so in the previous post i created a utility to get some exe’s and .pdb’s file GUID’s. However i need to get this information for the process that is currently executed (self). So what do we do? We apply they same technique as before and just preset the filename with the current executable:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
| #include "stdafx.h"
#include "exeguid.h"
void printGuid(GUID &g)
{
char buf[120];
sprintf(buf,"%08x-%04x-%04x-%02x%02x-%02x%02x%02x%02x%02x%02x", g.Data1,g.Data2,g.Data3,UINT(g.Data4[0]),UINT(g.Data4[1]),UINT(g.Data4[2]),UINT(g.Data4[3]),UINT(g.Data4[4]),UINT(g.Data4[5]),UINT(g.Data4[6]),UINT(g.Data4[7]));
printf("%s\n", buf);
}
int _tmain(int argc, _TCHAR* argv[])
{
GUID g;
if(!getmyGUID(g))
printGuid(g);
else
printf("error getting my GUID\n");
return 0;
} |
sounds easy, eh? 
so to proof its working:
>SelfGUID.exe
8b73f207-5509-4b7d-82c9-a4979208416f
executed the same binary again: GUID stays the same:
>SelfGUID.exe
8b73f207-5509-4b7d-82c9-a4979208416f
rebuild the binary two times and executed it after building:
>SelfGUID.exe
68ce44c6-5ca8-431d-a29b-5480cf7e53ac
>SelfGUID.exe
b91fae19-e200-4207-8468-a4ca54581949
EDIT: removed unused dependency from the project (zip updated)
Source code and executable: selfGUID.zip
So i was experimenting with windows and the debug functions and needed a fast utility that will give me the GUID’s for executables and .pdb files. This GUID is used to compare if the executable and the .pdb file fit together. (note: only in most recent [VS2008] versions, before a timestamp was used)
So with this little handy utility you can get the executable GUID:
>GUIDTool.exe exe RoR.exe
f8d102be-40e0-4409-8d94-c97e0bb4fce0
and its fitting .pdb file:
>GUIDTool.exe pdb ror.pdb
f8d102be-40e0-4409-8d94-c97e0bb4fce0
Please note that 99% of the code was copy-pasted together – origins are still in the source files.
Source code and executable: GUIDTool.zip
so i always wondered whats the fastest way is to iterate over a vector. So with this little snippet you can find out for yourself:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
| // some simple std::vector benchmark tool
// Jan 2010, thomas{AT}thomasfischer{DOT}biz
// also see http://stackoverflow.com/questions/776624/whats-faster-iterating-an-stl-vector-with-vectoriterator-or-with-at
#include <stdio.h>
#include <vector>
#include "Timer.h"
#define VECSIZE 500000
typedef struct foo_t {
char tmp[2];
} foo_t;
int main(int argc, char **argv)
{
int amount = VECSIZE;
int tests = 10;
printf("testing with %d elements a %d bytes (%0.2f MB) and %d runs per test\n", amount, sizeof(foo_t), (amount*sizeof(foo_t))/1024.0f/1024.0f, tests);
double time = 0.0f;
for(int i=0;i<tests;i++)
{
std::vector<foo_t> vec;
vec.reserve(amount);
foo_t f; // with random data in it
Timer *t = new Timer();
for(int i=0;i<amount;i++)
vec.push_back(f);
time += t->elapsed();
}
printf("add time (resized before): %f\n", time/tests);
time=0;
for(int i=0;i<tests;i++)
{
std::vector<foo_t> vec;
foo_t f; // with random data in it
Timer *t = new Timer();
for(int i=0;i<amount;i++)
