mozilla
/
pjs
зеркало из https://github.com/mozilla/pjs.git
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность
pjs/tools/trace-malloc/bloatblame.c

759 строки
28 KiB
C
Исходник Ответственный История

/* -*- Mode: C; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 4 -*-
*
* ***** BEGIN LICENSE BLOCK *****
* Version: MPL 1.1/GPL 2.0/LGPL 2.1
*
* The contents of this file are subject to the Mozilla Public License Version
* 1.1 (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
* http://www.mozilla.org/MPL/
*
* Software distributed under the License is distributed on an "AS IS" basis,
* WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
* for the specific language governing rights and limitations under the
* License.
*
* The Original Code is nsTraceMalloc.c/bloatblame.c code, released
* April 19, 2000.
*
* The Initial Developer of the Original Code is
* Netscape Communications Corporation.
* Portions created by the Initial Developer are Copyright (C) 2000
* the Initial Developer. All Rights Reserved.
*
* Contributor(s):
* Brendan Eich, 14-April-2000
*
* Alternatively, the contents of this file may be used under the terms of
* either the GNU General Public License Version 2 or later (the "GPL"), or
* the GNU Lesser General Public License Version 2.1 or later (the "LGPL"),
* in which case the provisions of the GPL or the LGPL are applicable instead
* of those above. If you wish to allow use of your version of this file only
* under the terms of either the GPL or the LGPL, and not to allow others to
* use your version of this file under the terms of the MPL, indicate your
* decision by deleting the provisions above and replace them with the notice
* and other provisions required by the GPL or the LGPL. If you do not delete
* the provisions above, a recipient may use your version of this file under
* the terms of any one of the MPL, the GPL or the LGPL.
*
* ***** END LICENSE BLOCK ***** */
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <ctype.h>
#include <errno.h>
#ifdef HAVE_GETOPT_H
#include <getopt.h>
#else
#ifdef XP_WIN
#include "getopt.c"
#else
extern int getopt(int argc, char *const *argv, const char *shortopts);
extern char *optarg;
extern int optind;
#endif
#endif
#include <math.h>
#include <time.h>
#include <sys/stat.h>
#include "prtypes.h"
#include "prlog.h"
#include "prprf.h"
#include "plhash.h"
#include "pldhash.h"
#include "nsTraceMalloc.h"
#include "tmreader.h"
static char *program;
static int sort_by_direct = 0;
static int js_mode = 0;
static int do_tree_dump = 0;
static int unified_output = 0;
static char *function_dump = NULL;
static uint32 min_subtotal = 0;
static void compute_callsite_totals(tmcallsite *site)
{
tmcallsite *kid;
site->allocs.bytes.total += site->allocs.bytes.direct;
site->allocs.calls.total += site->allocs.calls.direct;
for (kid = site->kids; kid; kid = kid->siblings) {
compute_callsite_totals(kid);
site->allocs.bytes.total += kid->allocs.bytes.total;
site->allocs.calls.total += kid->allocs.calls.total;
}
}
static void walk_callsite_tree(tmcallsite *site, int level, int kidnum, FILE *fp)
{
tmcallsite *parent;
tmgraphnode *comp, *pcomp, *lib, *plib;
tmmethodnode *meth, *pmeth;
int old_meth_low, old_comp_low, old_lib_low, nkids;
tmcallsite *kid;
parent = site->parent;
comp = lib = NULL;
meth = NULL;
if (parent) {
meth = site->method;
if (meth) {
pmeth = parent->method;
