Module Name: othersrc
Committed By: dyoung
Date: Sat Aug 22 05:08:48 UTC 2015
Modified Files:
othersrc/external/bsd/arfe: README
othersrc/external/bsd/arfe/dt: Makefile README dt.c testit
othersrc/external/bsd/arfe/dt/rr: ifconfig.0-ifconfig.1
ifconfig.0-logger-ifconfig.0 ifconfig.1-ifconfig.0
logger-ifconfig.0-ifconfig.0 wm0-a-wm0-b wm0-b-wm0-a
Log Message:
Locate in both inputs hexadecimal numbers starting 0x and make them
"wild" in the alignments dt computes. In dt, bitwise-AND the 0x-hex
numbers. In it, bitwise-OR them. Take care not to match a hexadecimal
with a decimal or vice versa!
TBD: identify hexadecimals that don't start 0x.
Remove a little dead code.
Split HB_DEBUG into HB_DEBUG and HB_ASSERT. The latter just enables the
assertions.
Update old test results for the new treatment of 0x-hexadecimal. Add
some new tests.
To generate a diff of this commit:
cvs rdiff -u -r1.1 -r1.2 othersrc/external/bsd/arfe/README
cvs rdiff -u -r1.1 -r1.2 othersrc/external/bsd/arfe/dt/Makefile \
othersrc/external/bsd/arfe/dt/README othersrc/external/bsd/arfe/dt/testit
cvs rdiff -u -r1.2 -r1.3 othersrc/external/bsd/arfe/dt/dt.c
cvs rdiff -u -r1.1 -r1.2 \
othersrc/external/bsd/arfe/dt/rr/ifconfig.0-ifconfig.1 \
othersrc/external/bsd/arfe/dt/rr/ifconfig.0-logger-ifconfig.0 \
othersrc/external/bsd/arfe/dt/rr/ifconfig.1-ifconfig.0 \
othersrc/external/bsd/arfe/dt/rr/logger-ifconfig.0-ifconfig.0 \
othersrc/external/bsd/arfe/dt/rr/wm0-a-wm0-b \
othersrc/external/bsd/arfe/dt/rr/wm0-b-wm0-a
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
Modified files:
Index: othersrc/external/bsd/arfe/README
diff -u othersrc/external/bsd/arfe/README:1.1 othersrc/external/bsd/arfe/README:1.2
--- othersrc/external/bsd/arfe/README:1.1 Mon Aug 10 21:10:59 2015
+++ othersrc/external/bsd/arfe/README Sat Aug 22 05:08:48 2015
@@ -1,3 +1,5 @@
+$ARFE: README 216 2015-08-22 05:04:28Z dyoung $
+
ARFE is a suite of tools for processing record- and field-oriented
digital texts. ARFE strives to make a useful set of automatic
text-processing functions available at a level of abstraction that both
Index: othersrc/external/bsd/arfe/dt/Makefile
diff -u othersrc/external/bsd/arfe/dt/Makefile:1.1 othersrc/external/bsd/arfe/dt/Makefile:1.2
--- othersrc/external/bsd/arfe/dt/Makefile:1.1 Mon Aug 10 21:10:59 2015
+++ othersrc/external/bsd/arfe/dt/Makefile Sat Aug 22 05:08:48 2015
@@ -1,9 +1,11 @@
+# $ARFE: Makefile 216 2015-08-22 05:04:28Z dyoung $
NOMAN=
.include <bsd.own.mk>
PROG=dt
SRCS+=dt.c
#CPPFLAGS+=-DHB_DEBUG
+CPPFLAGS+=-DHB_ASSERT
DBG+=-g -O3
#DBG+=-g -O0
#COPTS+=-pg
Index: othersrc/external/bsd/arfe/dt/README
diff -u othersrc/external/bsd/arfe/dt/README:1.1 othersrc/external/bsd/arfe/dt/README:1.2
--- othersrc/external/bsd/arfe/dt/README:1.1 Mon Aug 10 21:10:59 2015
+++ othersrc/external/bsd/arfe/dt/README Sat Aug 22 05:08:48 2015
@@ -1,3 +1,5 @@
+$ARFE: README 216 2015-08-22 05:04:28Z dyoung $
+
