[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[gawk-diffs] [SCM] gawk branch, master, updated. 783ae783275b5fab9104b71
|
From: |
Arnold Robbins |
|
Subject: |
[gawk-diffs] [SCM] gawk branch, master, updated. 783ae783275b5fab9104b711ac3d531dcc454d96 |
|
Date: |
Sun, 08 May 2011 17:53:09 +0000 |
This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "gawk".
The branch, master has been updated
via 783ae783275b5fab9104b711ac3d531dcc454d96 (commit)
via 7c20cc42ca1b77a2f3caddab543839efe897eeb0 (commit)
from e2b1395c84d6073b064dac58ccb28e1784577e75 (commit)
Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.
- Log -----------------------------------------------------------------
http://git.sv.gnu.org/cgit/gawk.git/commit/?id=783ae783275b5fab9104b711ac3d531dcc454d96
commit 783ae783275b5fab9104b711ac3d531dcc454d96
Author: Arnold D. Robbins <address@hidden>
Date: Sun May 8 20:52:33 2011 +0300
More array sorting fixes.
diff --git a/ChangeLog b/ChangeLog
index fde2cc6..fcc5f71 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,20 @@
+Sun May 8 20:38:03 2011 John Haque <address@hidden>
+
+ * eval.c (r_interpret): In case Op_sub_array, store only the
+ subarray index as 'vname'.
+ * array.c (make_aname): Redone for dynamic computation of
+ a subarray actual 'vname'.
+ (array_vname): Use make_aname() for (sub)array name.
+ (asort_actual): Performance optimization for asort(a).
+
+Sun May 8 20:29:17 2011 Arnold D. Robbins <address@hidden>
+
+ * debug.c (print_array): Sort in order of string indices, per
+ request from John Haque.
+ * array.c (sort_up_value_number): Use string value to provide
+ ordering when numeric values are equal. Ensures that tests come
+ out OK on different systems.
+
Sun May 8 20:27:27 2011 Arnold D. Robbins <address@hidden>
* dfa.c: Sync with GNU grep.
diff --git a/array.c b/array.c
index b6e72c8..5cf75d2 100644
--- a/array.c
+++ b/array.c
@@ -85,6 +85,41 @@ array_init()
hash = gst_hash_string;
}
+/* make_aname --- construct a 'vname' for a (sub)array */
+
+static char *
+make_aname(const NODE *symbol)
+{
+ static char *aname = NULL;
+ static size_t alen;
+ static size_t max_alen;
+#define SLEN 256
+
+ if (symbol->parent_array != NULL) {
+ size_t slen;
+
+ (void) make_aname(symbol->parent_array);
+ slen = strlen(symbol->vname); /* subscript in parent array */
+ if (alen + slen + 4 > max_alen) { /*
sizeof("[\"\"]") = 4 */
+ max_alen = alen + slen + 4 + SLEN;
+ erealloc(aname, char *, (max_alen + 1) * sizeof(char
*), "make_aname");
+ }
+ alen += sprintf(aname + alen, "[\"%s\"]", symbol->vname);
+ } else {
+ alen = strlen(symbol->vname);
+ if (aname == NULL) {
+ max_alen = alen + SLEN;
+ emalloc(aname, char *, (max_alen + 1) * sizeof(char *),
"make_aname");
+ } else if (alen > max_alen) {
+ max_alen = alen + SLEN;
+ erealloc(aname, char *, (max_alen + 1) * sizeof(char
*), "make_aname");
+ }
+ memcpy(aname, symbol->vname, alen + 1);
+ }
+ return aname;
+#undef SLEN
+}
+
/*
* array_vname --- print the name of the array
*
@@ -103,18 +138,33 @@ array_vname(const NODE *symbol)
int n;
const NODE *save_symbol = symbol;
const char *from = _("from %s");
+ const char *aname;
- if (symbol->type != Node_array_ref || symbol->orig_array->type !=
Node_var_array)
- return symbol->vname;
+ if (symbol->type != Node_array_ref
+ || symbol->orig_array->type != Node_var_array
+ ) {
+ if (symbol->type != Node_var_array || symbol->parent_array ==
NULL)
+ return symbol->vname;
+ return make_aname(symbol);
+ }
/* First, we have to compute the length of the string: */
- len = strlen(symbol->vname) + 2; /* "%s (" */
+
+ len = 2; /* " (" */
n = 0;
- do {
- symbol = symbol->prev_array;
+ while (symbol->type == Node_array_ref) {
len += strlen(symbol->vname);
n++;
- } while (symbol->type == Node_array_ref);
+ symbol = symbol->prev_array;
+ }
+
+ /* Get the (sub)array name */
+ if (symbol->parent_array == NULL)
+ aname = symbol->vname;
+ else
+ aname = make_aname(symbol);
+ len += strlen(aname);
+
/*
* Each node contributes by strlen(from) minus the length
* of "%s" in the translation (which is at least 2)
@@ -139,42 +189,21 @@ array_vname(const NODE *symbol)
* Ancient systems have sprintf() returning char *, not int.
