest.syms > /dev/null 2>&1 ; then
+    echo "failed"
+    exit 1;
+  fi
+
+  sizeof_ptr=0
+  sizeof_short=0
+  sizeof_int=0
+  sizeof_long=0
+  sizeof_longlong_t=0
+
+  while read symbol type offset size ; do
+    eval "size=$(( 0$size + 0 ))"
+    eval $(printf "%s=%d\n" "$symbol" "$size")
+  done < conftest.syms
+
+  rm -f conftest.syms conftest.o
+
+  if [ $sizeof_ptr -eq 0 ] ; then
+    printf "failed\n"
+    exit 1;
+  fi
+
+  if [ $sizeof_ptr -eq $sizeof_short ] ; then
+    intptr="short"
+  elif [ $sizeof_ptr -eq $sizeof_int ] ; then
+    intptr="int"
+  elif [ $sizeof_ptr -eq $sizeof_long ] ; then
+    intptr="long"
+  elif [ $sizeof_ptr -eq $sizeof_long_long_t ] ; then
+    intptr="longlong_t"
+  fi
+
+  if [ -z "$intptr" ] ; then
+    printf "failed\n"
+    exit 1;
+  fi
+fi
+
+printf '"%s"\n' "$intptr"
+printf "typedef $intptr int_ptr_t;\n" >> config.h
+intptr_max=$(( (1 << ( sizeof_ptr * 8  - 1 )) - 1 ))
+printf "#define INT_PTR_MAX %d\n" $intptr_max >> config.h
+printf "#define INT_PTR_MIN -%d\n" $intptr_max >> config.h
+
+#
+# Clean up
+#
+rm -f conftest conftest.c conftest.o conftest.syms
+
+#
+# Save configuration in config.log
+#
+cat > config.log <<!
+
+Configured on $(date) using
+
+   $cmdline
+
+!
+#
 # Parting message#
 #
 cat <<!
+
 Configuration seems to have been successful. The next
 step is one of these two.
 