DT---(d)ifferentiate (t)ext---finds a longest common subsequence (LCS)
of two texts where decimal numbers are "wild": a span of decimal digits
in the first text will match a digits span in the second text. Then it
Index: othersrc/external/bsd/arfe/dt/testit
diff -u othersrc/external/bsd/arfe/dt/testit:1.1 othersrc/external/bsd/arfe/dt/testit:1.2
--- othersrc/external/bsd/arfe/dt/testit:1.1 Mon Aug 10 21:10:59 2015
+++ othersrc/external/bsd/arfe/dt/testit Sat Aug 22 05:08:48 2015
@@ -1,4 +1,5 @@
#!/bin/sh
+# $ARFE: testit 216 2015-08-22 05:04:28Z dyoung $
results=rr
@@ -34,4 +35,6 @@ quackabequack quackdariaquack
quackdariaquack quackabequack
wm0-a wm0-b
wm0-b wm0-a
+if-re0 if-wm0
+if-wm0 if-re0
EOF
Index: othersrc/external/bsd/arfe/dt/dt.c
diff -u othersrc/external/bsd/arfe/dt/dt.c:1.2 othersrc/external/bsd/arfe/dt/dt.c:1.3
--- othersrc/external/bsd/arfe/dt/dt.c:1.2 Tue Aug 11 21:04:59 2015
+++ othersrc/external/bsd/arfe/dt/dt.c Sat Aug 22 05:08:48 2015
@@ -1,5 +1,5 @@
-/* $NetBSD: dt.c,v 1.2 2015/08/11 21:04:59 dyoung Exp $ */
-/* $Id: dt.c,v 1.2 2015/08/11 21:04:59 dyoung Exp $ */
+/* $NetBSD: dt.c,v 1.3 2015/08/22 05:08:48 dyoung Exp $ */
+/* $ARFE: dt.c 215 2015-08-22 05:02:46Z dyoung $ */
/*-
* Copyright (c) 2014,2015 David Young <[email protected]>
@@ -32,6 +32,7 @@
#include <sys/mman.h>
#include <sys/stat.h>
#include <assert.h>
+#include <ctype.h> /* for ispunct(), isspace(), isxdigit() */
#include <err.h>
#include <errno.h>
#include <fcntl.h>
@@ -44,7 +45,7 @@
#include <limits.h>
#include <string.h>
-#if defined(HB_DEBUG)
+#if defined(HB_ASSERT)
#define dbg_assert assert
#else
#define dbg_assert(__x) do { } while (false) /* do nothing */
@@ -55,6 +56,90 @@ typedef struct origin {
int j;
} origin_t;
+typedef enum hex_op {
+ HEX_OP_NONE = 0
+ , HEX_OP_PUSH
+ , HEX_OP_EMIT
+ , HEX_OP_START
+} hex_op_t;
+
+struct hex_state;
+typedef struct hex_state hex_state_t;
+
+typedef struct hex_transition {
+ const hex_state_t *t_newstate;
+ hex_op_t t_op;
+} hex_transition_t;
+
+struct hex_state {
+ const char *s_descr;
+ hex_transition_t s_other;
+ hex_transition_t s_hexdigit;
+ hex_transition_t s_0;
+ hex_transition_t s_x;
+ hex_transition_t s_delimiter;
+ hex_transition_t s_eof;
+};
+
+enum hex_state_idxs {
+ READ_0
+ , READ_X
+ , READ_HEXDIGIT
+ , READ_EOF
+ , READ_OTHER
+ , READ_DELIMITER
+};
+
+#define read_0 hex_states[READ_0]
+#define read_x hex_states[READ_X]
+#define read_hexdigit hex_states[READ_HEXDIGIT]
+#define read_eof hex_states[READ_EOF]
+#define read_other hex_states[READ_OTHER]
+#define read_delimiter hex_states[READ_DELIMITER]
+
+static const hex_state_t hex_states[] = {[READ_0] = {
+ .s_descr = "read 0",
+ .s_other = {&read_other, HEX_OP_NONE},
+ .s_x = {&read_x, HEX_OP_NONE},
+}, [READ_X] = {
+ .s_descr = "read x",
+ .s_other = {&read_other, HEX_OP_NONE},
+ .s_hexdigit = {&read_hexdigit, HEX_OP_PUSH},
+}, [READ_HEXDIGIT] = {
+ .s_descr = "read hexdigit",
+ .s_other = {&read_other, HEX_OP_NONE},
