Comments by ilostmyid2 (Top 200 by date)

We're using the fork, ffmpeg, with the same functionality.

So why if the function accepts "const obj *" and I pass an "obj *" it doesn't issue a warning?

This is good, but not what I'm looking for. It's just expanding and collapsing some parts of code, not separating a module into separate independent tabs.

If I use the loopback interface, I'll get the same packet on recv. I need to process the sent packet and change it as you see in the code so that recv will receive what I've generated or modified from the interface. For this, loopback interface may not be useful for this purpose.

Yeah I did. Preventing a header file from being included again is not what I asked! Please re-read my question. In header file I need to know the name and path of the file which has included it. This is really a different thing.

Oh man! I described why I'm hesitating to use this method.

Have no idea. The code is in front of you. It just sticks at accept. What can I do? The code above it runs without any problem. What to debug?!

to debug inside accept assembly code?!

Software that use Onvif to communicate with cameras may get motion from this camera too, like Onvif Device Manager. So the problem is out of the camera itself. So I think I should figure my problem out by relying on the Onvif site and documents and asking others who may be more familiar with the protocol. If you need to know the brand, it's an Avigilon camera.

The vendor tells it's Onvif compliant and you're enough to search in that site!

Maybe essentially I should develop a client like Telegram for Telegram servers to do it instead of a bot, because bots are designed for a request/response purpose. Am I right?

The links you provided didn't help me so much. It's not for persons as beginner as who I am.

Thanks for answering my question. The reasons I picked Telegram instead of a web site are:
1. A Telegram channel has no limit. You may upload as much files as you want.
2. There's no need to login to your web site. You've already done, because you use it with your cell phone number which is dedicated to you.
3. I may share it with anyone. In this case, my family.
4. I may be sure that they'll be never removed from the channel.
5. I may use Telegram to browse them. It's a good media viewer.
6. I may search for a specific file by entering its tag. The tag is to be the file's MD5 hash code. So, existence of a specific file is checkable at the moment.
7. I live with Telegram. ;)

I couldn't find a proper and active enough forum for this purpose

I updated the question to include my whole code.

I've exactly did the same. The problem arises when I use dxva2 and I think it's because the format changes to nv12.

Please concentrate on sw_play function. thanx

I wonder why you asked me to do this. Maybe you couldn't follow the link and see the code. I couldn't find an option for inserting the output image here. Just click on the link to see it!

ok thanks anyway

But then, another issue occurs. I get the following error while decoding packets:
[h264 @ 007ecf00] error while decoding MB 68 3, bytestream -6
[h264 @ 007ecf00] concealing 3341 DC, 3341 AC, 3341 MV errors in I frame
[h264 @ 007ecf00] error while decoding MB 67 4, bytestream -9
[h264 @ 007ecf00] concealing 3262 DC, 3262 AC, 3262 MV errors in I frame
....
This happens when I call FFmpegInvoke.avcodec_send_packet, not when I call FFmpegInvoke.avcodec_receive_frame.

Before I come here to see your comment, I tried using FFmpegInvoke wrapper in a C# project with the same work and this time it worked!
My final goal was to convert the code to C# after getting successful result from it. Now that C# code is working, I don't care about the C++ code.

Invalid argument

Unfortunately there's no ffmpeg forum! You believe? At least I couldn't find any. There's chat room or mailing list for ffmpeg, but I couldn't find a forum for this.

The link you introduced doesn't answer my question. Please be more specific. Thanks

I use on-board intel graphics card on some systems and I need my program to run on them as well. Is it possible?

Thank you for replying. Indeed there's no stream. We establish connections from Windows clients to Linux servers via web sockets. Then we process the received packets and extracts video frames, metadata, etc. from them. Then, having the extracted video frames, we want to show them on screen. I know that there's a way to send the frames to GPU so that the graphics card may decode + render them on screen. I need to know how is it possible with fastest and simplest method.
Thanks

In another words, we don't 'have' the video to save or transfer or to do anything with it. We may just 'see' it.

Oh my! Just at the moment I thought I could succeed getting a correct image, I realized that this solution is not proper for me. Our software structure is client/server. Client-side is C# and if we would intend to get the correct image we could use this solution. But at the server-side, which is Linux, we connect to the cameras. When I reviewed the C# codes, I understood that rendering the control's surface is beyond our source control. It's done by the Bosch OCX'es and we have no access to the images/video. What I need is that the video may be got in Linux, recorded as a video stream to HDD, and be sent to client as a video stream. The worst case is when the servers are federated and client will have no access to the stream/camera. For this, I think we should be looking for another solution.
Any idea?