if (pmeth && pmeth != meth) {
if (!meth->graphnode.low) {
meth->graphnode.allocs.bytes.total += site->allocs.bytes.total;
meth->graphnode.allocs.calls.total += site->allocs.calls.total;
}
if (!tmgraphnode_connect(&(pmeth->graphnode), &(meth->graphnode), site))
goto bad;
comp = meth->graphnode.up;
if (comp) {
pcomp = pmeth->graphnode.up;
if (pcomp && pcomp != comp) {
if (!comp->low) {
comp->allocs.bytes.total
+= site->allocs.bytes.total;
comp->allocs.calls.total
+= site->allocs.calls.total;
}
if (!tmgraphnode_connect(pcomp, comp, site))
goto bad;
lib = comp->up;
if (lib) {
plib = pcomp->up;
if (plib && plib != lib) {
if (!lib->low) {
lib->allocs.bytes.total
+= site->allocs.bytes.total;
lib->allocs.calls.total
+= site->allocs.calls.total;
}
if (!tmgraphnode_connect(plib, lib, site))
goto bad;
}
old_lib_low = lib->low;
if (!old_lib_low)
lib->low = level;
}
}
old_comp_low = comp->low;
if (!old_comp_low)
comp->low = level;
}
}
old_meth_low = meth->graphnode.low;
if (!old_meth_low)
meth->graphnode.low = level;
}
}
if (do_tree_dump) {
fprintf(fp, "%c%*s%3d %3d %s %lu %ld\n",
site->kids ? '+' : '-', level, "", level, kidnum,
meth ? tmmethodnode_name(meth) : "???",
(unsigned long)site->allocs.bytes.direct,
(long)site->allocs.bytes.total);
}
nkids = 0;
level++;
for (kid = site->kids; kid; kid = kid->siblings) {
walk_callsite_tree(kid, level, nkids, fp);
nkids++;
}
if (meth) {
if (!old_meth_low)
meth->graphnode.low = 0;
if (comp) {
if (!old_comp_low)
comp->low = 0;
if (lib) {
if (!old_lib_low)
lib->low = 0;
}
}
}
return;
bad:
perror(program);
exit(1);
}
/*
* Linked list bubble-sort (waterson and brendan went bald hacking this).
*
* Sort the list in non-increasing order, using the expression passed as the
* 'lessthan' formal macro parameter. This expression should use 'curr' as
* the pointer to the current node (of type nodetype) and 'next' as the next
* node pointer. It should return true if curr is less than next, and false
* otherwise.
*/
#define BUBBLE_SORT_LINKED_LIST(listp, nodetype, lessthan) \
PR_BEGIN_MACRO \
nodetype *curr, **currp, *next, **nextp, *tmp; \
\
currp = listp; \
while ((curr = *currp) != NULL && curr->next) { \
nextp = &curr->next; \
while ((next = *nextp) != NULL) { \
if (lessthan) { \
tmp = curr->next; \
*currp = tmp; \
if (tmp == next) { \
PR_ASSERT(nextp == &curr->next); \
curr->next = next->next; \
next->next = curr; \
} else { \
*nextp = next->next; \
curr->next = next->next; \
next->next = tmp; \
*currp = next; \
*nextp = curr; \
nextp = &curr->next; \
} \
curr = next; \
continue; \
} \
nextp = &next->next; \
} \
currp = &curr->next; \
} \
PR_END_MACRO
static PRIntn tabulate_node(PLHashEntry *he, PRIntn i, void *arg)
{
tmgraphnode *node = (tmgraphnode*) he;
tmgraphnode **table = (tmgraphnode**) arg;
table[i] = node;
BUBBLE_SORT_LINKED_LIST(&node->down, tmgraphnode,
(curr->allocs.bytes.total < next->allocs.bytes.total));
return HT_ENUMERATE_NEXT;
}
/* Sort in reverse size order, so biggest node comes first. */
static int node_table_compare(const void *p1, const void *p2)
{
const tmgraphnode *node1, *node2;
uint32 key1, key2;
node1 = *(const tmgraphnode**) p1;
node2 = *(const tmgraphnode**) p2;
if (sort_by_direct) {
key1 = node1->allocs.bytes.direct;
key2 = node2->allocs.bytes.direct;
} else {
key1 = node1->allocs.bytes.total;
key2 = node2->allocs.bytes.total;
}