+ .s_delimiter = {&read_delimiter, HEX_OP_EMIT},
+ .s_eof = {&read_eof, HEX_OP_EMIT},
+ .s_hexdigit = {&read_hexdigit, HEX_OP_PUSH},
+}, [READ_EOF] = {
+ .s_descr = "read eof",
+ .s_other = {&read_eof, HEX_OP_NONE},
+}, [READ_OTHER] = {
+ .s_descr = "read other",
+ .s_other = {&read_other, HEX_OP_NONE},
+ .s_delimiter = {&read_delimiter, HEX_OP_NONE},
+}, [READ_DELIMITER] = {
+ .s_descr = "read delimiter",
+ .s_0 = {&read_0, HEX_OP_START},
+ .s_other = {&read_other, HEX_OP_NONE},
+}};
+
+struct hex_parser;
+typedef struct hex_parser hex_parser_t;
+
+typedef void (*hex_emitter_t)(hex_parser_t *, void *);
+
+struct hex_parser {
+ struct {
+ int start, stop;
+ } p_column, p_idx;
+ char p_digits[16 + 1]; /* XXX how many hex digits in an intmax_t? */
+ char *p_top;
+ const hex_state_t *p_state;
+ hex_emitter_t p_emitter;
+ void *p_arg;
+};
+
typedef struct chainelt {
TAILQ_ENTRY(chainelt) ce_link;
int ce_column; /* 0: no column alignment,
@@ -74,17 +159,21 @@ typedef struct slice {
bool backward;
} slice_t;
+typedef enum {
+ KIND_NONE = 0
+ , KIND_INTMAX_DEC
+ , KIND_INTMAX_HEX
+ , KIND_STRING
+} clocc_kind_t;
+
typedef struct clocc {
int column;
int first;
int last;
- enum {
- KIND_NONE = 0
- , KIND_INTMAX
- , KIND_STRING
- } kind;
+ clocc_kind_t kind;
union {
intmax_t u_intmax;
+ uintmax_t u_uintmax;
} val_u;
} clocc_t;
@@ -101,7 +190,8 @@ typedef struct scratch {
cloccs_t Aoccs, Boccs;
slice_t *Abasis, *Bbasis;
size_t expected_lcs;
- char op;
+ char dec_op;
+ char hex_op;
} scratch_t;
union freeslice;
@@ -124,8 +214,6 @@ static bool freeslices_initialized __ali
static freeslice_t *first_freeslice __aligned(COHERENCY_UNIT);
static freeslice_t freeslices[2048];
-static int mainloop __aligned(COHERENCY_UNIT) = 0;
-
static int dbg_printf(const char *, ...) __printflike(1, 2);
static int dbg2_printf(const char *, ...) __printflike(1, 2);
@@ -166,6 +254,99 @@ dbg2_printf(const char *fmt, ...)
#endif
static void
+hex_parser_init(hex_parser_t *p, hex_emitter_t emitter, void *arg)
+{
+ memset(p, 0, sizeof(*p));
+ p->p_state = &read_delimiter;
+ p->p_emitter = emitter;
+ p->p_arg = arg;
+}
+
+static const char *
+hex_op_string(hex_op_t op)
+{
+ switch (op) {
+ case HEX_OP_NONE:
+ return "none";
+ case HEX_OP_EMIT:
+ return "emit";
+ case HEX_OP_PUSH:
+ return "push";
+ case HEX_OP_START:
+ return "start";
+ default:
+ return "unknown";
+ }
+}
+
+static void
+hex_parser_drive(hex_parser_t *p, int idx, int column, int c)
+{
+ const hex_transition_t *t;
+ const hex_state_t *ns, *s;
+
+ s = p->p_state;
+
+ if (c == -1 && (ns = s->s_eof.t_newstate) != NULL)
+ t = &s->s_eof;
+ else if (c == '0' && (ns = s->s_0.t_newstate) != NULL)
+ t = &s->s_0;
+ else if (c == 'x' && (ns = s->s_x.t_newstate) != NULL)
+ t = &s->s_x;
+ else if ((isspace(c) || ispunct(c)) &&
+ (ns = s->s_delimiter.t_newstate) != NULL)
+ t = &s->s_delimiter;
+ else if (isxdigit(c) && (ns = s->s_hexdigit.t_newstate) != NULL)
+ t = &s->s_hexdigit;