* If you have one of those, use sprintf(..); s += strlen(s) instead.
*/
+
s += sprintf(s, "%s (", symbol->vname);
for (;;) {
symbol = symbol->prev_array;
- sprintf(s, from, symbol->vname);
- s += strlen(s);
if (symbol->type != Node_array_ref)
break;
- sprintf(s, ", ");
- s += strlen(s);
+ s += sprintf(s, from, symbol->vname);
+ s += sprintf(s, ", ");
}
+ s += sprintf(s, from, aname);
strcpy(s, ")");
return message;
}
-/* make_aname --- construct a sub-array name for multi-dimensional array */
-
-char *
-make_aname(NODE *array, NODE *subs)
-{
- static char *aname = NULL;
- static size_t aname_len;
- size_t slen;
-
- slen = strlen(array->vname) + subs->stlen + 6;
- if (aname == NULL) {
- emalloc(aname, char *, slen, "make_aname");
- aname_len = slen;
- } else if (slen > aname_len) {
- erealloc(aname, char *, slen, "make_aname");
- aname_len = slen;
- }
- sprintf(aname, "%s[\"%.*s\"]", array->vname, (int) subs->stlen,
subs->stptr);
- return aname;
-}
-
/*
* get_array --- proceed to the actual Node_var_array,
@@ -702,7 +731,7 @@ do {
\
/* e.g.: a[1] = 1; delete a[1][1] */
free_subs(i);
fatal(_("attempt to use scalar `%s[\"%.*s\"]'
as an array"),
- symbol->vname,
+ array_vname(symbol),
(int) bucket->ahname_len,
bucket->ahname_str);
}
@@ -994,14 +1023,9 @@ dup_table(NODE *symbol, NODE *newsymb)
if (chain->ahvalue->type ==
Node_var_array) {
NODE *r;
- char *aname;
- size_t aname_len;
getnode(r);
r->type = Node_var_array;
- aname_len =
strlen(newsymb->vname) + chain->ahname_len + 4;
- emalloc(aname, char *,
aname_len + 2, "dup_table");
- sprintf(aname, "%s[\"%.*s\"]",
newsymb->vname, (int) chain->ahname_len, chain->ahname_str);
- r->vname = aname;
+ r->vname =
estrdup(chain->ahname_str, chain->ahname_len);
r->parent_array = newsymb;
bucket->ahvalue =
dup_table(chain->ahvalue, r);
} else
@@ -1136,23 +1160,34 @@ asort_actual(int nargs, SORT_CTXT ctxt)
/* We want the values of the source array. */
val = r->ahvalue;
- if (val->type == Node_val)
- *assoc_lookup(result, subs, FALSE) =
dupnode(val);
- else {
- const char *arr_name = make_aname(result, subs);
- NODE *arr;
-
- /*
- * There isn't any reference counting for
- * subarrays, so recursively copy subarrays
- * using dup_table().
- */
- getnode(arr);
- arr->type = Node_var_array;
- arr->var_array = NULL;
- arr->vname = estrdup(arr_name,
strlen(arr_name));
- arr->parent_array = array; /* actual parent,
not the temporary one. */
- *assoc_lookup(result, subs, FALSE) =
dup_table(val, arr);
+ if (result != dest) {
+ /* optimization for dest = NULL or dest = array
*/
+
+ if (val->type == Node_var_array) {
+ /* update subarray index in parent
array */
+ efree(val->vname);
+ val->vname = estrdup(subs->stptr,
subs->stlen);
+ }
+ *assoc_lookup(result, subs, FALSE) = val;
+ r->ahvalue = Nnull_string;
+ } else {
+ if (val->type == Node_val)
+ *assoc_lookup(result, subs, FALSE) =
dupnode(val);
+ else {
+ NODE *arr;
+
+ /*
+ * There isn't any reference counting
for
+ * subarrays, so recursively copy
subarrays
+ * using dup_table().