return (key2 < key1) ? -1 : (key2 > key1) ? 1 : 0;
}
static int mean_size_compare(const void *p1, const void *p2)
{
const tmgraphnode *node1, *node2;
double div1, div2, key1, key2;
node1 = *(const tmgraphnode**) p1;
node2 = *(const tmgraphnode**) p2;
div1 = (double)node1->allocs.calls.direct;
div2 = (double)node2->allocs.calls.direct;
if (div1 == 0 || div2 == 0)
return (int)(div2 - div1);
key1 = (double)node1->allocs.bytes.direct / div1;
key2 = (double)node2->allocs.bytes.direct / div2;
if (key1 < key2)
return 1;
if (key1 > key2)
return -1;
return 0;
}
static const char *prettybig(uint32 num, char *buf, size_t limit)
{
if (num >= 1000000000)
PR_snprintf(buf, limit, "%1.2fG", (double) num / 1e9);
else if (num >= 1000000)
PR_snprintf(buf, limit, "%1.2fM", (double) num / 1e6);
else if (num >= 1000)
PR_snprintf(buf, limit, "%1.2fK", (double) num / 1e3);
else
PR_snprintf(buf, limit, "%lu", (unsigned long) num);
return buf;
}
static double percent(uint32 num, uint32 total)
{
if (num == 0)
return 0.0;
return ((double) num * 100) / (double) total;
}
static void sort_graphlink_list(tmgraphlink **listp, int which)
{
BUBBLE_SORT_LINKED_LIST(listp, tmgraphlink,
(TM_LINK_TO_EDGE(curr, which)->allocs.bytes.total
< TM_LINK_TO_EDGE(next, which)->allocs.bytes.total));
}
static void dump_graphlink_list(tmgraphlink *list, int which, const char *name,
FILE *fp)
{
tmcounts bytes;
tmgraphlink *link;
tmgraphedge *edge;
char buf[16];
bytes.direct = bytes.total = 0;
for (link = list; link; link = link->next) {
edge = TM_LINK_TO_EDGE(link, which);
bytes.direct += edge->allocs.bytes.direct;
bytes.total += edge->allocs.bytes.total;
}
if (js_mode) {
fprintf(fp,
" %s:{dbytes:%ld, tbytes:%ld, edges:[\n",
name, (long) bytes.direct, (long) bytes.total);
for (link = list; link; link = link->next) {
edge = TM_LINK_TO_EDGE(link, which);
fprintf(fp,
" {node:%d, dbytes:%ld, tbytes:%ld},\n",
link->node->sort,
(long) edge->allocs.bytes.direct,
(long) edge->allocs.bytes.total);
}
fputs(" ]},\n", fp);
} else {
fputs("<td valign=top>", fp);
for (link = list; link; link = link->next) {
edge = TM_LINK_TO_EDGE(link, which);
fprintf(fp,
"<a href='#%s'>%s&nbsp;(%1.2f%%)</a>\n",
tmgraphnode_name(link->node),
prettybig(edge->allocs.bytes.total, buf, sizeof buf),
percent(edge->allocs.bytes.total, bytes.total));
}
fputs("</td>", fp);
}
}
static void dump_graph(tmreader *tmr, PLHashTable *hashtbl, const char *varname,
const char *title, FILE *fp)
{
uint32 i, count;
tmgraphnode **table, *node;
char *name;
size_t namelen;
char buf1[16], buf2[16], buf3[16], buf4[16];
static char NA[] = "N/A";
count = hashtbl->nentries;
table = (tmgraphnode**) malloc(count * sizeof(tmgraphnode*));
if (!table) {
perror(program);
exit(1);
}
PL_HashTableEnumerateEntries(hashtbl, tabulate_node, table);
qsort(table, count, sizeof(tmgraphnode*), node_table_compare);
for (i = 0; i < count; i++)
table[i]->sort = i;
if (js_mode) {
fprintf(fp,
"var %s = {\n name:'%s', title:'%s', nodes:[\n",
varname, varname, title);
} else {
fprintf(fp,
"<table border=1>\n"
"<tr>"
"<th>%s</th>"
"<th>Down</th>"
"<th>Next</th>"
"<th>Total/Direct (percents)</th>"
"<th>Allocations</th>"
"<th>Fan-in</th>"
"<th>Fan-out</th>"
"</tr>\n",
title);
}
for (i = 0; i < count; i++) {
/* Don't bother with truly puny nodes. */
node = table[i];
if (node->allocs.bytes.total < min_subtotal)
break;
name = tmgraphnode_name(node);
if (js_mode) {