+ else if ((ns = s->s_other.t_newstate) != NULL)
+ t = &s->s_other;
+ else
+ ns = NULL;
+
+ assert(ns != NULL);
+ if (c == -1) {
+ dbg_printf("%s: transition (%s, eof) -> %s (op %s)\n", __func__,
+ s->s_descr, ns->s_descr, hex_op_string(t->t_op));
+ } else {
+ dbg_printf("%s: transition (%s, %c) -> %s (op %s)\n", __func__,
+ s->s_descr, c, ns->s_descr, hex_op_string(t->t_op));
+ }
+
+ switch (t->t_op) {
+ case HEX_OP_PUSH:
+ if (p->p_top == NULL)
+ ;
+ else if (p->p_top - &p->p_digits[0] ==
+ __arraycount(p->p_digits) - 1)
+ p->p_top = NULL;
+ else {
+ *p->p_top = c;
+ p->p_top++;
+ }
+ p->p_idx.stop = idx;
+ p->p_column.stop = column;
+ break;
+ case HEX_OP_START:
+ p->p_column.start = column;
+ p->p_column.stop = column + 1;
+ p->p_idx.start = idx;
+ p->p_idx.stop = idx + 1;
+ p->p_top = &p->p_digits[0];
+ break;
+ case HEX_OP_EMIT:
+ if (p->p_top == NULL)
+ dbg_printf("hex string too long\n");
+ *p->p_top = '\0';
+ (*p->p_emitter)(p, p->p_arg);
+ break;
+ case HEX_OP_NONE:
+ break;
+ }
+
+ p->p_state = ns;
+}
+
+static void
init_freeslices(void)
{
unsigned int i;
@@ -323,7 +504,7 @@ subslice(const slice_t *s, unsigned int
}
static bool
-clocc_starts_at(const cloccs_t *o, int first, size_t *wlenp)
+clocc_starts_at(const cloccs_t *o, int first, size_t *wlenp, clocc_kind_t *kindp)
{
const clocc_t *occ;
int l, m, r;
@@ -344,13 +525,15 @@ clocc_starts_at(const cloccs_t *o, int f
return false;
found:
+ if (kindp != NULL)
+ *kindp = occ->kind;
if (wlenp != NULL)
*wlenp = occ->last - occ->first + 1;
return true;
}
static bool
-clocc_ends_at(const cloccs_t *o, int last, size_t *wlenp)
+clocc_ends_at(const cloccs_t *o, int last, size_t *wlenp, clocc_kind_t *kindp)
{
const clocc_t *occ;
int l, m, r;
@@ -371,6 +554,8 @@ clocc_ends_at(const cloccs_t *o, int las
return false;
found:
+ if (kindp != NULL)
+ *kindp = occ->kind;
if (wlenp != NULL)
*wlenp = occ->last - occ->first + 1;
return true;
@@ -390,18 +575,18 @@ clocc_starts_in_slice_at(const cloccs_t
{
const int iprime = index_versus_basis(basis, s, i);
- return (s->backward) ? clocc_ends_at(o, iprime, NULL)
- : clocc_starts_at(o, iprime, NULL);
+ return (s->backward) ? clocc_ends_at(o, iprime, NULL, NULL)
+ : clocc_starts_at(o, iprime, NULL, NULL);
}
static bool
clocc_ends_in_slice_at(const cloccs_t *o, const slice_t *basis,
- const slice_t *s, int i, size_t *wlenp)
+ const slice_t *s, int i, size_t *wlenp, size_t *kindp)
{
const int iprime = index_versus_basis(basis, s, i);
- return (s->backward) ? clocc_starts_at(o, iprime, wlenp)
- : clocc_ends_at(o, iprime, wlenp);
+ return (s->backward) ? clocc_starts_at(o, iprime, wlenp, kindp)
+ : clocc_ends_at(o, iprime, wlenp, kindp);
}
static chain_t *
@@ -424,34 +609,66 @@ newchain(char *content, int column)
return c;
}
+/* Return 0 if the `a' and `b' are not left- or right-aligned on
+ * the same column, the column they right-align on, or zero minus
+ * the column they left-align on.