diff --git a/dep.mk b/dep.mk
index 4ef2d79f..c5aefeae 100644
--- a/dep.mk
+++ b/dep.mk
@@ -1,12 +1,15 @@
-parser.tab.o: $(top_srcdir)/lib.h $(top_srcdir)/regex.h $(top_srcdir)/parser.h
-utf8.o: $(top_srcdir)/lib.h $(top_srcdir)/utf8.h
-lib.o: $(top_srcdir)/lib.h $(top_srcdir)/gc.h $(top_srcdir)/unwind.h $(top_srcdir)/stream.h $(top_srcdir)/utf8.h
-lex.yy.o: y.tab.h $(top_srcdir)/lib.h $(top_srcdir)/gc.h $(top_srcdir)/stream.h $(top_srcdir)/utf8.h $(top_srcdir)/parser.h
-regex.o: $(top_srcdir)/lib.h $(top_srcdir)/unwind.h $(top_srcdir)/regex.h
-y.tab.o: $(top_srcdir)/lib.h $(top_srcdir)/regex.h $(top_srcdir)/utf8.h $(top_srcdir)/parser.h
-unwind.o: $(top_srcdir)/lib.h $(top_srcdir)/gc.h $(top_srcdir)/stream.h $(top_srcdir)/txr.h $(top_srcdir)/unwind.h
-txr.o: $(top_srcdir)/lib.h $(top_srcdir)/stream.h $(top_srcdir)/gc.h $(top_srcdir)/unwind.h $(top_srcdir)/parser.h $(top_srcdir)/match.h $(top_srcdir)/utf8.h $(top_srcdir)/txr.h
-match.o: $(top_srcdir)/lib.h $(top_srcdir)/gc.h $(top_srcdir)/unwind.h $(top_srcdir)/regex.h $(top_srcdir)/stream.h $(top_srcdir)/parser.h $(top_srcdir)/txr.h $(top_srcdir)/utf8.h $(top_srcdir)/match.h
-stream.o: $(top_srcdir)/lib.h $(top_srcdir)/gc.h $(top_srcdir)/unwind.h $(top_srcdir)/stream.h $(top_srcdir)/utf8.h
-gc.o: $(top_srcdir)/lib.h $(top_srcdir)/stream.h $(top_srcdir)/hash.h $(top_srcdir)/txr.h $(top_srcdir)/gc.h
-hash.o: $(top_srcdir)/lib.h $(top_srcdir)/gc.h $(top_srcdir)/unwind.h $(top_srcdir)/hash.h
+parser.tab.o: $(top_srcdir)/config.h $(top_srcdir)/lib.h $(top_srcdir)/regex.h $(top_srcdir)/parser.h
+utf8.o: $(top_srcdir)/config.h $(top_srcdir)/lib.h $(top_srcdir)/utf8.h
+lib.o: $(top_srcdir)/config.h $(top_srcdir)/lib.h $(top_srcdir)/gc.h $(top_srcdir)/hash.h $(top_srcdir)/unwind.h $(top_srcdir)/stream.h $(top_srcdir)/utf8.h
+lex.yy.o: $(top_srcdir)/config.h $(top_srcdir)/lib.h y.tab.h $(top_srcdir)/gc.h $(top_srcdir)/stream.h $(top_srcdir)/utf8.h $(top_srcdir)/parser.h
+popen_getwc.o:
+test-sprintf.o:
+test.o:
+regex.o: $(top_srcdir)/config.h $(top_srcdir)/lib.h $(top_srcdir)/unwind.h $(top_srcdir)/regex.h
+y.tab.o: $(top_srcdir)/config.h $(top_srcdir)/lib.h $(top_srcdir)/regex.h $(top_srcdir)/utf8.h $(top_srcdir)/parser.h
+unwind.o: $(top_srcdir)/config.h $(top_srcdir)/lib.h $(top_srcdir)/gc.h $(top_srcdir)/stream.h $(top_srcdir)/txr.h $(top_srcdir)/unwind.h
+txr.o: $(top_srcdir)/config.h $(top_srcdir)/lib.h $(top_srcdir)/stream.h $(top_srcdir)/gc.h $(top_srcdir)/unwind.h $(top_srcdir)/parser.h $(top_srcdir)/match.h $(top_srcdir)/utf8.h $(top_srcdir)/txr.h
+match.o: $(top_srcdir)/config.h $(top_srcdir)/lib.h $(top_srcdir)/gc.h $(top_srcdir)/unwind.h $(top_srcdir)/regex.h $(top_srcdir)/stream.h $(top_srcdir)/parser.h $(top_srcdir)/txr.h $(top_srcdir)/utf8.h $(top_srcdir)/match.h
+stream.o: $(top_srcdir)/config.h $(top_srcdir)/lib.h $(top_srcdir)/gc.h $(top_srcdir)/unwind.h $(top_srcdir)/stream.h $(top_srcdir)/utf8.h
+gc.o: $(top_srcdir)/config.h $(top_srcdir)/lib.h $(top_srcdir)/stream.h $(top_srcdir)/hash.h $(top_srcdir)/txr.h $(top_srcdir)/gc.h
+hash.o: $(top_srcdir)/config.h $(top_srcdir)/lib.h $(top_srcdir)/gc.h $(top_srcdir)/unwind.h $(top_srcdir)/hash.h
diff --git a/gc.c b/gc.c
index 0de0c3f2..ec192a9e 100644
--- a/gc.c
+++ b/gc.c
@@ -31,6 +31,7 @@
 #include <setjmp.h>
 #include <dirent.h>
 #include <wchar.h>
+#include "config.h"
 #include "lib.h"
 #include "stream.h"
 #include "hash.h"
@@ -231,7 +232,7 @@ tail_call:
     {
       val alloc_size = obj->v.vec[-2];
       val fill_ptr = obj->v.vec[-1];
-      long i, fp = c_num(fill_ptr);
+      cnum i, fp = c_num(fill_ptr);
 
       mark_obj(alloc_size);
       mark_obj(fill_ptr);
diff --git a/hash.c b/hash.c
index 81d2709b..965dd4d3 100644
--- a/hash.c
+++ b/hash.c
@@ -32,6 +32,7 @@
 #include <assert.h>
 #include <setjmp.h>
 #include <limits.h>
+#include "config.h"
 #include "lib.h"
 #include "gc.h"
 #include "unwind.h"
@@ -48,8 +49,8 @@ struct hash {
   hash_flags_t flags;
   struct hash *next;
   val table;
-  long modulus;
-  long count;
+  cnum modulus;
+  cnum count;
 };
 