+ */
+ getnode(arr);
+ arr->type = Node_var_array;
+ arr->var_array = NULL;
+ arr->vname = estrdup(subs->stptr,
subs->stlen);
+ arr->parent_array = array; /* actual
parent, not the temporary one. */
+ *assoc_lookup(result, subs, FALSE) =
dup_table(val, arr);
+ }
}
}
@@ -1383,6 +1418,16 @@ sort_up_value_number(const void *p1, const void *p2)
else
ret = (n1->numbr > n2->numbr);
+ if (ret == 0) {
+ /*
+ * Use string value to guarantee same sort order on all
+ * versions of qsort().
+ */
+ n1 = force_string(n1);
+ n2 = force_string(n2);
+ ret = cmp_string(n1, n2);
+ }
+
return ret;
}
diff --git a/awk.h b/awk.h
index a14f20a..d05e07c 100644
--- a/awk.h
+++ b/awk.h
@@ -1113,7 +1113,6 @@ typedef enum sort_context { SORTED_IN = 1, ASORT, ASORTI
} SORT_CTXT;
extern NODE **assoc_list(NODE *array, const char *sort_str, SORT_CTXT
sort_ctxt);
extern NODE *get_array(NODE *symbol, int canfatal);
extern char *array_vname(const NODE *symbol);
-extern char *make_aname(NODE *array, NODE *subs);
extern void array_init(void);
extern void set_SUBSEP(void);
extern NODE *concat_exp(int nargs, int do_subsep);
diff --git a/debug.c b/debug.c
index 1543a46..9b9db34 100644
--- a/debug.c
+++ b/debug.c
@@ -1074,7 +1074,7 @@ print_array(volatile NODE *arr, char *arr_name)
num_elems = arr->table_size;
/* sort indices, sub_arrays are also sorted! */
- list = assoc_list((NODE *) arr, "@unsorted", SORTED_IN);
+ list = assoc_list((NODE *) arr, "@ind_str_asc", SORTED_IN);
PUSH_BINDING(pager_quit_tag_stack, pager_quit_tag,
pager_quit_tag_valid);
if (setjmp(pager_quit_tag) == 0) {
@@ -1262,12 +1262,12 @@ do_set_var(CMDARG *arg, int cmd ATTRIBUTE_UNUSED)
}
} else {
if (value == NULL) {
- const char *aname = make_aname(r, subs);
NODE *array;
+
getnode(array);
array->type = Node_var_array;
array->var_array = NULL;
- array->vname = estrdup(aname,
strlen(aname));
+ array->vname = estrdup(subs->stptr,
subs->stlen);
*assoc_lookup(r, subs, FALSE) = array;
r = array;
} else if (value->type != Node_var_array) {
diff --git a/eval.c b/eval.c
index 02ca13e..4132474 100644
--- a/eval.c
+++ b/eval.c
@@ -1763,18 +1763,15 @@ top:
t1 = POP_ARRAY();
r = in_array(t1, t2);
if (r == NULL) {
- const char *arr_name = make_aname(t1, t2);
getnode(r);
r->type = Node_var_array;
r->var_array = NULL;
- r->vname = estrdup(arr_name, strlen(arr_name));
+ r->vname = estrdup(t2->stptr, t2->stlen);
/* the subscript in parent array */
r->parent_array = t1;
*assoc_lookup(t1, t2, FALSE) = r;
- } else if (r->type != Node_var_array) {
- const char *arr_name = make_aname(t1, t2);
- DEREF(t2);
- fatal(_("attempt to use scalar `%s' as an
array"), arr_name);
- }
+ } else if (r->type != Node_var_array)
+ fatal(_("attempt to use scalar `%s[\"%.*s\"]'
as an array"),
+ array_vname(t1), (int)
t2->stlen, t2->stptr);
DEREF(t2);
PUSH(r);
break;
@@ -1783,11 +1780,9 @@ top:
t2 = mk_sub(pc->sub_count);
t1 = POP_ARRAY();
lhs = assoc_lookup(t1, t2, pc->do_reference);
- if ((*lhs)->type == Node_var_array) {