fprintf(fp,
" {name:'%s', dbytes:%ld, tbytes:%ld,"
" dallocs:%ld, tallocs:%ld,\n",
name,
(long) node->allocs.bytes.direct,
(long) node->allocs.bytes.total,
(long) node->allocs.calls.direct,
(long) node->allocs.calls.total);
} else {
namelen = strlen(name);
fprintf(fp,
"<tr>"
"<td valign=top><a name='%s'>%.*s%s</a></td>",
name,
(namelen > 40) ? 40 : (int)namelen, name,
(namelen > 40) ? "<i>...</i>" : "");
if (node->down) {
fprintf(fp,
"<td valign=top><a href='#%s'><i>down</i></a></td>",
tmgraphnode_name(node->down));
} else {
fputs("<td></td>", fp);
}
if (node->next) {
fprintf(fp,
"<td valign=top><a href='#%s'><i>next</i></a></td>",
tmgraphnode_name(node->next));
} else {
fputs("<td></td>", fp);
}
fprintf(fp,
"<td valign=top>%s/%s (%1.2f%%/%1.2f%%)</td>"
"<td valign=top>%s/%s (%1.2f%%/%1.2f%%)</td>",
prettybig(node->allocs.bytes.total, buf1, sizeof buf1),
prettybig(node->allocs.bytes.direct, buf2, sizeof buf2),
percent(node->allocs.bytes.total,
tmr->calltree_root.allocs.bytes.total),
percent(node->allocs.bytes.direct,
tmr->calltree_root.allocs.bytes.total),
prettybig(node->allocs.calls.total, buf3, sizeof buf3),
prettybig(node->allocs.calls.direct, buf4, sizeof buf4),
percent(node->allocs.calls.total,
tmr->calltree_root.allocs.calls.total),
percent(node->allocs.calls.direct,
tmr->calltree_root.allocs.calls.total));
}
/* NB: we must use 'fin' because 'in' is a JS keyword! */
sort_graphlink_list(&node->in, TM_EDGE_IN_LINK);
dump_graphlink_list(node->in, TM_EDGE_IN_LINK, "fin", fp);
sort_graphlink_list(&node->out, TM_EDGE_OUT_LINK);
dump_graphlink_list(node->out, TM_EDGE_OUT_LINK, "out", fp);
if (js_mode)
fputs(" },\n", fp);
else
fputs("</tr>\n", fp);
}
if (js_mode) {
fputs("]};\n", fp);
} else {
fputs("</table>\n<hr>\n", fp);
qsort(table, count, sizeof(tmgraphnode*), mean_size_compare);
fprintf(fp,
"<table border=1>\n"
"<tr><th colspan=4>Direct Allocators</th></tr>\n"
"<tr>"
"<th>%s</th>"
"<th>Mean&nbsp;Size</th>"
"<th>StdDev</th>"
"<th>Allocations<th>"
"</tr>\n",
title);
for (i = 0; i < count; i++) {
double allocs, bytes, mean, variance, sigma;
node = table[i];
allocs = (double)node->allocs.calls.direct;
if (!allocs)
continue;
/* Compute direct-size mean and standard deviation. */
bytes = (double)node->allocs.bytes.direct;
mean = bytes / allocs;
variance = allocs * node->sqsum - bytes * bytes;
if (variance < 0 || allocs == 1)
variance = 0;
else
variance /= allocs * (allocs - 1);
sigma = sqrt(variance);
name = tmgraphnode_name(node);
namelen = strlen(name);
fprintf(fp,
"<tr>"
"<td valign=top>%.*s%s</td>"
"<td valign=top>%s</td>"
"<td valign=top>%s</td>"
"<td valign=top>%s</td>"
"</tr>\n",
(namelen > 65) ? 45 : (int)namelen, name,
(namelen > 65) ? "<i>...</i>" : "",
prettybig((uint32)mean, buf1, sizeof buf1),
prettybig((uint32)sigma, buf2, sizeof buf2),
prettybig(node->allocs.calls.direct, buf3, sizeof buf3));
}
fputs("</table>\n", fp);
}
free((void*) table);
}
static void my_tmevent_handler(tmreader *tmr, tmevent *event)
{
switch (event->type) {
case TM_EVENT_STATS:
if (js_mode)
break;
fprintf(stdout,
"<p><table border=1>"
"<tr><th>Counter</th><th>Value</th></tr>\n"
"<tr><td>maximum actual stack depth</td><td align=right>%lu</td></tr>\n"
"<tr><td>maximum callsite tree depth</td><td align=right>%lu</td></tr>\n"
"<tr><td>number of parent callsites</td><td align=right>%lu</td></tr>\n"
"<tr><td>maximum kids per parent</td><td align=right>%lu</td></tr>\n"