 /*
@@ -76,7 +77,7 @@ static long hash_c_str(const wchar_t *str)
   return h;
 }
 
-static long ll_hash(val obj)
+static cnum ll_hash(val obj)
 {
   if (obj == nil)
     return NUM_MAX;
@@ -94,14 +95,14 @@ static long ll_hash(val obj)
     return c_num(obj) & NUM_MAX;
   case SYM:
   case PKG:
-    return ((long) obj) & NUM_MAX;
+    return ((cnum) obj) & NUM_MAX;
   case FUN:
-    return ((long) obj->f.f.interp_fun + ll_hash(obj->f.env)) & NUM_MAX;
+    return ((cnum) obj->f.f.interp_fun + ll_hash(obj->f.env)) & NUM_MAX;
   case VEC:
     {
       val fill = obj->v.vec[vec_fill];
-      long i, h = ll_hash(obj->v.vec[vec_fill]);
-      long len = c_num(fill);
+      cnum i, h = ll_hash(obj->v.vec[vec_fill]);
+      cnum len = c_num(fill);
 
       for (i = 0; i < len; i++)
         h = (h + ll_hash(obj->v.vec[i])) & NUM_MAX;
@@ -116,7 +117,7 @@ static long ll_hash(val obj)
   case COBJ:
     if (obj->co.ops->hash)
       return obj->co.ops->hash(obj);
-    return ((long) obj) & NUM_MAX;
+    return ((cnum) obj) & NUM_MAX;
   }
 
   internal_error("unhandled case in equal function");
@@ -140,7 +141,7 @@ void hash_destroy(val hash)
 void hash_mark(val hash)
 {
   struct hash *h = (struct hash *) hash->co.handle;
-  long i;
+  cnum i;
 
   switch (h->flags) {
   case hash_weak_none:
@@ -195,8 +196,8 @@ static struct cobj_ops hash_ops = {
 
 void hash_grow(struct hash *h)
 {
-  long i;
-  long new_modulus = 2 * h->modulus;
+  cnum i;
+  cnum new_modulus = 2 * h->modulus;
   val new_table = vector(num(new_modulus));
 
   bug_unless (new_modulus > h->modulus);
@@ -276,7 +277,7 @@ val remhash(val hash, val key)
 void hash_process_weak(void)
 {
   struct hash *h;
-  long i;
+  cnum i;
 
   for (h = reachable_weak_hashes; h != 0; h = h->next) {
     switch (h->flags) {
diff --git a/lib.c b/lib.c
index 1be02723..55913d70 100644
--- a/lib.c
+++ b/lib.c
@@ -34,6 +34,7 @@
 #include <dirent.h>
 #include <setjmp.h>
 #include <wchar.h>
+#include "config.h"
 #include "lib.h"
 #include "gc.h"
 #include "hash.h"
@@ -410,7 +411,7 @@ val flatten(val list)
   return mappend(func_f1(nil, flatten_helper), list);
 }
 
-long c_num(val num);
+cnum c_num(val num);
 
 val equal(val left, val right)
 {
@@ -489,7 +490,7 @@ val equal(val left, val right)
     return nil;
   case VEC:
     if (type(right) == VEC) {
-      long i, fill;
+      cnum i, fill;
       if (!equal(left->v.vec[vec_fill], right->v.vec[vec_fill]))
         return nil;
       fill = c_num(left->v.vec[vec_fill]);
@@ -620,7 +621,7 @@ val proper_listp(val obj)
 
 val length(val list)
 {
-  long len = 0;
+  cnum len = 0;
   while (consp(list)) {
     len++;
     list = cdr(list);
@@ -628,17 +629,17 @@ val length(val list)
   return num(len);
 }
 
-val num(long n)
+val num(cnum n)
 {
   numeric_assert (n >= NUM_MIN && n <= NUM_MAX);
   return (val) ((n << TAG_SHIFT) | TAG_NUM);
 }
 
-long c_num(val num)
+cnum c_num(val num)
 {
   if (!is_num(num))
     type_mismatch(lit("~s is not a number"), num, nao);
-  return ((long) num) >> TAG_SHIFT;
+  return ((cnum) num) >> TAG_SHIFT;
 }
 
 val nump(val num)
@@ -648,8 +649,8 @@ val nump(val num)
 
 val plus(val anum, val bnum)
 {
-  long a = c_num(anum);
-  long b = c_num(bnum);
+  cnum a = c_num(anum);
+  cnum b = c_num(bnum);
 