- const char *arr_name = make_aname(t1, t2);
- DEREF(t2);
- fatal(_("attempt to use array `%s' in a scalar
context"), arr_name);
- }
+ if ((*lhs)->type == Node_var_array)
+ fatal(_("attempt to use array `%s[\"%.*s\"]' in
a scalar context"),
+ array_vname(t1), (int)
t2->stlen, t2->stptr);
DEREF(t2);
PUSH_ADDRESS(lhs);
break;
@@ -2031,11 +2026,9 @@ post:
t1 = get_array(pc->memory, TRUE); /* array */
t2 = mk_sub(pc->expr_count); /* subscript */
lhs = assoc_lookup(t1, t2, FALSE);
- if ((*lhs)->type == Node_var_array) {
- const char *arr_name = make_aname(t1, t2);
- DEREF(t2);
- fatal(_("attempt to use array `%s' in a scalar
context"), arr_name);
- }
+ if ((*lhs)->type == Node_var_array)
+ fatal(_("attempt to use array `%s[\"%.*s\"]' in
a scalar context"),
+ array_vname(t1), (int)
t2->stlen, t2->stptr);
DEREF(t2);
unref(*lhs);
*lhs = POP_SCALAR();
diff --git a/test/ChangeLog b/test/ChangeLog
index 03c9359..c7a01cc 100644
--- a/test/ChangeLog
+++ b/test/ChangeLog
@@ -1,3 +1,7 @@
+Sun May 8 20:32:59 2011 Arnold D. Robbins <address@hidden>
+
+ * arraysort.ok, sort1.ok: Updated.
+
Wed May 4 23:37:27 2011 Arnold D. Robbins <address@hidden>
Revise tests for array sorting.
diff --git a/test/arraysort.ok b/test/arraysort.ok
index 454badf..d2d8ca7 100644
--- a/test/arraysort.ok
+++ b/test/arraysort.ok
@@ -71,13 +71,13 @@ y 1
|4|4|
|5|5|
--asc val num--
-|D3|D3|
|0|0|
+|D3|D3|
|1|1|
|2|2|
-|3D|3D|
| 3| 3|
|3|3|
-|4|4|
+|3D|3D|
| 4 | 4 |
+|4|4|
|5|5|
diff --git a/test/sort1.ok b/test/sort1.ok
index 2d43e31..00ed661 100644
--- a/test/sort1.ok
+++ b/test/sort1.ok
@@ -58,14 +58,14 @@
[6][3]: tttalb
[6][4]: zrab
---end PROCINFO["sorted_in"] = "@val_str_asc", IGNORECASE = 0---
+ [3]: Zebra
[1]: barz
[2]: blattt
- [3]: Zebra
[5]: 234
[4]: 1234
- [6][4]: zrab
[6][2]: arbeZ
[6][3]: tttalb
+ [6][4]: zrab
[6][1]: 4321
---end PROCINFO["sorted_in"] = "@val_num_asc", IGNORECASE = 0---
[6][4]: zrab
@@ -80,12 +80,12 @@
---end PROCINFO["sorted_in"] = "@val_str_desc", IGNORECASE = 0---
[6][1]: 4321
[6][4]: zrab
- [6][2]: arbeZ
[6][3]: tttalb
+ [6][2]: arbeZ
[4]: 1234
[5]: 234
- [1]: barz
[2]: blattt
+ [1]: barz
[3]: Zebra
---end PROCINFO["sorted_in"] = "@val_num_desc", IGNORECASE = 0---
[5]: 234
@@ -173,9 +173,9 @@
[3]: Zebra
[5]: 234
[4]: 1234
- [6][4]: zrab
[6][2]: arbeZ
[6][3]: tttalb
+ [6][4]: zrab
[6][1]: 4321
---end PROCINFO["sorted_in"] = "@val_num_asc", IGNORECASE = 1---
[6][4]: zrab
@@ -190,13 +190,13 @@
---end PROCINFO["sorted_in"] = "@val_str_desc", IGNORECASE = 1---
[6][1]: 4321
[6][4]: zrab
- [6][2]: arbeZ
[6][3]: tttalb
+ [6][2]: arbeZ
[4]: 1234
[5]: 234
- [1]: barz
- [2]: blattt
[3]: Zebra
+ [2]: blattt
+ [1]: barz
---end PROCINFO["sorted_in"] = "@val_num_desc", IGNORECASE = 1---
[5]: 234
[4]: 1234
http://git.sv.gnu.org/cgit/gawk.git/commit/?id=7c20cc42ca1b77a2f3caddab543839efe897eeb0
commit 7c20cc42ca1b77a2f3caddab543839efe897eeb0
Author: Arnold D. Robbins <address@hidden>
Date: Sun May 8 20:28:13 2011 +0300
Sync dfa.c and regex.h with their upstreams.