"<tr><td>hits looking for a kid</td><td align=right>%lu</td></tr>\n"
"<tr><td>misses looking for a kid</td><td align=right>%lu</td></tr>\n"
"<tr><td>steps over other kids</td><td align=right>%lu</td></tr>\n"
"<tr><td>callsite recurrences</td><td align=right>%lu</td></tr>\n"
"<tr><td>number of stack backtraces</td><td align=right>%lu</td></tr>\n"
"<tr><td>backtrace failures</td><td align=right>%lu</td></tr>\n"
"<tr><td>backtrace malloc failures</td><td align=right>%lu</td></tr>\n"
"<tr><td>backtrace dladdr failures</td><td align=right>%lu</td></tr>\n"
"<tr><td>malloc calls</td><td align=right>%lu</td></tr>\n"
"<tr><td>malloc failures</td><td align=right>%lu</td></tr>\n"
"<tr><td>calloc calls</td><td align=right>%lu</td></tr>\n"
"<tr><td>calloc failures</td><td align=right>%lu</td></tr>\n"
"<tr><td>realloc calls</td><td align=right>%lu</td></tr>\n"
"<tr><td>realloc failures</td><td align=right>%lu</td></tr>\n"
"<tr><td>free calls</td><td align=right>%lu</td></tr>\n"
"<tr><td>free(null) calls</td><td align=right>%lu</td></tr>\n"
"</table>",
(unsigned long) event->u.stats.tmstats.calltree_maxstack,
(unsigned long) event->u.stats.tmstats.calltree_maxdepth,
(unsigned long) event->u.stats.tmstats.calltree_parents,
(unsigned long) event->u.stats.tmstats.calltree_maxkids,
(unsigned long) event->u.stats.tmstats.calltree_kidhits,
(unsigned long) event->u.stats.tmstats.calltree_kidmisses,
(unsigned long) event->u.stats.tmstats.calltree_kidsteps,
(unsigned long) event->u.stats.tmstats.callsite_recurrences,
(unsigned long) event->u.stats.tmstats.backtrace_calls,
(unsigned long) event->u.stats.tmstats.backtrace_failures,
(unsigned long) event->u.stats.tmstats.btmalloc_failures,
(unsigned long) event->u.stats.tmstats.dladdr_failures,
(unsigned long) event->u.stats.tmstats.malloc_calls,
(unsigned long) event->u.stats.tmstats.malloc_failures,
(unsigned long) event->u.stats.tmstats.calloc_calls,
(unsigned long) event->u.stats.tmstats.calloc_failures,
(unsigned long) event->u.stats.tmstats.realloc_calls,
(unsigned long) event->u.stats.tmstats.realloc_failures,
(unsigned long) event->u.stats.tmstats.free_calls,
(unsigned long) event->u.stats.tmstats.null_free_calls);
if (event->u.stats.calltree_maxkids_parent) {
tmcallsite *site =
tmreader_callsite(tmr, event->u.stats.calltree_maxkids_parent);
if (site && site->method) {
fprintf(stdout, "<p>callsite with the most kids: %s</p>",
tmmethodnode_name(site->method));
}
}
if (event->u.stats.calltree_maxstack_top) {
tmcallsite *site =
tmreader_callsite(tmr, event->u.stats.calltree_maxstack_top);
fputs("<p>deepest callsite tree path:\n"
"<table border=1>\n"
"<tr><th>Method</th><th>Offset</th></tr>\n",
stdout);
while (site) {
fprintf(stdout,
"<tr><td>%s</td><td>0x%08lX</td></tr>\n",
site->method ? tmmethodnode_name(site->method) : "???",
(unsigned long) site->offset);
site = site->parent;
}
fputs("</table>\n<hr>\n", stdout);
}
break;
}
}
int main(int argc, char **argv)
{
int c, i, j, rv;
tmreader *tmr;
FILE *fp;
program = *argv;
tmr = tmreader_new(program, NULL);
if (!tmr) {
perror(program);
exit(1);
}
while ((c = getopt(argc, argv, "djtuf:m:")) != EOF) {
switch (c) {
case 'd':
sort_by_direct = 1;
break;
case 'j':
js_mode = 1;
break;
case 't':
do_tree_dump = 1;
break;
case 'u':
unified_output = 1;
break;
case 'f':
function_dump = optarg;
break;
case 'm':