   numeric_assert (a <= 0 || b <= 0 || NUM_MAX - b >= a);
   numeric_assert (a >= 0 || b >= 0 || NUM_MIN - b >= a);
@@ -659,8 +660,8 @@ val plus(val anum, val bnum)
 
 val minus(val anum, val bnum)
 {
-  long a = c_num(anum);
-  long b = c_num(bnum);
+  cnum a = c_num(anum);
+  cnum b = c_num(bnum);
 
   numeric_assert (b != NUM_MIN || NUM_MIN == -NUM_MAX);
   numeric_assert (a <= 0 || -b <= 0 || NUM_MAX + b >= a);
@@ -671,7 +672,7 @@ val minus(val anum, val bnum)
 
 val neg(val anum)
 {
-  long n = c_num(anum);
+  cnum n = c_num(anum);
   return num(-n);
 }
 
@@ -838,8 +839,8 @@ val search_str(val haystack, val needle, val start_num, val from_end)
     return nil;
   } else {
     val h_is_lazy = lazy_stringp(haystack);
-    long start = c_num(start_num);
-    long good = -1, pos = -1;
+    cnum start = c_num(start_num);
+    cnum good = -1, pos = -1;
     const wchar_t *n = c_str(needle), *h;
 
     if (!h_is_lazy) {
@@ -915,10 +916,10 @@ val sub_str(val str_in, val from, val to)
 
 val cat_str(val list, val sep)
 {
-  long total = 0;
+  cnum total = 0;
   val iter;
   wchar_t *str, *ptr;
-  long len_sep = sep ? c_num(length_str(sep)) : 0;
+  cnum len_sep = sep ? c_num(length_str(sep)) : 0;
 
   for (iter = list; iter != nil; iter = cdr(iter)) {
     val item = car(iter);
@@ -943,7 +944,7 @@ val cat_str(val list, val sep)
 
   for (ptr = str, iter = list; iter != nil; iter = cdr(iter)) {
     val item = car(iter);
-    long len;
+    cnum len;
     if (!item)
       continue;
     if (stringp(item)) {
@@ -1364,7 +1365,7 @@ val chain(val fun1_list)
 
 val vector(val alloc)
 {
-  long alloc_plus = c_num(alloc) + 2;
+  cnum alloc_plus = c_num(alloc) + 2;
   val vec = make_obj();
   val *v = (val *) chk_malloc(alloc_plus * sizeof *v);
   vec->v.type = VEC;
@@ -1385,14 +1386,14 @@ val vec_set_fill(val vec, val fill)
   type_check(vec, VEC);
 
   {
-    long new_fill = c_num(fill);
-    long old_fill = c_num(vec->v.vec[vec_fill]);
-    long old_alloc = c_num(vec->v.vec[vec_alloc]);
-    long fill_delta = new_fill - old_fill;
-    long alloc_delta = new_fill - old_alloc;
+    cnum new_fill = c_num(fill);
+    cnum old_fill = c_num(vec->v.vec[vec_fill]);
+    cnum old_alloc = c_num(vec->v.vec[vec_alloc]);
+    cnum fill_delta = new_fill - old_fill;
+    cnum alloc_delta = new_fill - old_alloc;
 
     if (alloc_delta > 0) {
-      long new_alloc = max(new_fill, 2*old_alloc);
+      cnum new_alloc = max(new_fill, 2*old_alloc);
       val *newvec = (val *) chk_realloc(vec->v.vec - 2,
                                         (new_alloc + 2)*sizeof *newvec);
       vec->v.vec = newvec + 2;
@@ -1400,7 +1401,7 @@ val vec_set_fill(val vec, val fill)
     }
 
     if (fill_delta > 0) {
-      long i;
+      cnum i;
       for (i = old_fill; i < new_fill; i++)
         vec->v.vec[i] = nil;
     }
@@ -2011,7 +2012,7 @@ void obj_print(val obj, val out)
     return;
   case VEC:
     {
-      long i, fill = c_num(obj->v.vec[vec_fill]);
+      cnum i, fill = c_num(obj->v.vec[vec_fill]);
       put_string(out, lit("#("));
       for (i = 0; i < fill; i++) {
         obj_print(obj->v.vec[i], out);
@@ -2081,7 +2082,7 @@ void obj_pprint(val obj, val out)
     return;
   case VEC:
     {
-      long i, fill = c_num(obj->v.vec[vec_fill]);
+      cnum i, fill = c_num(obj->v.vec[vec_fill]);
       put_string(out, lit("#("));
       for (i = 0; i < fill; i++) {
         obj_pprint(obj->v.vec[i], out);
diff --git a/lib.h b/lib.h
index 71fd4668..c3e9e4a7 100644
--- a/lib.h
+++ b/lib.h
@@ -30,8 +30,10 @@
 #define TAG_NUM 1
 #define TAG_CHR 2
 #define TAG_LIT 3
-#define NUM_MAX (LONG_MAX/4)
-#define NUM_MIN (LONG_MIN/4)
+#define NUM_MAX (INT_PTR_MAX/4)
+#define NUM_MIN (INT_PTR_MIN/4)
+
+typedef int_ptr_t cnum;
 