+ */
+static int
+cloccs_to_column(const clocc_t * const a, const clocc_t * const b)
+{
+ if (a->column + a->last - a->first ==
+ b->column + b->last - b->first)
+ return b->column + b->last - b->first;
+ else if (a->column == b->column)
+ return -a->column;
+ else
+ return 0;
+}
+
static chain_t *
-clocc(const scratch_t *scratch, int idx_a, int idx_b)
+hex_clocc(const scratch_t *scratch, int idx_a, int idx_b)
{
char *ret;
int column;
- intmax_t result;
+ uintmax_t result;
const clocc_t * const a = &scratch->Aoccs.occs[idx_a],
* const b = &scratch->Boccs.occs[idx_b];
- if (a->kind != b->kind)
- return newchain(strdup("<clocc>"), 0);
+ column = cloccs_to_column(a, b);
- if (b->kind != KIND_INTMAX)
- return newchain(strdup("<word>"), 0);
+ dbg2_printf("%#jx @ %d / %#jx @ %d -> column %d\n",
+ a->val_u.u_uintmax, a->column,
+ b->val_u.u_uintmax, b->column, column);
- if (a->column + a->last - a->first ==
- b->column + b->last - b->first)
- column = b->column + b->last - b->first;
- else if (a->column == b->column)
- column = -a->column;
+ if (scratch->hex_op == '&')
+ result = b->val_u.u_uintmax & a->val_u.u_uintmax;
else
- column = 0;
+ result = b->val_u.u_uintmax | a->val_u.u_uintmax;
+
+ if (result == 0)
+ return newchain(strdup("0x0"), column);
+ if (asprintf(&ret, "%#jx", result) != -1)
+ return newchain(ret, column);
+
+ return newchain(strdup("<hexmax>"), column);
+}
+
+static chain_t *
+int_clocc(const scratch_t *scratch, int idx_a, int idx_b)
+{
+ char *ret;
+ int column;
+ intmax_t result;
+ const clocc_t * const a = &scratch->Aoccs.occs[idx_a],
+ * const b = &scratch->Boccs.occs[idx_b];
+
+ column = cloccs_to_column(a, b);
dbg2_printf("%jd @ %d / %jd @ %d -> column %d\n",
a->val_u.u_intmax, a->column,
b->val_u.u_intmax, b->column, column);
- if (scratch->op == '-')
+ if (scratch->dec_op == '-')
result = b->val_u.u_intmax - a->val_u.u_intmax;
else
result = b->val_u.u_intmax + a->val_u.u_intmax;
@@ -463,6 +680,24 @@ clocc(const scratch_t *scratch, int idx_
}
static chain_t *
+clocc(const scratch_t *scratch, int idx_a, int idx_b)
+{
+ const clocc_t * const a = &scratch->Aoccs.occs[idx_a],
+ * const b = &scratch->Boccs.occs[idx_b];
+
+ if (a->kind != b->kind)
+ return newchain(strdup("<clocc>"), 0);
+
+ if (b->kind == KIND_INTMAX_HEX)
+ return hex_clocc(scratch, idx_a, idx_b);
+
+ if (b->kind == KIND_INTMAX_DEC)
+ return int_clocc(scratch, idx_a, idx_b);
+
+ return newchain(strdup("<word>"), 0);
+}
+
+static chain_t *
empty(void)
{
return newchain(strdup(""), 0);
@@ -477,6 +712,9 @@ algb(const slice_t *A, const slice_t *B,
unsigned int i, j;
size_t *K[2], *Kclocc, *Kcur, *Kprev, *Ktmp;
struct {
+ clocc_kind_t A, B;
+ } kind;
+ struct {
size_t A, B;
} wlen;
bool inside_clocc = false;
@@ -510,7 +748,7 @@ algb(const slice_t *A, const slice_t *B,
const bool clocc_ends_this_row = inside_clocc &&
clocc_ends_in_slice_at(&scratch->Aoccs, scratch->Abasis,
- A, i, &wlen.A);
+ A, i, &wlen.A, &kind.A);
for (j = 0; j < n; j++) {
@@ -522,7 +760,7 @@ algb(const slice_t *A, const slice_t *B,
if (clocc_ends_this_row &&
clocc_ends_in_slice_at(&scratch->Boccs,
scratch->Bbasis, B, j,
- &wlen.B)) {
+ &wlen.B, &kind.B) && kind.A == kind.B) {
