[SOLVED] multibyte input in terminal

babbab · 08-24-2019, 09:42 AM

Hello friends

I'm doing c programming in linux and trying to use fgetws with xfce4-terminal. and the problem is after typing multibyte string into terminal using fgetws function, when I delete with backspace, it only deletes half of the character. after pressing enter, fgetws(line,MAXLINES, stdin) != NULL dies off prematurely.

is there alternate to fgetws or terminal library or alternate to stdin with wide stream or other possible solution to this?

NevemTeve · 08-24-2019, 10:08 AM

As a start, please quote the output of command `locale`

babbab · 08-24-2019, 10:23 AM

Code:

LANG=ko_KR.UTF-8
LC_CTYPE="ko_KR.UTF-8"
LC_NUMERIC="ko_KR.UTF-8"
LC_TIME="ko_KR.UTF-8"
LC_COLLATE=C
LC_MONETARY="ko_KR.UTF-8"
LC_MESSAGES="ko_KR.UTF-8"
LC_PAPER="ko_KR.UTF-8"
LC_NAME="ko_KR.UTF-8"
LC_ADDRESS="ko_KR.UTF-8"
LC_TELEPHONE="ko_KR.UTF-8"
LC_MEASUREMENT="ko_KR.UTF-8"
LC_IDENTIFICATION="ko_KR.UTF-8"
LC_ALL=

babbab · 08-24-2019, 11:35 AM

I have a question. what function of library do xfce4-terminal use for command line input? does anyone know?

NevemTeve · 08-24-2019, 11:37 AM

Did you try other terminal-emulators like xterm, rxvt, konsole, gnome-terminal, eterm, etc?
You might find out that the problem isn't related with this xfce4-terminal.

Edit: Did you call setlocale(LC_ALL,"") in your code?

babbab · 08-24-2019, 11:58 AM

I did include setlocale at the beginning of the main() in source code.
when I use xfce4-terminal, it seems to me in the command line that for multibyte letters, it deletes twice for functioning properly.
for ascii, the terminal deletes one character, for multibytes, it deletes two characters for each backspace stroke.

wonder what function from of what library the terminal uses..

NevemTeve · 08-24-2019, 01:03 PM

Actually it is bash (actually: readline) that does this editing.

Did you try other terminal-emulators like xterm, rxvt, konsole, gnome-terminal, eterm, etc?
You might find out that the problem isn't related with this xfce4-terminal.

babbab · 08-24-2019, 01:03 PM

I searched. and the answer was to use readline library
Thank NevemTeve for fast replying

edit: I used xterm with utf8 setting, and fgetws did the same
it deletes only half the character when pressing backspace but internally deletes a wide character

NevemTeve · 08-26-2019, 01:40 AM

I've just started experimenting, first test: read(2):

Code:

terminal LC_CTYPE    stty   test    result
8-bit    .ISO-8859-2 -iutf8 read(2) single-byte, backspace ok
utf8     .UTF-8      iutf8  read(2) utf8-sequences, backspace ok
utf8     .UTF-8      -iutf8 read(2) utf8-sequences, backspace NOT ok

fread(3), fgets(3): same as read(2) -- well, they do call read(2)

Edit: also tried on AIX: very similar, except it doesn't have 'stty iutf8' option, so backscpace (and Ctrl+W) always does the wrong thing.

Edit: Testing fgetws

Code:

terminal LC_CTYPE    stty   test      result
8-bit    .ISO-8859-2 -iutf8 fgetws(3) wide characters, backspace ok
utf8     .UTF8       iutf8  fgetws(3) wide characters, backspace ok
utf8     .UTF8       -iutf8 fgetws(3) errno=84: Invalid or incomplete multibyte or wide character

Note: 'fgetws' doesn't work after 'fgets'; so I created a copy via dup+fdopen

Code:

        int h= dup (0);
        FILE *wstdin= fdopen (h, "r");