I'd like to print a string, either const char* or std::string, using lldb so that it is human readable. Most importantly, \n's would be printed as a newline. Does anyone know how to do this? I tried the advice given for gdb in this post, however it doesn't seem to work with lldb.
Edit: I'm aware that you can issue the print myString command to print the string, however it doesn't format newline characters (at least not by default):
Most of the time you want to see the literal contents of your strings, so the default lldb behavior for print is correct. However, it would be useful to have a format option to "render" the output in the same way the standard libraries would do a string. That's basically what the gdb "printf" command is. Please file a bug with the lldb.llvm.org bug reporter asking for this.
Just like with gdb, you can get the standard library to render the text for you:
(lldb) expr (void) printf("Some text\nMore text\nEven more text\n")
Some text
More text
Even more text
(lldb)
I cast it to void in this case because I didn't care about the return value, and it makes it harder to see the text.
As was pointed out in the post you referred to, if your output is not going to a terminal somewhere, that isn't helpful, so some explicit "render the output" option would be a good idea as well. But that should only happen if you attach to, rather than run, your program in the debugger.
Related
I have a function in C++ that I am testing, and after careful inspection I'm pretty sure everything is correct. However, I'm still getting a mysterious error relating to the "return" statement at the end of the function.
Where "population" is a real matrix (using the armadillo matrix package).
Looks like the error code represents a Unicode value. Check if the file is clean of characters which shouldn't be there (copy paste into notepad and then copy paste back).
You have accidentally managed to enter the Device Control 3 character (which has the unicode value U+0031) before return and after ;. The character is probably invisible for you, which is why you aren't seeing anything.
Replace those with spaces. You can probably turn your editor into some kind of "show invisibles" mode which might help.
If you are used to using Emacs keybindings and tried to Cx-s Cx-c to save and quit in another IDE ie Xcode it will insert odd unicode characters.
I have a fairly simple program with a vector of characters which is then outputted to a .txt file.
ofstream op ("output.txt");
vector <char> outp;
for(int i=0;i<outp.size();i++){
op<<outp[i]; //the final output of this is incorrect
cout<<outp[i]; //this output is correct
}
op.close();
the text that is output by cout is correct, but when I open the text file that was created, the output is wrong with what look like Chinese characters that shouldn't have been an option for the program to output. For example, when the program should output:
O dsof
And cout prints the right output, the .txt file has this:
O獤景
I have even tried adding the characters into a string before outputting it but it doesn't help. My best guess is that the characters are combining together and getting a different value for unicode or ascii but I don't know enough about character codes to know for sure or how to stop this from happening. Is there a way to correct the output so that it doesn't do this? I am currently using a windows 8.1 computer with code::blocks 12.11 and the GNU GCC compiler in case that helps.
Some text editors try to guess the encoding of a file and occasionally get it wrong. This can particularly happen with very small amounts of text because whatever statistical analysis is being used just doesn't have enough data to make a good conclusion. Window's Notepad has/had an infamous example with the text "Bush hid the facts".
More advanced text editors (for example Notepad++) may either not experience the same problem or may give you options to change what encoding is being assumed. You could use such to verify that the contents of the file are actually correct.
Hex editors/viewers are another way, since they allow you to examine the raw bytes of the file without interpretation. For instance, HxD is a hex editor that I have used in the past.
Alternatively, you can simply output more text. The more there is, generally the less likely something will guess wrong. From some of my experiences, newlines are particularly helpful in convincing the text editor to assume the correct encoding.
there is nothing wrong with your code.
maybe the text editor you use has a default encoding.
use more advanced editors and you will get the right output.
I find it hard to explain but I will try my best. Some times in Linux- in the Terminal- things get printed but you can still write over them. eg when using wget you get a progress bar like this:
[===================> ]
Now if you type something while it is doing this it will 'overwrite' it. My question is how to recreate this in c++.
Will you use something like
cout <<
or something else?
I hope you understand what I am getting at...
btw I am using the most recent version of Arch with xfce4
Printing a carriage return character \r is typically interpreted in Linux as returning you to the beginning of the line. Try this, for example:
std::cout << "Hello\rJ";
The output will be:
Jello
This does depend on your terminal, however, so you should look up the meaning of particular control characters for your terminal.