const size_t step = wlen.A * wlen.B + 1,
nscore = Kclocc[j - wlen.B] + step;
assert(wlen.B <= j + 1);
@@ -633,7 +871,7 @@ split_interval(cloccs_t *o, int origin,
*/
static int
widest_clocc_in_interval(cloccs_t *o, int origin, int first, int last,
- int *startp, int *endp)
+ int *startp, int *endp, clocc_kind_t kind)
{
int i, widest = -1, maxw = 0;
@@ -645,6 +883,9 @@ widest_clocc_in_interval(cloccs_t *o, in
if (o->occs[i].last > origin + last)
break;
+ if (o->occs[i].kind != kind)
+ continue;
+
w = o->occs[i].last - o->occs[i].first + 1;
if (w > maxw) {
maxw = w;
@@ -657,14 +898,17 @@ widest_clocc_in_interval(cloccs_t *o, in
}
static int
-interval_to_clocc_index(cloccs_t *o, int origin, int first, int last)
+interval_to_clocc_index(cloccs_t *o, int origin, int first, int last, clocc_kind_t *kindp)
{
int i;
for (i = 0; i < o->noccs; i++) {
if (origin + first == o->occs[i].first &&
- origin + last == o->occs[i].last)
+ origin + last == o->occs[i].last) {
+ if (kindp != NULL)
+ *kindp = o->occs[i].kind;
return i;
+ }
}
return -1;
}
@@ -740,20 +984,22 @@ algc(const slice_t *A, const slice_t *B,
return algc_return(m, n, origin, lcsp, 0, empty());
}
- ivlidx.A = interval_to_clocc_index(&scratch->Aoccs, origin.i, 0, m - 1);
+ clocc_kind_t kind;
+
+ ivlidx.A = interval_to_clocc_index(&scratch->Aoccs, origin.i, 0, m - 1, &kind);
/* If A is a clocc occurrence, must try to match A with a clocc in B! */
if (ivlidx.A != -1) {
int start = 0, end = 0;
ivlidx.B = widest_clocc_in_interval(&scratch->Boccs, origin.j,
- 0, n - 1, &start, &end);
+ 0, n - 1, &start, &end, kind);
if (ivlidx.B != -1) {
dbg_assert(expected_lcs == no_expected_lcs ||
- (size_t)(end - start + 2) == expected_lcs);
+ (size_t)(end - start + 1) * m + 1 == expected_lcs);
return algc_return(m, n, origin, lcsp,
- (size_t)(end - start + 2),
+ (size_t)(end - start + 1) * m + 1,
clocc(scratch, ivlidx.A, ivlidx.B));
}
}
@@ -904,11 +1150,75 @@ slicestr(slice_t *s, const char *t)
}
static void
+emit_hex_clocc(hex_parser_t *p, void *arg)
+{
+ cloccs_t *o = arg;
+ intmax_t val;
+ char *end;
+
+ dbg_printf("found hex string %s\n", p->p_digits);
+
+ if (o->noccs >= (int)__arraycount(o->occs))
+ return;
+
+ val = strtoimax(p->p_digits, &end, 16);
+
+ if (*end != '\0' && errno == ERANGE)
+ warnx("%s: over/underflow at %d", __func__, p->p_column.start);
+
+#if 1
+ o->occs[o->noccs].column = p->p_column.start;
+ o->occs[o->noccs].first = p->p_idx.start;
+ o->occs[o->noccs].last = p->p_idx.stop;
+ o->occs[o->noccs].kind = KIND_INTMAX_HEX;
+ o->occs[o->noccs].val_u.u_uintmax = val;
+ o->noccs++;
+#endif
+}
+
+static int
+clocc_cmp(const void *l0, const void *r0)
+{
+ const clocc_t *l = l0, *r = r0;
+
+ return l->first - r->first;
+}
+
+static int
+clocc_width(const clocc_t *c)
+{
+ assert(c->last >= c->first);
+ return c->last - c->first + 1;
+}
+
+static void
+cloccs_dedup(cloccs_t *o)
+{
+ int i, j;
+
+ for (i = 1; i < o->noccs;) {
+ clocc_t *l, *r;
+ l = &o->occs[i - 1];
+ r = &o->occs[i];
+ if (l->last < r->first) {
+ i++;
+ continue;
+ }
+ if (clocc_width(r) > clocc_width(l))
+ *l = *r;
+ for (j = i + 1; j < o->noccs; j++)
+ o->occs[j - 1] = o->occs[j];
+ o->noccs--;
+ }