vec.push_back(f);
time += t->elapsed();
}
printf("add time (dynamic resize): %f\n", time/tests);
time=0;
for(int i=0;i<tests;i++)
{
std::vector<foo_t> vec;
foo_t f; // with random data in it
for(int i=0;i<amount;i++)
vec.push_back(f);
Timer *t = new Timer();
for(std::vector<foo_t>::iterator it=vec.begin(); it!=vec.end(); it++)
{
// some example usage
it->tmp[1] = 0;
}
time += t->elapsed();
}
printf("iterate time (version #1): %f\n", time/tests);
time=0;
for(int i=0;i<tests;i++)
{
std::vector<foo_t> vec;
foo_t f; // with random data in it
for(int i=0;i<amount;i++)
vec.push_back(f);
Timer *t = new Timer();
std::vector<foo_t>::iterator vecEnd = vec.end();
for(std::vector<foo_t>::iterator it=vec.begin(); it!=vecEnd; ++it)
{
// some example usage
it->tmp[1] = 0;
}
time += t->elapsed();
}
printf("iterate time (version #2): %f\n", time/tests);
time=0;
for(int i=0;i<tests;i++)
{
std::vector<foo_t> vec;
foo_t f; // with random data in it
for(int i=0;i<amount;i++)
vec.push_back(f);
Timer *t = new Timer();
for(int i=0; i<amount; i++)
{
// some example usage
vec[i].tmp[1] = 0;
}
time += t->elapsed();
}
printf("iterate time (version #3): %f\n", time/tests);
time=0;
for(int i=0;i<tests;i++)
{
std::vector<foo_t> vec;
foo_t f; // with random data in it
for(int i=0;i<amount;i++)
vec.push_back(f);
Timer *t = new Timer();
for(unsigned int i=0; i<vec.size(); ++i)
{
// some example usage
vec.at(i).tmp[1] = 0;
}
time += t->elapsed();
}
printf("iterate time (version #4): %f\n", time/tests);
time=0;
for(int i=0;i<tests;i++)
{
foo_t vec[VECSIZE];
Timer *t = new Timer();
for(int i=0; i<amount; ++i)
{
// some example usage
vec[i].tmp[1] = 0;
}
time += t->elapsed();
}
printf("iterate over array (version #1): %f\n", time/tests);
time=0;
for(int i=0;i<tests;i++)
{
foo_t *vec = (foo_t *)malloc(sizeof(foo_t) * amount);
Timer *t = new Timer();
for(int i=0; i<amount; ++i)
{
// some example usage
vec[i].tmp[1] = 0;
}
time += t->elapsed();
free(vec);
}
printf("iterate over array (version #2): %f\n", time/tests);
return 0;
} |
to compile, add the Timer.h from my previous post
which results in the following output on my windows machine (/Oi /O2)
testing with 500000 elements a 2 bytes (0.95 MB) and 10 runs per test
add time (resized before): 0.005281
add time (dynamic resize): 0.007693
iterate time (version #1): 0.004180
iterate time (version #2): 0.004107
iterate time (version #3): 0.000903
iterate time (version #4): 0.004891
iterate over array (version #1): 0.000277
iterate over array (version #2): 0.000773
what are your results?
So, as it seems there is no usable class in boost which can measure time down to milliseconds, so i rolled my own. (in fact i was a bit shocked that the timer class that ships with Boost only measures used CPU time) I tested it under linux and windows, and its working well:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
| // simple Timer class, Jan 2010, thomas{AT}thomasfischer{DOT}biz
// tested under windows and linux
// license: do whatever you want to do with it ;)
#ifndef TIMER_H__
#define TIMER_H__
// boost timer is awful, measures cpu time on linux only...