I agree, but I think this is just an alternative. This camera as well as many other cameras from Bosch support Onvif. At least Bosch claims that they do. If I would know Onvif and how to work with the camera by using this protocol, I could get advantage of this without having to use Bosch Video SDK.

I think you answered my question with mentioning to "Bosch Video SDK". After installing it, I found some sources in C# which may be used to solve the problem. Although this solution is specific to Bosch only, our customer is not going to use any other cameras at the moment.

happy xmas 2u2. i still wonder why they avoid a good support while this decreases their customers.

I don't know why they persist not to publicly deliver any useful technical document. I don't know what is the advantage of hiding them and what's the disadvantage of delivering them easily.

Calling me a fainthearted person and persisting on it is not certainly a good solution for the problem, is it?
Anyway, I've spent enough time on the available documents on the website and found not a valuable thing.

my search led me to:
https://ipp.boschsecurity.com/en/news-events/product-news/video-sdk-6-04/video-sdk
with everything other than suitable solution for this purpose

I couldn't find Bosch Video SDK in the search result.

I had seen the url, but the output is not what I expect. It gives just a better shaped of the whole image. It doesn't do rectilinear dewarping.

Thanks Eric for your description and effort is this problem. I believe in most of what you say and don't know what to say more. I also reviewed your code and don't know why it should work properly and mine doesn't. I'm also an experienced programmer and have had many challenges mostly with unmanaged codes in C/C++. I think we don't have to worry about most of them in managed codes like this, although it persist on 'most'. Anyway, after some investigations, I found some clues which I describe as a separate solution as it worked for me in such cases. I found it on StackOverflow.com. Still though, the problem is there and we may never find why it happens.

:)

In the first paragraph, I prefer "not yet" to "no longer". Maybe this happens before the item is completely added to the list box. For the second and third paragraph, I say that maybe it's a bug rather than a problem in my code.
I couldn't find your meaning with wired or mis-wired event. If some other info is required, tell me to share.
Calling dt.Rows.Add is a technique which add columns with their default values, not null!
Yeah, I change the list box data source with a change in the combobox index, because the combobox holds tables names and evidently when a table is changed a new structure should be shown in the list box.
I thought the same. So I put Debug.WriteLine where I thought might be called during calling dt.Rows.Add(). But I found that no other parts of my code are called.
About the DataTable, there's no explicit expectation in this function. The columns may be whatever they are. Although they're in good manner, it can't be the source of the problem.
Since other parts of my code is not getting involved logically and practically during calling dt.Rows.Add, simplifying them can't affect on the symptom.
Thanks for suggesting the commenting out code method. It works because of the simplifying I mentioned. But as I said, it can't help in this case.
At last, I worry about this being a bug. When I googlized it, someone had suggested to add the handler for SelectedIndexChanged after calling dt.Rows.Add, although it's not called anyway!

Thanks Eric for taking time in reviewing my code and for your notes about it. I updated the question and added the stack trace at the point the debugger specifies as the source of the exception. I hope this info may help. As I told before the exception is pointed by compiler to be at line I call dt.Rows.Add().
About disposing the IDisposable objects, I thought GC will dispose them later. You mean it doesn't?
About the combobox content, I fill it with two items in .Designer.

oh my god!
i search for it so much!

can't be simpler than a 2x2 layout ;D

I updated the question. Please refer to the code. It works partially. When I double-click the panels in the first row, they get maximized, but when I click on the 2nd row panels, they expand in the second row. I don't know why they don't get maximized in the whole panel!

i need a way not to change anything regard to visible panels. they get hidden and automatically exposed when i maximize one of them and restore it back.

it doesn't change size of panel which remains shown

indeed i was looking for this. the exception had been muted by the Debug menu. please put it as a solution so that i may accept it. thx

thanks for the trick, i'm going to test it.
but i don't get an exception, this has been the problem. it's just reported in the output window and the code keeps running as there has been no exception.

the problem is that i can't find the code block which raises the exception, because it's being muted.

i had read it in the internet, but as i said to Eric Lynch, the exception is muted and i can't find it.