+}
+
+static void
cloccs_init(cloccs_t *o, slice_t *s)
{
+ hex_parser_t hex_parser;
size_t column, i, n;
unsigned int j, k;
- char digits[64 / 3 + 1];
+ char digits[64 / 3 + 1]; /* XXX how many digits in an intmax_t? */
o->noccs = 0;
for (i = 0; i < __arraycount(names); i++) {
@@ -926,9 +1236,23 @@ cloccs_init(cloccs_t *o, slice_t *s)
o->noccs++;
}
- j = 0;
+ hex_parser_init(&hex_parser, emit_hex_clocc, o);
+
column = 0;
n = length(s);
+ for (j = 0; j < n; j++) {
+ char c;
+
+ c = get(s, j);
+ ++column;
+ hex_parser_drive(&hex_parser, j, column, c);
+ if (c == '\n')
+ column = 0;
+ }
+ hex_parser_drive(&hex_parser, n, column + 1, -1);
+
+ j = 0;
+ column = 0;
while (o->noccs < (int)__arraycount(o->occs)) {
intmax_t val;
char *end;
@@ -966,7 +1290,7 @@ cloccs_init(cloccs_t *o, slice_t *s)
o->occs[o->noccs].column = column;
o->occs[o->noccs].first = j;
o->occs[o->noccs].last = k - 1;
- o->occs[o->noccs].kind = KIND_INTMAX;
+ o->occs[o->noccs].kind = KIND_INTMAX_DEC;
o->occs[o->noccs].val_u.u_intmax = val;
dbg_printf("found %jd at %d - %d\n",
val, o->occs[o->noccs].first,
@@ -978,6 +1302,8 @@ next:
column += k - 1 - j;
j = k;
}
+ mergesort(o->occs, o->noccs, sizeof(*o->occs), clocc_cmp);
+ cloccs_dedup(o);
}
static void
@@ -1026,11 +1352,13 @@ main(int argc, char **argv)
B = file_to_slice(argv[2]);
n = length(B);
scratch_init(&scratch, n, A, B);
- if (strcmp(getprogname(), "dt") == 0)
- scratch.op = '-';
- else if (strcmp(getprogname(), "it") == 0)
- scratch.op = '+';
- else
+ if (strcmp(getprogname(), "dt") == 0) {
+ scratch.dec_op = '-';
+ scratch.hex_op = '&';
+ } else if (strcmp(getprogname(), "it") == 0) {
+ scratch.dec_op = '+';
+ scratch.hex_op = '|';
+ } else
errx(EXIT_FAILURE, "not implemented");
cloccs_init(&scratch.Aoccs, A);
@@ -1042,7 +1370,5 @@ main(int argc, char **argv)
printchain(C);
dbg_printf("\n");
- dbg_printf("\n%d times through main loop\n", mainloop);
-
return EXIT_SUCCESS;
}
Index: othersrc/external/bsd/arfe/dt/rr/ifconfig.0-ifconfig.1
diff -u othersrc/external/bsd/arfe/dt/rr/ifconfig.0-ifconfig.1:1.1 othersrc/external/bsd/arfe/dt/rr/ifconfig.0-ifconfig.1:1.2
--- othersrc/external/bsd/arfe/dt/rr/ifconfig.0-ifconfig.1:1.1 Mon Aug 10 21:10:59 2015
+++ othersrc/external/bsd/arfe/dt/rr/ifconfig.0-ifconfig.1 Sat Aug 22 05:08:48 2015
@@ -6,5 +6,5 @@ wm0: flags= 0<UP,BROADCAST,RUNNING,SIM
status: active
input: 61 packets, 5432 bytes, 3 multicasts
output: 30 packets, 3436 bytes, 0 multicasts
- inet 0.0.0. 0 netmask 0xffffff 0 broadcast 0.0.0. 0
- inet0 fe 0:: 0a:bff:fecd:0ef%wm0 prefixlen 0 scopeid 0x0
+ inet 0.0.0. 0 netmask 0xffffff00 broadcast 0.0.0. 0
+ inet0 fe 0:: 0a:bff:fecd:0ef%wm0 prefixlen 0 scopeid 0x1
Index: othersrc/external/bsd/arfe/dt/rr/ifconfig.0-logger-ifconfig.0
diff -u othersrc/external/bsd/arfe/dt/rr/ifconfig.0-logger-ifconfig.0:1.1 othersrc/external/bsd/arfe/dt/rr/ifconfig.0-logger-ifconfig.0:1.2
--- othersrc/external/bsd/arfe/dt/rr/ifconfig.0-logger-ifconfig.0:1.1 Mon Aug 10 21:10:59 2015
+++ othersrc/external/bsd/arfe/dt/rr/ifconfig.0-logger-ifconfig.0 Sat Aug 22 05:08:48 2015
@@ -6,5 +6,5 @@ media: Ethernet autoselect (0baseT full-
status: active
input: 0 packets, 0 bytes, 0 multicasts