// thus we have to hack together some cross platform timer :(
#ifndef WIN32
#include <sys/time.h>
#else
#include <windows.h>
#endif
class Timer
{
protected:
#ifdef WIN32
LARGE_INTEGER start;
#else
struct timeval start;
#endif
public:
Timer()
{
restart();
}
double elapsed()
{
#ifdef WIN32
LARGE_INTEGER tick, ticksPerSecond;
QueryPerformanceFrequency(&ticksPerSecond);
QueryPerformanceCounter(&tick);
return ((double)tick.QuadPart - (double)start.QuadPart) / (double)ticksPerSecond.QuadPart;
#else
struct timeval now;
gettimeofday(&now, NULL);
return (now.tv_sec - start.tv_sec) + (now.tv_usec - start.tv_usec)/1000000.0;
#endif
}
void restart()
{
#ifdef WIN32
QueryPerformanceCounter(&start);
#else
gettimeofday(&start, NULL);
#endif
}
};
#endif //TIMER_H__ |
how you use it:
1
2
3
4
5
6
7
8
9
10
| #include <stdio.h>
#include "Timer.h"
int main(int argc, char **argv)
{
Timer *t = new Timer();
Sleep(1000);
printf("time gone: %f\n", t->elapsed());
return 0;
} |
which results in my windows machine into this output:
time gone: 0.999699
this is a simple example (based on some 2006 rwhois code) how to get further information about an IP address:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
| import os, sys, string, time, getopt, socket, select, re, errno, copy, signal
def queryWhois(query, server='whois.ripe.net'):
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
while 1:
try:
s.connect((server, 43))
except socket.error, (ecode, reason):
if ecode==errno.EINPROGRESS:
continue
elif ecode==errno.EALREADY:
continue
else:
raise socket.error, (ecode, reason)
pass
break
ret = select.select ([s], [s], [], 30)
if len(ret[1])== 0 and len(ret[0]) == 0:
s.close()
raise TimedOut, "on data"
s.setblocking(1)
s.send("%s\n" % query)
page = ""
while 1:
data = s.recv(8196)
if not data: break
page = page + data
pass
s.close()
if string.find(page, "IANA-BLK") != -1:
raise 'no match'
if string.find(page, "Not allocated by APNIC") != -1:
raise 'no match'
return page
if __name__ == "__main__":
if len(sys.argv) != 2:
print "usage: %s <IP address>" % sys.argv[0]
sys.exit(1)
ip = sys.argv[1]
for server in ['whois.arin.net', 'whois.ripe.net', 'whois.apnic.net', 'whois.lacnic.net', 'whois.afrinic.net']:
try:
res = queryWhois(ip, server)
print '======', server
print res
break # we only need the info once
except:
pass |
just run it with the IP as argument
i just coded on this little bit, and i thought it might be worth to share the information as an example of how to roll your own assert, how to jump into the debugger and how to add log messages with proper calling information:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
| // this is the master swith to debug the stream locking/unlocking
#define DEBUGSTREAMFACTORIES
#define OGREFUNCTIONSTRING String(__FUNCTION__)+" @ "+String(__FILE__)+":"+StringConverter::toString(__LINE__)
#ifdef DEBUGSTREAMFACTORIES
# define LOCKSTREAMS() do { LogManager::getSingleton().logMessage("***LOCK: "+OGREFUNCTIONSTRING); lockStreams(); } while(0)
# define UNLOCKSTREAMS() do { LogManager::getSingleton().logMessage("***UNLOCK: "+OGREFUNCTIONSTRING); unlockStreams(); } while(0)
# ifdef WIN32
// __debugbreak will break into the debugger in visual studio
# define MYASSERT(x) do { if(!x) { LogManager::getSingleton().logMessage("***ASSERT FAILED: "+OGREFUNCTIONSTRING); __debugbreak(); }; } while(0)
# else //!WIN32
# define MYASSERT(x) assert(x)
# endif //WIN32
#else //!DEBUGSTREAMFACTORIES
# define LOCKSTREAMS() ((void)0)
# define UNLOCKSTREAMS() ((void)0)
# define MYASSERT(x) ((void)0)
#endif //DEBUGSTREAMFACTORIES |
also, you might enjoy this very well written tips and tricks for asserts: http://cnicholson.net/2009/02/stupid-c-tricks-adventures-in-assert/
btw, happy new year 
this snipped using the windows debugger (shipped with win 2k and after) was very helpful in tracking a problem:
set _NT_DEBUG_LOG_FILE_OPEN=debug-log.txt
ntsd -v -c "kb;q" .exe
also, you might want to read this awesome article:
http://blogs.msdn.com/pfedev/archive/2008/09/26/all-the-ways-to-capture-a-dump.aspx
and generally:
http://blogs.msdn.com/pfedev/
have fun debugging
“A Pattern for Secure Graphical User Interface Systems” by “Thomas Fischer and Ahmad-Reza Sadeghi and Marcel Winandy”
in “3rd International Workshop on Secure systems methodologies using patterns (SPattern’09)”. Publisher: IEEE
read this interesting article to get to know: http://www.xboxist.com/xbox-360/games/xna-community-games-is-a-failed-experiment-010122.php
nice quote:
As Fouts revealed in his blog post, Microsoft takes a base cut of 30% plus an additional 10-30% for advertising. Simply put, 40-60% of profits goes to Microsoft, and the remainder to the developers.