i think there's a problem. the exception is thrown somewhere in the code, but it's catched there, without being handled. so i can't find it. the exception is muted this way.

sure, but don't know where it occurs in a solution with lots of projects. so that i don't know where to put the try/catch block. this is the problem.

oh ic
yeah, it's better to build it myself. for now, i mark it as answer, because you helped me in this problem. and i hope you'll remain in the thread until it will be resolved.
thx

how to get the latest version? i use ffmpeginvoke.cs

i tried it. the return value of avcodec_find_decoder_by_name is NULL.

thanx, i'll check it. before checking, let me know whether it's required to use ffmpeg or i may access the card directly for this purpose using direcx or something.

i think that the functions like ws_stream_on_connect should have "C" signature. so the module is of type .c not .cpp. so i can't use try/catch directly here. but i did it in another cpp module which this function calls it.

yeah i thought it would help, but nothing is logged. it seems that the process is killed abnormally because of a coredump or something

I need to solve the problem not by using any means, just softwarely.

i expect it anyway... :)

thx, looked for this.
after implode, i may call the function again:
error_reporting(E_ALL);
so that other errors won't be suppressed.
but essentially why this should have been an error?!...

the line below it:
0, Array, 1

oh thanks. I did. this is the result:
NOTICE Array to string conversion on line number 5

so maybe something is different in php config file(s)?

sorry, i thought it was clear enough. now i've updated it with an example.

I see that you didn't get the architecture. Please re-read what I wrote. Id of a row in MonitorGroups table is a reference for a row in Monitors table. For example, if we want to know what monitors are defined in a specified group we run this query:
SELECT * FROM Monitors WHERE GroupId=<id>;

no help. existence of a row in table MonitorGroups with Id 0 is inevitable for creating the foreign key.

the question is exactly this, what causes the date-time not to be synced?

no, it's an array which is returned and i don't change it nor fill another collection or array based on it. i just use it as it's returned. I'll try BindingList for this purpose. thanx. I know few about it, but I hope I'll succeed.

we're programming in C# winforms, web services are accessed via NuSoap in Linux. the return value of web service is a simple structure.

thank you. is there any way other than using BindingList? indeed the data source comes from a web service hosting on linux. can't make it return a BindingList.

it's indeed a RadListView control

it's my own account at dropbox, but ok, where do you want to be uploaded?

indeed i need authentication, but not to define a separate user/pass than the defined user for the linux system itself. i mean how to avoid separation and have not to define separate users for samba?

now what can i do to introduce network as a trusted source?
MS better not to make barriers than making the world a safer place!

I didn't think this may be the cause. A drive is a drive. If it's trusted enough to run/debug a program, it should be trusted to access the PDB file for this purpose!

omg! i copied the whole solution from network drive into a local drive and the problem was resolved! why? the drive is essentially provided by SFTP Net Drive which is mapped from a linux machine. maybe i've to change the settings of Windows/SFTP Net Drive? maybe the problem would happen even if the network drive would be a mapped drive from a Windows share?

they're not absolute paths. so include no drive letter. they're relative paths.

the directories hierarchy is the same. i did a clean and rebuild with no success.

as i told before, the warning is:
The breakpoint will not currently be hit. No symbols have been loaded for this document.
now what can i do? i see the PDB files there.

build mode is Debug. i use F5. i'm not going to step into assemblies. the line with the breakpoint is located in my own code.

The breakpoint will not currently be hit. No symbols have been loaded for this document.

thank you. i read the document and didn't find any clue about how to call the web service ASYNCHRONOUSLY. calling web services is what we're currently doing. but all of them are synchronous calls.

yes. i told it was clear.

1. many programs are written with void main(void) which do many things!
2. the question is clear. i described in glimpse.

ur right, but since programmers write extension modules for apache, i thought it would be a good idea to ask my question here. if you know a more proper site, i would glad if u give me the link.

it can't solve the problem. considering i want to see errors and don't want to see output, this command redirects stderr to stdout. apxs2 writes its output as well as errors to stderr. so i'll get both on stdout. what differ?!

thank u

thank u. it's better that i read and learn more about sql script language.

how?

oh ic
thx
is there any way not to allow invalid values entered manually? ie. without using spin control.

thanks for the code. i wonder why my code didn't work:

dateTimePicker1.Value.AddMinutes(29);

i did. but it doesn't include what i asked. i need it to spin up and down in half-an-hour boundaries.