 typedef enum type {
   NUM = TAG_NUM, CHR = TAG_CHR, LIT = TAG_LIT, CONS,
@@ -145,7 +147,7 @@ struct cobj_ops {
   void (*print)(val self, val stream);
   void (*destroy)(val self);
   void (*mark)(val self);
-  long (*hash)(val self);
+  cnum (*hash)(val self);
 };
 
 union obj {
@@ -161,7 +163,7 @@ union obj {
   struct cobj co;
 };
 
-inline long tag(val obj) { return ((long) obj) & TAG_MASK; }
+inline cnum tag(val obj) { return ((cnum) obj) & TAG_MASK; }
 inline int is_ptr(val obj) { return obj && tag(obj) == TAG_PTR; }
 inline int is_num(val obj) { return tag(obj) == TAG_NUM; }
 inline int is_chr(val obj) { return tag(obj) == TAG_CHR; }
@@ -174,20 +176,20 @@ inline type_t type(val obj)
 
 inline val auto_str(const wchar_t *str)
 {
-  return (val) ((long) (str) | TAG_LIT);
+  return (val) ((cnum) (str) | TAG_LIT);
 }
 
 inline val static_str(const wchar_t *str)
 {
-  return (val) ((long) (str) | TAG_LIT);
+  return (val) ((cnum) (str) | TAG_LIT);
 }
 
 inline wchar_t *litptr(val obj)
 {
- return (wchar_t *) ((long) obj & ~TAG_MASK);
+ return (wchar_t *) ((cnum) obj & ~TAG_MASK);
 }
 
-#define lit_noex(strlit) ((obj_t *) ((long) (L ## strlit) | TAG_LIT))
+#define lit_noex(strlit) ((obj_t *) ((cnum) (L ## strlit) | TAG_LIT))
 #define lit(strlit) lit_noex(strlit)
 
 extern val keyword_package;
@@ -245,7 +247,7 @@ val tree_find(val obj, val tree);
 val some_satisfy(val list, val pred, val key);
 val all_satisfy(val list, val pred, val key);
 val none_satisfy(val list, val pred, val key);
-long c_num(val num);
+cnum c_num(val num);
 val nump(val num);
 val equal(val left, val right);
 unsigned char *chk_malloc(size_t size);
@@ -259,8 +261,8 @@ val atom(val obj);
 val listp(val obj);
 val proper_listp(val obj);
 val length(val list);
-val num(long val);
-long c_num(val num);
+val num(cnum val);
+cnum c_num(val num);
 val plus(val anum, val bnum);
 val minus(val anum, val bnum);
 val neg(val num);
diff --git a/match.c b/match.c
index 937d3223..9cb30ee9 100644
--- a/match.c
+++ b/match.c
@@ -33,6 +33,7 @@
 #include <setjmp.h>
 #include <stdarg.h>
 #include <wchar.h>
+#include "config.h"
 #include "lib.h"
 #include "gc.h"
 #include "unwind.h"
@@ -633,7 +634,7 @@ fpip_t complex_open(val name, val output)
   fpip_t ret = { 0, 0 };
 
   const wchar_t *namestr = c_str(name);
-  long len = c_num(length_str(name));
+  cnum len = c_num(length_str(name));
 
   if (len == 0)
     return ret;
@@ -802,7 +803,7 @@ void do_output_line(val bindings, val specline,
             val bind_a = mapcar(func_n1(bind_car), bind_cp);
             do_output_line(bind_a, single_clauses, spec_lineno, out);
           } else if (!zerop(max_depth)) {
-            long i;
+            cnum i;
 