output: 0 packets, 0 bytes, 0 multicasts
-inet 0.0.0.0 netmask 0xffffff0 broadcast 0.0.0.0
-inet0 fe0::0a:bff:fecd:0ef%wm0 prefixlen 0 scopeid 0x0
+inet 0.0.0.0 netmask 0xffffff00 broadcast 0.0.0.0
+inet0 fe0::0a:bff:fecd:0ef%wm0 prefixlen 0 scopeid 0x1
Index: othersrc/external/bsd/arfe/dt/rr/ifconfig.1-ifconfig.0
diff -u othersrc/external/bsd/arfe/dt/rr/ifconfig.1-ifconfig.0:1.1 othersrc/external/bsd/arfe/dt/rr/ifconfig.1-ifconfig.0:1.2
--- othersrc/external/bsd/arfe/dt/rr/ifconfig.1-ifconfig.0:1.1 Mon Aug 10 21:10:59 2015
+++ othersrc/external/bsd/arfe/dt/rr/ifconfig.1-ifconfig.0 Sat Aug 22 05:08:48 2015
@@ -6,5 +6,5 @@ wm0: flags= 0<UP,BROADCAST,RUNNING,SIM
status: active
input: -61 packets, -5432 bytes, -3 multicasts
output: -30 packets, -3436 bytes, 0 multicasts
- inet 0.0.0. 0 netmask 0xffffff 0 broadcast 0.0.0. 0
- inet0 fe 0:: 0a:bff:fecd:0ef%wm0 prefixlen 0 scopeid 0x0
+ inet 0.0.0. 0 netmask 0xffffff00 broadcast 0.0.0. 0
+ inet0 fe 0:: 0a:bff:fecd:0ef%wm0 prefixlen 0 scopeid 0x1
Index: othersrc/external/bsd/arfe/dt/rr/logger-ifconfig.0-ifconfig.0
diff -u othersrc/external/bsd/arfe/dt/rr/logger-ifconfig.0-ifconfig.0:1.1 othersrc/external/bsd/arfe/dt/rr/logger-ifconfig.0-ifconfig.0:1.2
--- othersrc/external/bsd/arfe/dt/rr/logger-ifconfig.0-ifconfig.0:1.1 Mon Aug 10 21:10:59 2015
+++ othersrc/external/bsd/arfe/dt/rr/logger-ifconfig.0-ifconfig.0 Sat Aug 22 05:08:48 2015
@@ -6,5 +6,5 @@ media: Ethernet autoselect (0baseT full-
status: active
input: 0 packets, 0 bytes, 0 multicasts
output: 0 packets, 0 bytes, 0 multicasts
-inet 0.0.0.0 netmask 0xffffff0 broadcast 0.0.0.0
-inet0 fe0::0a:bff:fecd:0ef%wm0 prefixlen 0 scopeid 0x0
+inet 0.0.0.0 netmask 0xffffff00 broadcast 0.0.0.0
+inet0 fe0::0a:bff:fecd:0ef%wm0 prefixlen 0 scopeid 0x1
Index: othersrc/external/bsd/arfe/dt/rr/wm0-a-wm0-b
diff -u othersrc/external/bsd/arfe/dt/rr/wm0-a-wm0-b:1.1 othersrc/external/bsd/arfe/dt/rr/wm0-a-wm0-b:1.2
--- othersrc/external/bsd/arfe/dt/rr/wm0-a-wm0-b:1.1 Mon Aug 10 21:10:59 2015
+++ othersrc/external/bsd/arfe/dt/rr/wm0-a-wm0-b Sat Aug 22 05:08:48 2015
@@ -6,7 +6,7 @@ wm0: flags= 0<UP,BROADCAST,RUNNING,SIM
status: active
input: 112 packets, 11380 bytes, 11 multicasts
output: 60 packets, 11440 bytes, 0 multicasts
- inet 0.0.0. 0 netmask 0xffffff 0 broadcast 0.0.0. 0
- inet0 fe 0:: 0a:bff:fecd:0ef%wm0 prefixlen 0 scopeid 0x0
+ inet 0.0.0. 0 netmask 0xffffff00 broadcast 0.0.0. 0
+ inet0 fe 0:: 0a:bff:fecd:0ef%wm0 prefixlen 0 scopeid 0x1
Index: othersrc/external/bsd/arfe/dt/rr/wm0-b-wm0-a
diff -u othersrc/external/bsd/arfe/dt/rr/wm0-b-wm0-a:1.1 othersrc/external/bsd/arfe/dt/rr/wm0-b-wm0-a:1.2
--- othersrc/external/bsd/arfe/dt/rr/wm0-b-wm0-a:1.1 Mon Aug 10 21:10:59 2015
+++ othersrc/external/bsd/arfe/dt/rr/wm0-b-wm0-a Sat Aug 22 05:08:48 2015
@@ -6,7 +6,7 @@ wm0: flags= 0<UP,BROADCAST,RUNNING,SIM
status: active
input: -112 packets, -11380 bytes, -11 multicasts
output: -60 packets, -11440 bytes, 0 multicasts
- inet 0.0.0. 0 netmask 0xffffff 0 broadcast 0.0.0. 0
- inet0 fe 0:: 0a:bff:fecd:0ef%wm0 prefixlen 0 scopeid 0x0
+ inet 0.0.0. 0 netmask 0xffffff00 broadcast 0.0.0. 0
+ inet0 fe 0:: 0a:bff:fecd:0ef%wm0 prefixlen 0 scopeid 0x1