i gave the link to my code in my comment above.
the behavior of VS2010 is weird. the same code is working in the VS2015 and may be debugged also there. just in VS2010 and just when the program is paused by a breakpoint it can't be continued and the exception is thrown, regardless of where i put the breakpoint.

i say that i can't debug! you say the only way to fix it is to debug?!

in such a case i can't run the program in debug mode without an exception, but i do. i just can't debug.

http://pastebin.com/NC34qRWc

ok, i'm trying to reproduce it in a simpler code. the code is not the cause, because it's being run without any problem.

since it doesn't belong to me and i'm programming for someone else, i can't. but i wish i could.

nothing yet. how do u know that?

oh my!
i changed the Windows to 7 with the same work of success! something seems to be wrong in XP

As I told before, it's before coding. Unless I may connect to the database directly by mentioning the Access file name from inside the program.

it's for when one want to access the db from withing the program. my problem is before beginning programming.

still not in programming to use a class.

Note that I'm still not going to access the database from inside the program. No program is still written.
About 64-bit issues, it's not applicable. I'm using Windows XP and it's 32-bit in architecture. So, the access issue is also not applicable.
About the DB Engine. I thought no other thing is required when I'm using an Access file as my database. When VS has such a dialog to connect to an Access file as a database, it should recognize the database directly without using any engine. I thought all connections to such file should be implemented internally. If I'm in mistake, please let me know what exactly I've to do. What's ACE? Should I install something other than Access and VS?

ok, i got the sense of what you mean. so you're sure that the Telerik code has no problem and if i get the exception it's just because of bad input. right?

rehi
now i'm reading your text. let me say that i'm not so much newbie in programming. if you check the call stack i sent, you see that no clue of my functions is there. this happen inside the Windows and telerik codes itself. i've not their code to set a breakpoint at them and debug. i'm also not responsible for their code. if this call stack causes the error, so others should have the same problem.
so it's better to ask you for a clue.
thx

i think i couldn't describe my problem well. the vmware virtual machine is located on a dual-cpu xeon each have 8 cores. no cpu usage is high. the machine is assigned 10 cpu's, 5 cores from each cpu. when i run 10 threads at a time, i see only the first cpu is involved with my process and threads, while all cores of the other cpu are idle.

thank you for the response. since it's late night now, and i'll not be at work tomorrow, i will check it the next day.
regards

this is really a good answer and includes all guides a complete solution should.
thx

yeah, exactly. thank u. now i've upgraded my vs to 2015 and project is converted. but in vs 2015 the same error happened and when i changed _T to _t the problem was solved. thanks

i switched to vs 2015. in 2010, i got an error like this:
undefined variable ar
sorry if i was lazy

u did it in vs 2010?

the question is clear. cpp compiler which is used by vs2015 can compile the code, while the cpp compiler of vs2010 can't. it's a bug or lack is what i don't know. but there should be a way to update the vs2010 compiler to be able to compile the code as vs2013 and vs2015 do.
upgrading vs environment is a thing and updating the compiler it uses in completely another thing.
for example, gcc maybe upgraded, while ubuntu is the same.

u mean there's no update for VS 2010 to solve the problem and i've no choice other than changing VS version?

hmm, in glimpse, my question is that is there a way to upgrade c++ compiler in VS 2010?

1. I'm not familiar with HDD hardware or firmware. I judge based on what I see. I see that this happens.
2. I'm not going to copy the HDD in an image file. My idea is to copy a hard into another, /dev/sdc to /dev/sda. I'm not the one who decrypts the second HDD. I will give it to Windows and when it finds the same content, I hope it will be able to decrypt it.
3. How switching the power on and off may increase bad sectors?!
4. I found that the HDD becomes more unreadable as much as I use it. Maybe heads are closer to surface than expected and this caused them to scratch the surface more and more. So I thought it may be better not to read it and take it to a service company. I hope they may recover it.