             for (i = 0; i < c_num(max_depth); i++) {
               val bind_a = mapcar(func_n1(bind_car), bind_cp);
@@ -868,7 +869,7 @@ void do_output(val bindings, val specs, val out)
           val bind_a = mapcar(func_n1(bind_car), bind_cp);
           do_output(bind_a, single_clauses, out);
         } else if (!zerop(max_depth)) {
-          long i;
+          cnum i;
 
           for (i = 0; i < c_num(max_depth); i++) {
             val bind_a = mapcar(func_n1(bind_car), bind_cp);
@@ -899,7 +900,7 @@ val match_files(val spec, val files,
                 val data_linenum)
 {
   val data = nil;
-  long data_lineno = 0;
+  cnum data_lineno = 0;
 
   if (listp(first_file_parsed)) {
     data = first_file_parsed;
@@ -950,8 +951,8 @@ repeat_spec_same_data:
 
       if (sym == skip) {
         val max = first(rest(first_spec));
-        long cmax = nump(max) ? c_num(max) : 0;
-        long reps = 0;
+        cnum cmax = nump(max) ? c_num(max) : 0;
+        cnum reps = 0;
 
         if (rest(specline))
           sem_error(spec_linenum,
@@ -1265,7 +1266,7 @@ repeat_spec_same_data:
           if (success) {
             if (consp(success)) {
               cons_bind (new_data, new_line, success);
-              long new_lineno = c_num(new_line);
+              cnum new_lineno = c_num(new_line);
 
               bug_unless (new_lineno >= data_lineno);
 
diff --git a/parser.h b/parser.h
index 7bad9451..2a19a349 100644
--- a/parser.h
+++ b/parser.h
@@ -25,7 +25,7 @@
  */
 
 #include <stdio.h>
-extern long lineno;
+extern cnum lineno;
 extern int errors;
 extern val yyin_stream;
 extern const wchar_t *spec_file;
diff --git a/parser.l b/parser.l
index 564f9730..433c0489 100644
--- a/parser.l
+++ b/parser.l
@@ -34,8 +34,9 @@
 #include <errno.h>
 #include <dirent.h>
 #include <wchar.h>
-#include "y.tab.h"
+#include "config.h"
 #include "lib.h"
+#include "y.tab.h"
 #include "gc.h"
 #include "stream.h"
 #include "utf8.h"
@@ -58,7 +59,7 @@
 
 val yyin_stream;
 
-long lineno = 1;
+cnum lineno = 1;
 int opt_loglevel = 1;   /* 0 - quiet; 1 - normal; 2 - verbose */
 int opt_nobindings = 0;
 int opt_arraydims = 1;
@@ -190,7 +191,7 @@ UONLY   {U2}{U}|{U3}{U}{U}|{U4}{U}{U}{U}
 %%
 
 <SPECIAL,NESTED>{TOK}   {
-                          long val;
+                          cnum val;
                           char *errp;
 
 
@@ -217,6 +218,9 @@ UONLY   {U2}{U}|{U3}{U}{U}|{U4}{U}{U}{U}
                               && errno == ERANGE)
                             yyerror("numeric overflow in token");
 
+                          if (val < NUM_MIN || val > NUM_MAX)
+                            yyerror("numeric overflow in token");
+
                           yylval.num = val;
                           return NUMBER;
                         }
diff --git a/parser.y b/parser.y
index c69c1b33..51f4e02e 100644
--- a/parser.y
+++ b/parser.y
@@ -30,6 +30,7 @@
 #include <assert.h>
 #include <limits.h>
 #include <dirent.h>
+#include "config.h"
 #include "lib.h"
 #include "regex.h"
 #include "utf8.h"
@@ -50,7 +51,7 @@ static val parsed_spec;
   wchar_t *lexeme;
   union obj *obj;
   wchar_t chr;
-  long num;
+  cnum num;
 }
 
 %token <lexeme> TEXT IDENT KEYWORD ALL SOME NONE MAYBE CASES AND OR END COLLECT
diff --git a/regex.c b/regex.c
index 19bbd785..d315c15f 100644
--- a/regex.c
+++ b/regex.c
@@ -32,6 +32,7 @@
 #include <dirent.h>
 #include <setjmp.h>
 #include <dirent.h>
+#include "config.h"
 #include "lib.h"
 #include "unwind.h"
 #include "regex.h"
@@ -912,7 +913,7 @@ int nfa_move(nfa_state_t **in, int nin, nfa_state_t **out, wchar_t ch)
  * determines the match length (defaulting to zero
  * if no acceptance states were encountered).
  */
-long nfa_run(nfa_t nfa, const wchar_t *str)
+cnum nfa_run(nfa_t nfa, const wchar_t *str)
 {
   const wchar_t *last_accept_pos = 0, *ptr = str;
   unsigned visited = nfa.start->a.visited + 1;
@@ -953,7 +954,7 @@ long nfa_run(nfa_t nfa, const wchar_t *str)
   return last_accept_pos ? last_accept_pos - str : -1;
 }
 
