Comments by Vaclav_ (Top 35 by date)

OpenGL has a "feature" glFlush described as " force OpenGL to complete the previous code IMMEDIATELY" I am not interested in finding out why , but with current experience with printf and a suggestion to use fflush stdout it looks as this "flush stdout " has similar function. So if this is just another version of cout /cerr lacking synchronization with main process I guess "...And that's the way it is..." (W,Cronkite )

If I am reading the real replies correctly - the scale type has no bearing on automatic type conversion when scale = 1/2; I suppose an example of automatic type conversion would bring another RTFM "reply". Case closed.

You have to declare an instance in one and only one place. Then everywhere else the variable is accessed you use the extern modifier.

Perfect.
Thanks

Internet and forums, including this one, should not be replacement for education - so your best course of action for " where to start" would be to "ask Dr. Thompson". Assuming he /she is a real educator - most forum participants are not.

Perfect solution , beats fiddling with "string".

YES it is a problem
shows that using "2>&1" is bogus IN THIS CASE.
Hence original question #1 answer is incomplete - FOR (MY) LEARNING purpose.

Now BOTH of these EXAMPLES output to console.
BTW I have used HCI as an example, I guess I need to spell that out next time Gees.

cout <<"\t\t\t hciconfig -a 2>&1"<< endl;
system("hciconfig -a 2>&1");
cout <<"\t\t\t\hciconfig -a "<< endl;
system("hciconfig -a ");

stderr to stdout. The "|" (pipe symbol) says to pass all output (i.e the combined stderr and stdout
BUT at that point the stderr was redirected to stdout so there in no stderr as "all output" so what is being commbined / piped ?

I like to add this "comment".
I have asked "how to do it right ..."
CPallini's is the best solution , and with combination of last Richard's comment actually explains why - irregardless of my mistake in using same names - creating ifstream object or creating ifstream object with parameter is the main difference . Thanks for all the comments - SOLVED and more.

Well - you are free to "believe" whatever, but I have noticed that such "funny" commentaries are becoming more and more on this site.Did not used to be that way.
Perhaps periodic refresher reading the instruction would out these commentaries where they belong - on social media.
Happy coding - over and out...

Apparently fstream OPENS the file hence I was actually doing it twice and that
helped" to create the conflict. Using "shadowing" instead of "duplicate " helps with my vocabulary , however I would prefer "duplicate"....

Richard - as always you have a valid point.
What I do object in this case is "smart " comments on # of lines , AI etc.
People like this really should read the instruction "how to answer etc " FIRST.
Such superficial comments do not contribute to the SOLUTION.
" better be silent and be thought off as a fool or speak-up removing any doubts "

So how do I specify that I want to work on parameter (file) passed ?

Be little more understanding would help.
Posting (....you world choice here... such as silly etc) comments DOES NOT HELP the OP.
Posting code and specifying the errors on lines without posting the error or identifying the code lines...
should be pointed out to the poster
or
should be passed over without silly comments
"read the posting instruction " - FTFM

Both nice solutions, but...
it's the terminology which is puzzling
"char *" is a string in C
and
"string" x is a string in C++
Thanks

Richard,
I am looking at two options in
setsockopt(FileDescriptor_socket, SOL_SOCKET, SO_LINGER | SO_REUSEADDR,
(const char*) &lin, sizeof(int));

Looks as SO_LINGER works at "completion of running program" - pretty vague or directly when exit() is processed. Not tested.
SO_REUSEADDR is specifically to allow bind() to reuse the address -that is where the "address already in use " is generated. So far no success.
The man specifically states "for AF_INET " so I am not sure if it will work for AF_BLUTTOTH.
socket(AF_BLUETOOTH, SOCK_STREAM, BTPROTO_RFCOMM);

I have tried both char and string - failed same way.

To "find out what is going on" comment is superfluous. Not really helpful...As far as I remember "open" creates a file IF it does not exist then opens it for whatever mode is specified or in case of stream by the actual "command".
The file location in this case is immaterial, perhaps I will get a different error if file cannot be created "in folder". ( I'll will select non existent folder next ) I am trying to debug this in some logical way by altering analysing the ONLY thing which is changing - the file name.

Here is my latest

ofstream file_1;
file_1.open("file_1.txt"); //open a file
if (file_1.is_open())
cout << "open " << endl;
else
cout << "closed " << endl;

socket.perror_cpp("File 1 state ");
file_1 << "Hello file\n" << 75; //write to it
file_1.close(); //close it

fstream file_2;
file_2.open("file_2.txt"); //open a file
if (file_2.is_open())
cout << "open " << endl;
else
cout << "closed " << endl;
socket.perror_cpp("File 2 state ");
file_2 << "Hello file\n" << 75; //write to it
file_2.close(); //close it

The first "open" works, the second fails - consistently.

Why would different name format behave this way ?

Thanks.
I am not sure how this "forum" works, but I like to ask follow-up questions Since I am not sure how the "mother" project actually links to "bluetooth" library - for this to work - does GCC wants full file name libbluetooth.a?
Do I have to supply path to the whatever file name is correct ( full or "castrated" )? It is falling now ( g++: error: libbluetooth.a: No such file or directory) and I am just fumbling around trying different ways.
I have implemented bluetooth library long time ago and I am sure it has "system " path - no (-L).

Go thru your code and remove the duplicates.

"Note: there were 11 duplicate class definitions."

???????????????????????

After further research I woudl add - I am missing link to object file - perhaps impossible to link object files between projects.

I was hoping I do not have to make it a library.

Which would "generate" another question - how does "normal project" links to object files?

SOLVED
Purfect...perror_cpp prints EXACTLY where it does most good.
Tracking socket behaviour in REAL time!
Thanks

I feel this may shed some light on the issue.
Here are quotes from elsewhere:
"by default cout is buffered " and "new line at the end of cout will clear the buffer"
With that - the issue is NOT with cout anyway , but with cerr / perror.

I can add endl or /n to cerr or add flush to whole cerr line.

Not sure how to "add" such to perror - since it prints the optional message first then the actual text of the "errno" .

I am sorry, since this has been a long time issue, I totally forgot about "flush". Yes, I did use it and it did not help. But I'll try it again.
I'll give strerror a try too. Thanks

While tracking other issue noticed this.
It looks as msbbox was not initialized properly,
however the output to the screen contained additional printable
"characters. I'll check the latest "save" file next to verify if
"Received SIGHUP or SIGTERM" generated it.
++ sudo nano /etc/network/interfaces
Received SIGHUP or SIGTERM
+ whiptail --title 'Command sudo nano /etc/network/interfaces output 1403' --sep' 17 80 10ut --scrolltext --msgbox '
+ echo 'STOP HERE 1413'
STOP HERE 1413

I asked for explanation when to use command substitution or redirection.
Since everything in Linux is a file, I assume I can "redirect" from stdout to VARIABLE - in this case "index".
I did specify that I did attempted to do so.
I will eventually figure out how.

I am a little late in posting my comments, but basically I had a failing hard drive which eventually crashed for good.
Lesson learned - if in doubt - toss it out!

Thanks, this will get me started.
I guess I still need to make sure the SDK is valid for the IDE.
For now I'll stay with VC++ 6.0