Hi Sergey. I'm happy to have a reply of you. I may code in C# in Windows or C in Ubuntu. Which one is better? I think since i may work with the HDD in lower-level API's in Ubuntu, it's preferred. But both are Ok for me.

i'm not looking for a tool. i'm looking for a way to write the tool.

i need something simpler preferred in ubuntu, like:
void main()
{
char buf[512];
for (int i=begin; i<end; i++)
{
absread(i, "/dev/sdc", 512, buf);
abswrite(i, "/dev/sda", 512, buf);
}
}

i'll certainly try the solution. but the problem is something different. consider and example. sectors 0 to 100 may be read successfully. then sector 101 is read. it's defective. (it's an example). dd or ddrescue tries to read, can't and so skips. my word is that reading from sector 101 causes the drive to fail any further reads, even 0 to 100. for this, i need a program and i think i've to write it to have control over it. if so, i turn the hdd off, and on again, and relaunch my program and tell it to begin from sector 102. maybe 102 to 157 reads successfully and the process will begin with sector 159.

how can i make sure that the problem is the disk surface, head etc. or the controller?
anyway, you don't give me a method for writing the program.

this is the higher level method. first, i've to try to make an exact copy.

yes, the holes will exist anyway. i say the result will be the best approach to the data i can retrieve.

how u say that? regardless of the source hdd how encrypted, the resulting hdd should be the same, because contains the same data. if Windows may unlock the source hdd (which is the case), the same Windows may unlock the destination hdd too, because it encounters the same data. maybe i'm talking about a lower-level treating with an hdd.
please just tell me the way to programmatically copy the source hdd into the dest hdd. this is a programming forum.

i hope ddrescue doesn't fall into the state which it can't read any sector and i've to turn the hdd off and then on.

since the partition is indeed an NTFS partition encrypted by BitLocker, i prefer not to use any higher level program like testdisk and dislocker, etc. i prefer to make a sector-by-sector copy to another hard disk. for the effect i told, which causes the source hdd not to be readable even from safe sectors after i read from special sectors by dd, i still prefer to write a program for this purpose. ddrescue is doing something, but at the same i want to begin writing the program.

Tomas: no, i didn't. how can i install it?
Richard: I'll do it after my tries fail.
Jochen: thx, going to view

there's no option for this. ms also doesn't reply. i tried ms forums.

so, if a new partition of type ntfs is made, this is the Windows that sets the permissions. right?

i followed the link. as well i searched for such an option in accounts section. it doesn't exist.

it's indeed a local account. indeed the local built-in Administrator account!

oh i found that it was not dependent to BitLocker at all!
it was a security issue.
when i referred to security settings i found that the same users (Administrator and System) are granted to access the content of the hard disk. this shows that they use the same SID's. but it couldn't show the owner. i changed it to "Administrators". but this didn't solve the problem. the only thing which caused the content to get browsable and solved the problem was that i change owner to "Everyone"! but why?

thx 4 replies
SID i meant, which is indeed a Globally Unique IDentifier.
when u install Office, for example, you see that its registry entries for CLSID's are the same in all machines. i expect the built-in Administrator has the same SID in all machines. am i right?
i changed all files and folder's permissions with no success.
if the problem is file-system related, even total commander should not be able to open items too.

i answered in the following section.

you should say the same for C++ environment. so why did they put Ctrl+T there?!

consider a situation in which the code is not compilable. ie. when there're some errors in it. eg. when adding a whole new part which code is not yet complete.
i encountered this requirement when i saw the different is aspect while designing and when i run my program.

thanks to both
so i think simply testing a form just to see its aspect before running the program and examine how it really appears at run-time leads to a complex solution. i'm not familiar with "unit test" concept, and as i see after your comments, i've to investigate in it and improve my knowledge about it.
i'll begin with your links and hope to find interesting things.

i'm not expert in this context. would u explain pls?

i'll try .net unit test. is it a utility?
how to setup a precompile command?

domain's group policy should not be applied to local users. should be?

it's mine. i found the reason. but don't know why!
it was because it was joined to a domain. when i disconnected it from domain, all fields got changeable.
i wonder why this should be the cause while it was a local user!