-long nfa_machine_match_span(nfa_machine_t *nfam)
+cnum nfa_machine_match_span(nfa_machine_t *nfam)
 {
   return nfam->last_accept_pos;
 }
@@ -1080,12 +1081,12 @@ val search_regex(val haystack, val needle_regex, val start,
     return nil;
   } else {
     if (from_end) {
-      long i;
-      long s = c_num(start);
+      cnum i;
+      cnum s = c_num(start);
       const wchar_t *h = c_str(haystack);
 
       for (i = c_num(length_str(haystack)) - 1; i >= s; i--) {
-        long span = nfa_run(*pnfa, h + i);
+        cnum span = nfa_run(*pnfa, h + i);
         if (span >= 0)
           return cons(num(i), num(span));
       }
diff --git a/regex.h b/regex.h
index accb0386..d33f5c84 100644
--- a/regex.h
+++ b/regex.h
@@ -145,22 +145,22 @@ typedef enum nfam_result {
 } nfam_result_t;
 
 typedef struct nfa_machine {
-  long last_accept_pos;
+  cnum last_accept_pos;
   unsigned visited;
   nfa_state_t **move, **clos, **stack;
   int nmove, nclos;
-  long count;
+  cnum count;
   nfa_t nfa;
 } nfa_machine_t;
 
 nfa_t nfa_compile_regex(val regex);
 void nfa_free(nfa_t);
-long nfa_run(nfa_t nfa, const wchar_t *str);
+cnum nfa_run(nfa_t nfa, const wchar_t *str);
 void nfa_machine_reset(nfa_machine_t *);
 void nfa_machine_init(nfa_machine_t *, nfa_t);
 void nfa_machine_cleanup(nfa_machine_t *);
 nfam_result_t nfa_machine_feed(nfa_machine_t *, wchar_t ch);
-long nfa_machine_match_span(nfa_machine_t *);
+cnum nfa_machine_match_span(nfa_machine_t *);
 val regex_compile(val regex_sexp);
 val regexp(val);
 nfa_t *regex_nfa(val);
diff --git a/stream.c b/stream.c
index b57a57b1..b76d037d 100644
--- a/stream.c
+++ b/stream.c
@@ -35,6 +35,7 @@
 #include <wchar.h>
 #include <unistd.h>
 #include <sys/wait.h>
+#include "config.h"
 #include "lib.h"
 #include "gc.h"
 #include "unwind.h"
@@ -612,6 +613,44 @@ val get_byte(val stream)
   }
 }
 
+struct fmt {
+  size_t minsize;
+  const char *dec;
+  const char *oct;
+  const char *hex;
+  const char *HEX;
+};
+
+static struct fmt fmt_tab[] = {
+  { sizeof(short),"%hd",   "%ho",   "%hx",   "%hX"   },
+  { sizeof(int),  "%d",    "%o",    "%x",    "%X"    },
+  { sizeof(long), "%ld",   "%lo",   "%lx",   "%llX"  },
+  { sizeof(cnum), "%lld",  "%llo",  "%llx",  "%llX"  },
+  { sizeof(cnum), "%Ld",   "%Lo",   "%Lx",   "%llX"  },
+  { sizeof(cnum), "%qd",   "%qo",   "%qx",   "%qX",  },
+  { sizeof(cnum), "%I64d", "%I64o", "%I64x", "%I64X" }, 
+  { 0,            0,       0,       0,       0       }
+};
+
+static struct fmt *num_fmt;
+
+static void detect_format_string(void)
+{
+  struct fmt *f;
+  char buf[64];
+  cnum num = 1234;
+
+  for (f = fmt_tab; f->minsize != 0; f++) {
+    memset(buf, 0, sizeof buf);
+    if (f->minsize != sizeof num)
+      continue;
+    if (sprintf(buf, f->dec, num) == 4 && strcmp(buf, "1234") == 0) {
+      num_fmt = f;
+      break;
+    }
+  }
+}
+
 static val vformat_num(val stream, const char *str,
                        int width, int left, int pad, int precision)
 {
@@ -681,7 +720,7 @@ val vformat(val stream, val fmtstr, va_list vl)
     } state = vf_init, saved_state = vf_init;
     int width = 0, precision = 0, digits = 0;
     int left = 0, zeropad = 0;
-    long value;
+    cnum value;
     void *ptr;
     char num_buf[64];
 