min_subtotal = atoi(optarg);
break;
default:
fprintf(stderr,
"usage: %s [-dtu] [-f function-dump-filename] [-m min] [output.html]\n",
program);
exit(2);
}
}
if (!js_mode) {
time_t start = time(NULL);
fprintf(stdout,
"<script language=\"JavaScript\">\n"
"function onload() {\n"
" document.links[0].__proto__.onmouseover = new Function("
"\"window.status ="
" this.href.substring(this.href.lastIndexOf('#') + 1)\");\n"
"}\n"
"</script>\n");
fprintf(stdout, "%s starting at %s", program, ctime(&start));
fflush(stdout);
}
argc -= optind;
argv += optind;
if (argc == 0) {
if (tmreader_eventloop(tmr, "-", my_tmevent_handler) <= 0)
exit(1);
} else {
for (i = j = 0; i < argc; i++) {
fp = fopen(argv[i], "r");
if (!fp) {
fprintf(stderr, "%s: can't open %s: %s\n",
program, argv[i], strerror(errno));
exit(1);
}
rv = tmreader_eventloop(tmr, argv[i], my_tmevent_handler);
if (rv < 0)
exit(1);
if (rv > 0)
j++;
fclose(fp);
}
if (j == 0)
exit(1);
}
compute_callsite_totals(&tmr->calltree_root);
walk_callsite_tree(&tmr->calltree_root, 0, 0, stdout);
if (js_mode) {
fprintf(stdout,
"<script language='javascript'>\n"
"// direct and total byte and allocator-call counts\n"
"var dbytes = %ld, tbytes = %ld,"
" dallocs = %ld, tallocs = %ld;\n",
(long) tmr->calltree_root.allocs.bytes.direct,
(long) tmr->calltree_root.allocs.bytes.total,
(long) tmr->calltree_root.allocs.calls.direct,
(long) tmr->calltree_root.allocs.calls.total);
}
dump_graph(tmr, tmr->libraries, "libraries", "Library", stdout);
if (!js_mode)
fputs("<hr>\n", stdout);
dump_graph(tmr, tmr->components, "classes", "Class or Component", stdout);
if (js_mode || unified_output || function_dump) {
if (js_mode || unified_output || strcmp(function_dump, "-") == 0) {
fp = stdout;
if (!js_mode)
fputs("<hr>\n", fp);
} else {
struct stat sb, fsb;
fstat(fileno(stdout), &sb);
if (stat(function_dump, &fsb) == 0 &&
fsb.st_dev == sb.st_dev && fsb.st_ino == sb.st_ino) {
fp = stdout;
fputs("<hr>\n", fp);
} else {
fp = fopen(function_dump, "w");
if (!fp) {
fprintf(stderr, "%s: can't open %s: %s\n",
program, function_dump, strerror(errno));
exit(1);
}
}
}
dump_graph(tmr, tmr->methods, "methods", "Function or Method", fp);
if (fp != stdout)
fclose(fp);
if (js_mode) {
fputs("function viewnode(graph, index) {\n"
" view.location = viewsrc();\n"
"}\n"
"function viewnodelink(graph, index) {\n"
" var node = graph.nodes[index];\n"
" return '<a href=\"javascript:viewnode('"
" + graph.name.quote() + ', ' + node.sort"
" + ')\" onmouseover=' + node.name.quote() + '>'"
" + node.name + '</a>';\n"
"}\n"
"function search(expr) {\n"
" var re = new RegExp(expr);\n"
" var src = '';\n"
" var graphs = [libraries, classes, methods]\n"
" var nodes;\n"
" for (var n = 0; n < (nodes = graphs[n].nodes).length; n++) {\n"
" for (var i = 0; i < nodes.length; i++) {\n"
" if (re.test(nodes[i].name))\n"
" src += viewnodelink(graph, i) + '\\n';\n"
" }\n"
" }\n"
" view.location = viewsrc();\n"
"}\n"
"function ctrlsrc() {\n"
" return \"<form>\\n"
"search: <input size=40 onchange='search(this.value)'>\\n"
"</form>\\n\";\n"
"}\n"
"function viewsrc() {\n"
" return 'hiiiii'\n"
"}\n"
"</script>\n"
"<frameset rows='10%,*'>\n"
" <frame name='ctrl' src='javascript:top.ctrlsrc()'>\n"
" <frame name='view' src='javascript:top.viewsrc()'>\n"
"</frameset>\n",
stdout);
}
}
exit(0);
}
Ссылка в новой задаче Показать git blame Копировать постоянную ссылку