:) i've talked to him today. it seems that the company's policy is not to open the ports.
see, i've a server outside. so client/server solutions may work. is there any?

which FAQ u mean? maybe i didn't bring up a problem and i just exposed a question to discussion. but it doesn't mean that it needed any clearance.

how do you say that? it's not in my mind. i've told it. time syncing is the problem. ok. but i should use a proxy server to connect to the internet. so the solution "settings on my Windows clock" doesn't work, because there's no way to configure it to use a proxy. it just works when i connect to the internet directly. another problem is that the proxy server just routes http packets, not ntp packets. so i can't use software which use this protocol.
you see that i had described the problem well.
the solution is to use htpd.
http://www.rkeene.org/oss/htp/

in stackoverflow, i told the problem and got the answer, and here you're still looking for the problem clarification!
it's really clear. describe what?
thx anyway

yeah, it seems that ur right. it seems to be unicode. but what happened to the original that made the final? i need a rollback algorithm.

thank u :)

but my project is in VS and i need the behavior of gcc in VS. what can i do? do newer versions of VS resolve this issue?

look at my source. with this i still get 0,2:
#include "iostream"

#pragma pack(1)

using namespace std;

void f2()
{
struct s1
{
char b:2;
bool c:1;
long d:4;
short a:9;
};
union s2
{
s1 c;
short d;
};

s2 e;
e.c.d = 5;
e.d = 0; // i expect that a & b of c both get zero
cout << (int) e.c.d; // but b doesn't

cout << "," << sizeof(s1) << endl; // because d doesn't cover b
}

int main()
{
f2();
return 0;
}
shame on u MS! shame on u!

wow man! i even changed places of a and b in s1 and the result is still 0,2! haha! this is the compiler i'm looking for.
i had discovered another elegant capability in gcc before. 0b001100.. is meaningful for gcc, while it's not for msvc. and again another capability is now discovered.

haha, g++ works great. This is the difference between a professional compiler and a commercial one. No Microsoft! Everything is not business. Try to be a bit professional. I tried the code in Ubuntu 12.04. The g++ outputs 0,2 as i expected. Shame on you, Microsoft! Ah!

i prefer not to use the same types to indicate for the compiler what variables must group. for example, i want to set one to 'bool' to assign 'false' or 'true' to it instead of integral values.
since you said that you're certain that there's no way to tell MSVC to change this behavior, i mark your answer as the solution, but i'm still angry that there's no way.
now i'm going to test it in gcc. did u examine BC?

i think i know enough about bitfields after 20 years of programming in C. the fact is that the compiler acts incorrectly. the behavior in unexpected. it seems that MS doesn't believe in packing. as the default structures member alignment is not 1 byte.
i need an option to change this behavior.
any idea?

thanks for your investigation and description. this describes in details what actually is happening, but doesn't explain what to do to reach the same behavior i expect with the same data types. if i'm not wrong, when i was programming in BC++, it would behave correctly. regardless of data types, i've defined explicitly how many bits the variable should occupy. wasting other bits is not justifiable. then, the bits should be looked as the data types i've specified.
anyone knows how to change the vs 2010 compiler behavior to act correctly?!

yeah, but not exactly source control. i need to have my sources everywhere i go. so i put them in DropBox. and i don't want to include garbage files in the folder.

unfortunately none of them include path for creating APS file.

Thanks for the answer.
The behavior is that the view loses the focus.
Let me describe in more details. The pane has a title, which changes in color when it's, let's say, activated. Controls in the pane are a tab control and a custom view of type CSrollView which holds the buttons (images which are simply CWnd's). What i want is that the pane and its children can't get focus and can't get activated. For example, now, when i click on the tab control, it gets activated and receives keyboard events and i may change the selected page in the pane by arrow left and arrow right. I want the "view" remain focused, the pane title color don't change and MFC suppose that the pane and its children can't get focused.

is there any debugging tool to assist me in finding the source of the problem?
can i use crt functions like _CrtSetDbgFlag for this purpose?
there must be a way in locating the problem automatically.

this is true only if my process has created the child process. if the process has been running before my process, or after my process but has been brought up by a user, then i can't check it.

then i should end the process? this way, i need to first execute the program and then figure out that it has been running. what i want is to prevent it to be executed again. there should be an IN parameter for this purpose, like a flag to specify it should be run as single-instance.

thanks for the link. but it's for C#, not C/C++.

in task manager we see list of all running programs/processes. we may also bring a specific program to front. this is exactly what i need to do in my program.

this may help only if the only one who may execute the program be my own program. it doesn't help when the program is already running, even before my program. i want to bring it to front in such a case. it's enough that it be running. in such a case, i don't want to execute it again.

thanks
such a management seems somehow complicated.
isn't there an easier way to check it?
let's concentrate on what i really want. i'm going to execute a program or if it's already running (by my program or user), bring it to front.

yes, i do. indeed it returns the handle to the new process and its process id as well as these info about its primary thread. but how can it help?