diff --git a/ChangeLog b/ChangeLog
index c1275db..fde2cc6 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,9 @@
+Sun May 8 20:27:27 2011 Arnold D. Robbins <address@hidden>
+
+ * dfa.c: Sync with GNU grep.
+ * regex.h: Sync with GLIBC in preparation for submitting updates
+ back.
+
Thu May 5 21:22:44 2011 Arnold D. Robbins <address@hidden>
* eval.c (r_interpret): PROCINFO sorting only takes effect
diff --git a/dfa.c b/dfa.c
index 439f964..650ceac 100644
--- a/dfa.c
+++ b/dfa.c
@@ -598,7 +598,8 @@ setbit_case_fold (
else
{
#if MBS_SUPPORT
- if (wctob ((unsigned char)b) == b)
+ int b2 = wctob ((unsigned char) b);
+ if (b2 == EOF || b2 == b)
#endif
setbit (b, c);
}
diff --git a/regex.h b/regex.h
index 86f875c..d54876a 100644
--- a/regex.h
+++ b/regex.h
@@ -155,10 +155,18 @@ typedef unsigned long int reg_syntax_t;
If not set, then the GNU regex operators are recognized. */
# define RE_NO_GNU_OPS (RE_NO_POSIX_BACKTRACKING << 1)
+/* If this bit is set, turn on internal regex debugging.
+ If not set, and debugging was on, turn it off.
+ This only works if regex.c is compiled -DDEBUG.
+ We define this bit always, so that all that's needed to turn on
+ debugging is to recompile regex.c; the calling code can always have
+ this bit set, and it won't affect anything in the normal case. */
+# define RE_DEBUG (RE_NO_GNU_OPS << 1)
+
/* If this bit is set, a syntactically invalid interval is treated as
a string of ordinary characters. For example, the ERE 'a{1' is
treated as 'a\{1'. */
-# define RE_INVALID_INTERVAL_ORD (RE_NO_GNU_OPS << 1)
+# define RE_INVALID_INTERVAL_ORD (RE_DEBUG << 1)
/* If this bit is set, then ignore case when matching.
If not set, then case is significant. */
@@ -175,7 +183,7 @@ typedef unsigned long int reg_syntax_t;
/* If this bit is set, then no_sub will be set to 1 during
re_compile_pattern. */
-#define RE_NO_SUB (RE_CONTEXT_INVALID_DUP << 1)
+# define RE_NO_SUB (RE_CONTEXT_INVALID_DUP << 1)
#endif
/* This global variable defines the particular regexp syntax to use (for
-----------------------------------------------------------------------
Summary of changes:
ChangeLog | 23 ++++++++
array.c | 155 ++++++++++++++++++++++++++++++++++-------------------
awk.h | 1 -
debug.c | 6 +-
dfa.c | 3 +-
eval.c | 27 ++++------
regex.h | 12 ++++-
test/ChangeLog | 4 ++
test/arraysort.ok | 6 +-
test/sort1.ok | 16 +++---
10 files changed, 163 insertions(+), 90 deletions(-)
hooks/post-receive
--
gawk
| [Prev in Thread] |
Current Thread |
[Next in Thread] |
- [gawk-diffs] [SCM] gawk branch, master, updated. 783ae783275b5fab9104b711ac3d531dcc454d96,
Arnold Robbins <=