@@ -798,17 +837,17 @@ val vformat(val stream, val fmtstr, va_list vl)
         case 'x':
           obj = va_arg(vl, val);
           value = c_num(obj);
-          sprintf(num_buf, "%lx", value);
+          sprintf(num_buf, num_fmt->hex, value);
           goto output_num;
         case 'X':
           obj = va_arg(vl, val);
           value = c_num(obj);
-          sprintf(num_buf, "%lX", value);
+          sprintf(num_buf, num_fmt->HEX, value);
           goto output_num;
         case 'o':
           obj = va_arg(vl, val);
           value = c_num(obj);
-          sprintf(num_buf, "%lo", value);
+          sprintf(num_buf, num_fmt->oct, value);
           goto output_num;
         case 'a':
           obj = va_arg(vl, val);
@@ -816,7 +855,7 @@ val vformat(val stream, val fmtstr, va_list vl)
             goto premature;
           if (nump(obj)) {
             value = c_num(obj);
-            sprintf(num_buf, "%ld", value);
+            sprintf(num_buf, num_fmt->dec, value);
             goto output_num;
           } else if (stringp(obj)) {
             if (!vformat_str(stream, obj, width, left, precision))
@@ -831,7 +870,7 @@ val vformat(val stream, val fmtstr, va_list vl)
             goto premature;
           if (nump(obj)) {
             value = c_num(obj);
-            sprintf(num_buf, "%ld", value);
+            sprintf(num_buf, num_fmt->dec, value);
             if (!vformat_num(stream, num_buf, 0, 0, 0, 0))
               return nil;
             continue;
@@ -841,7 +880,8 @@ val vformat(val stream, val fmtstr, va_list vl)
         case 'p':
           ptr = va_arg(vl, void *);
           value = (int) ptr;
-          sprintf(num_buf, "0x%lx", value);
+          strcpy(num_buf, "0x");
+          sprintf(num_buf + 2, num_fmt->hex, value);
           goto output_num;
         default:
           abort();
@@ -920,4 +960,5 @@ void stream_init(void)
   std_input = make_stdio_stream(stdin, string(L"stdin"), t, nil);
   std_output = make_stdio_stream(stdout, string(L"stdout"), nil, t);
   std_error = make_stdio_stream(stderr, string(L"stderr"), nil, t);
+  detect_format_string();
 }
diff --git a/txr.c b/txr.c
index b6b49b9c..efa03b27 100644
--- a/txr.c
+++ b/txr.c
@@ -33,6 +33,7 @@
 #include <setjmp.h>
 #include <stdarg.h>
 #include <wchar.h>
+#include "config.h"
 #include "lib.h"
 #include "stream.h"
 #include "gc.h"
diff --git a/unwind.c b/unwind.c
index 1365c7c2..1c56ba4d 100644
--- a/unwind.c
+++ b/unwind.c
@@ -30,6 +30,7 @@
 #include <setjmp.h>
 #include <dirent.h>
 #include <stdarg.h>
+#include "config.h"
 #include "lib.h"
 #include "gc.h"
 #include "stream.h"
diff --git a/unwind.h b/unwind.h
index 546c42ac..d04da53c 100644
--- a/unwind.h
+++ b/unwind.h
@@ -148,7 +148,7 @@ noreturn val type_mismatch(val, ...);
 
 #define internal_error(STR)             \
   do {                                  \
-    extern obj_t *num(long);            \
+    extern obj_t *num(cnum);            \
     uw_throwf(internal_err,             \
               lit("~a:~a ~a"),          \
               lit(__FILE__),            \
diff --git a/utf8.c b/utf8.c
index 56f533a2..eedd503d 100644
--- a/utf8.c
+++ b/utf8.c
@@ -28,6 +28,7 @@
 #include <stdio.h>
 #include <stdlib.h>
 #include <wchar.h>
+#include "config.h"
 #include "lib.h"
 #include "utf8.h"