For a more cross-platform solution and the ability to do more complex text-based user interfaces, take a look at ncurses.
You can print the special character \b to go back one space. Then you can print a space to blank it out, or another character to overwrite what was there. You can also use \r to return to the beginning of the current output line and write again from there.
Controlling the terminal involved sending various escape sequences to it, in order to move the cursor around and such.
http://www.ibiblio.org/pub/historic-linux/ftp-archives/tsx-11.mit.edu/Oct-07-1996/info/vt102.codes
You could also use ncurses to do this.
I need to be able to use utf-8-encoded strings with log4cxx. I can print the strings just fine with std::cout (the characters are displayed correctly). Using log4cxx, i.e. putting the strings into the LOG4CXX_DEBUG() macro with a ConsoleAppender will output "??" instead of the special character. I found one solution:
LOG4CXX_DECODE_CHAR(logstring, str);
LOG4CXX_DEBUG(logstring);
where str is my input string, but this does not work. Anyone have an idea how this might work? I google'd around a bit, but I couldn't find anything useful.
You can use
setlocale(LC_CTYPE, "UTF-8");
to set only the character encoding, without changing any other information about the locale.
I met the same problem and searched and searched. I found this post, It may work, but I don't like the setlocaleish solution. so i made more research, finally the solution came out.
I reconfigure log4cxx and build it, the problem was solved!
add two more configure options in log4cxx:
./configure --prefx=blabla --with-apr=blabla --with-apr-util=blabla --with-charset=utf-8 --with-logchar=utf-8
hope this will help anyone who need it.
One solution is to use
setlocale(LC_ALL, "en_US.UTF-8");
in my main function. This is OK for me, but if you want more localizable applications, this will probably become hard to track/use.
The first answer didn't work for me, the second one is more than i want. So I combined the two answers:
setlocale(LC_CTYPE, "xx_XX.UTF-8"); // or "xx_XX.utf8", it means the same
where xx_XX is some language tag. I tried to log strings in many languages with different alphabets (on LINUX, including Chinese, language left-to-right and rigth-to-left); so I tried:
setlocale(LC_CTYPE, "it_IT.UTF-8");
and it worked with any tested language. I cannot understand why the simple "UTF-8" without indicating a language xx_XX doesn't work, since i use UTF8 to be language-independent and one shouldn't indicate one. (If somebody know the reason also for that, would be an interesting improvement to the answer). Maybe this also depends by Operatin System.
Finally, on Linux you can get a list of the encodings by typing on shell:
# locale -a | grep utf
Okay, I have been researching on how to do this, but say I am running a program that has a whole bit of output on the terminal, how would I clear the screen from within my program so that I can keep my program running?
I know I can just type clear in terminal and it clears it fine, but like I said, for this program it would be more beneficial for me.
I found something that works, however, I'm not sure what it is or what it is doing.
cout << "\033[2J\033[1;1H";
That works but I have no clue what it is, if you could explain it, than I would much appreciate it.
These are ANSI escape codes. The first one (\033[2J) clears the entire screen (J) from top to bottom (2). The second code (\033[1;1H) positions the cursor at row 1, column 1.
All ANSI escapes begin with the sequence ESC[, have zero or more parameters delimited by ;, and end with a command letter (J and H in your case). \033 is the C-style octal sequence for the escape character.
See here for the full roadshow.
Instead of depending on specific escape sequences that may break in unexpected situations (though accepting that trade-off is fine, if it's what you want), you can just do the same thing you'd do at your shell:
std::system("clear");
Though generally system() is to be avoided, for a user-interactive program neither the extra shell parsing nor process overhead is significant. There's no problem with shell escaping either, in this case.
You could always fork/exec to call clear if you did want to avoid system(). If you're already using [n]curses or another terminal library, use that.
For portability you should get the string from termcap's cl (clear) capability (Clear screen and cursor home). (Or use std::system("clear") as told by Roger Pate).
man 3 termcap (in ncurses)
man 5 termcap
set | grep TERMCAP
you can write in a terminal "clear > data" and read in data the escapes sequance
0x1B[H0x1B[2J0x1B[3J
so
std::cout << "\033[H\033[2J\033[3J" ;