2. good idea, but i can't find out when the child program is closed so that i may enable it again.

sorry, it's in C++. i've no control over the child application. i've no access to its source. it's not my program.

thank you for the solution, but it's not applicable in my case, because i've no control over the proxy server.

support team has opened http on port 8080 and not going to open other ports/protocols. i've to use this port and if needed convert other protocols/ports to this.

hmm, i think i could not imply what exactly i meant, sorry.
i'm not going to view sites or access materials in the internet which are not accessible and are beyond the proxy server set in the company i work in.
there are many requirements for accessing the internet directly. instead of all programs to be able to utilize the proxy server, i prefer a separate program redirect them to internet. for example, consider i want to sync Windows time with time servers. here, i can't ask Windows to connect to the time server via a proxy. it won't work. so, i need a program to direct accesses to the internet, while i'm still beyond proxy. Windows should think that it's directly connected to the internet, while it's not and the program is handling the redirection and connect Windows and whole running programs via the proxy server.

thx :)

thanx, but my problem is before this. if a scrollbar receives mouse move or other mouse events like hover, etc. its appearance change, eg. its buttons at start and end and also the thumb button is repainted to show that they're ready to be clicked on or dragged. in my case, no such visual effects appear. it seems that they're not receiving mouse events and the pane instead receives them.
another pane in my program includes a property grid. its scrollbar has no problem. when i referred to its implementation, i saw that the scrollbar there is a separate ctrl. i think i've to use a separate ctrl for scrollbars and can't enable the toolbox ctrl's scrollbars for this purpose. why? i don't know. and i'm also not sure about this idea.

so you suggest to remove the question itself instead of finding a good answer to it?!

ok. thank you for your reply.
although it's not a special code and indeed my code didn't cause the problem and the MFC itself seems to cause the flicker, i updated the question to include to code too.

besides, i realized that the same function may be used to convert a single unicode char to a single multibyte one too.

ok, now i checked WideCharToMultiByte and it works :)
but the problem still exists. i can't force fputws to use this function for converting the input argument to multibyte. it internally uses wcstombs. it seems that the only way is to convert it to multibyte in memory and write the result into the file by fputs, not to use fputws.
anyway, your last comment solved my problem. please add it as a separate solution, so that i can rate (accept) it.
thx

no, i mean a conversion of a single char instead of a string of chars.

thx, i'll check it. does it have a single byte version?

i added the following line:
_locale_t l = _get_current_locale();
i see that l.locinfo.lc_codepage is zero while the system locate is set to 1256. maybe visual studio (2010) doesn't know and may not adapt itself to system locale? wCodePage in l.locinfo.lc_id is also 0!

or r in this code:
wchar_t wc=L'\x633';
char c;
int r=wctomb(&c,wc);
why gets -1?

so the problem gets simpler. in the following code:
wchar_t *uctext=L"\x633";
int sz=wcstombs(NULL, uctext, 1);
what's wrong that causes sz to get -1 and errno get EILSEQ?

thanx to both. yes, i checked errno and found the same value. but it's weird. what's in the string which cannot be converted? we know that all unicode characters have equivalents in multibyte sequence form. i changed it to a simpler unicode string, a single arabic letter seen (U+0633). i also thought the editor may not correctly provide the string. so i used "\x633" instead of entering the letter directly into the string. but the result was the same.
what's wrong?!

yes, but for this trim must be added to the dialog size not to the CStatic size.

thanx 4 the solution :)
i couldn't communicate with the second paragraph. how that i've to add the trim to the size of the static ctrl?!

it's just for me to know when i've posted the question. because i'm not familiar with gregorian date, i use our own calendar. today is 1391/12/24.

ur right. this is a good solution to add a parameter to all web services to know who is the caller and simulate sessions this way. but i like to know whether it may be done automatically somehow eg. by filling some attribute in http protocol header, etc.

client is C# and web services are located on Linux Ubuntu.

in AssemblyInfo.cs whenever i change -1 to 0 and try to build, this file is regenerated (by who?!) and -1 is returned back.
version is important for me. i referred to the link, but it's not applicable for me, because there's no Microsoft.VersionNumber.targets file.

so u believe that the source is the MDB file. how can i check that it's setup to auto-decrement? i found no option when i opened it in access.

thanx for the solution
u know, i include name of a color in an obj because it's enough. there's no need to include all properties of a color also in obj. what can i do if i want to do so?