I have written a patch for this issue.
Logged in Ubuntu: https://bugs.launchpad.net/ubuntu/+source/procps/+bug/318221
I have tried to contact the upstream, but have failed to do so - the
email addresses in the man page do not seem to be read, and the
mailing list contains only spam.
--
Jarrod Lowe
diff -du -r procps-3.2.7-orig/AUTHORS procps-3.2.7/AUTHORS
--- procps-3.2.7-orig/AUTHORS 2002-10-10 22:14:34.000000000 +0100
+++ procps-3.2.7/AUTHORS 2008-10-04 14:50:10.000000000 +0100
@@ -47,4 +47,5 @@
watch:
Tony Rems <re...@unisoft.com>
Mike Coleman <m...@acm.org>
+Jarrod Lowe <pro...@rrod.net>
diff -du -r procps-3.2.7-orig/Makefile procps-3.2.7/Makefile
--- procps-3.2.7-orig/Makefile 2006-06-24 10:02:25.000000000 +0100
+++ procps-3.2.7/Makefile 2008-10-04 15:05:50.000000000 +0100
@@ -68,6 +68,7 @@
_TARFILES := Makefile
CURSES := -lncurses
+CURSESW := -lncursesw
# This seems about right for the dynamic library stuff.
# Something like this is probably needed to make the SE Linux
@@ -119,7 +120,7 @@
# Unlike the kernel one, this check_gcc goes all the way to
# producing an executable. There might be a -m64 that works
# until you go looking for a 64-bit curses library.
-check_gcc = $(shell if $(CC) $(ALL_CPPFLAGS) $(ALL_CFLAGS) dummy.c $(ALL_LDFLAGS) $(1) -o /dev/null $(CURSES) > /dev/null 2>&1; then echo "$(1)"; else echo "$(2)"; fi ;)
+check_gcc = $(shell if $(CC) $(ALL_CPPFLAGS) $(ALL_CFLAGS) dummy.c $(ALL_LDFLAGS) $(1) -o /dev/null $(CURSES) $(CURSESW) > /dev/null 2>&1; then echo "$(1)"; else echo "$(2)"; fi ;)
# Be 64-bit if at all possible. In a cross-compiling situation, one may
# do "make m64=-m32 lib64=lib" to produce 32-bit executables. DO NOT
@@ -250,7 +251,7 @@
$(CC) $(ALL_CFLAGS) $^ $(ALL_LDFLAGS) -o $@ $(CURSES)
watch: % : %.o
- $(CC) $(ALL_CFLAGS) $^ $(ALL_LDFLAGS) -o $@ $(CURSES)
+ $(CC) $(ALL_CFLAGS) $^ $(ALL_LDFLAGS) -o $@ $(CURSESW)
############ progX --> progY
diff -du -r procps-3.2.7-orig/watch.1 procps-3.2.7/watch.1
--- procps-3.2.7-orig/watch.1 2003-02-09 07:05:25.000000000 +0000
+++ procps-3.2.7/watch.1 2008-10-04 14:46:46.000000000 +0100
@@ -79,9 +79,21 @@
.PP
Non-printing characters are stripped from program output. Use "cat -v" as
part of the command pipeline if you want to see them.
+.PP
+Combining Characters that are supposed to display on the character at the
+last column on the screen may display one column early, or they may not
+display at all.
+.PP
+Combining Characters never count as different in
+.I --differences
+mode. Only the base character counts.
+.PP
+Blank lines directly after a line which ends in the last column do not
+display.
.SH AUTHORS
The original
.B watch
was written by Tony Rems <re...@unisoft.com> in 1991, with mods and
corrections by Francois Pinard. It was reworked and new features added by
-Mike Coleman <m...@acm.org> in 1999.
+Mike Coleman <m...@acm.org> in 1999. Unicode support was added in 2009
+by Jarrod Lowe <pro...@rrod.net>.
diff -du -r procps-3.2.7-orig/watch.c procps-3.2.7/watch.c
--- procps-3.2.7-orig/watch.c 2006-06-17 10:18:38.000000000 +0100
+++ procps-3.2.7/watch.c 2008-10-04 15:06:09.000000000 +0100
@@ -8,14 +8,17 @@
* Mike Coleman <m...@acm.org>.
*
* Changes by Albert Cahalan, 2002-2003.
+ *
+ * Unicode Support added by Jarrod Lowe <pro...@rrod.net> in 2009.
*/
-#define VERSION "0.2.0"
+#define VERSION "0.3.0"
+#include <wchar.h>
#include <ctype.h>
#include <getopt.h>
#include <signal.h>
-#include <ncurses.h>
+#include <ncursesw/ncurses.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
@@ -25,6 +28,7 @@
#include <termios.h>
#include <locale.h>
#include "proc/procps.h"
+#include <errno.h>
#ifdef FORCE_8BIT
#undef isprint
@@ -134,6 +138,32 @@
}
}
+// read a wide character from a popen'd stream
+#define MAX_ENC_BYTES 16
+wint_t my_getwc(FILE *s);
+wint_t my_getwc(FILE *s) {
+ char i[MAX_ENC_BYTES]; //assuming no encoding ever consumes more than 16 bytes
+ int byte = 0;
+ int convert;
+ int x;
+ wchar_t rval;
+ while(1) {
+ i[byte] = getc(s);
+ if (i[byte]==EOF) { return WEOF; }
+ byte++;
+ errno = 0;
+ mbtowc(NULL, NULL, 0);
+ convert = mbtowc(&rval, i, byte);
+ x = errno;
+ if(convert > 0) { return rval; } //legal conversion
+ if(byte == MAX_ENC_BYTES) {
+ while(byte > 1) { ungetc(i[--byte], s); } //at least *try* to fix up
+ errno = -EILSEQ;
+ return WEOF;
+ }
+ }
+}
+
int
main(int argc, char *argv[])
{
@@ -143,7 +173,10 @@
option_help = 0, option_version = 0;
double interval = 2;
char *command;
+ wchar_t *wcommand = NULL;
int command_length = 0; /* not including final \0 */
+ int wcommand_columns = 0; /* not including final \0 */
+ int wcommand_characters = 0; /* not including final \0 */
setlocale(LC_ALL, "");
progname = argv[0];
@@ -216,6 +249,23 @@
command[command_length] = '\0';
}
+ // convert to wide for printing purposes
+ //mbstowcs(NULL, NULL, 0);
+ wcommand_characters = mbstowcs(NULL, command, 0);
+ if(wcommand_characters < 0) {
+ fprintf(stderr, "Unicode Handling Error\n");
+ exit(1);
+ }
+ wcommand = (wchar_t*)malloc((wcommand_characters+1) * sizeof(wcommand));
+ if(wcommand == NULL) {
+ fprintf(stderr, "Unicode Handling Error (malloc)\n");
+ exit(1);
+ }
+ mbstowcs(wcommand, command, wcommand_characters+1);
+ wcommand_columns = wcswidth(wcommand, -1);
+
+
+
get_terminal_size();
/* Catch keyboard interrupts so we can put tty back in a sane state. */
@@ -252,12 +302,44 @@
if (show_title) {
// left justify interval and command,
// right justify time, clipping all to fit window width
- asprintf(&header, "Every %.1fs: %.*s",
- interval, min(width - 1, command_length), command);
- mvaddstr(0, 0, header);
- if (strlen(header) > (size_t) (width - tsl - 1))
- mvaddstr(0, width - tsl - 4, "... ");
- mvaddstr(0, width - tsl + 1, ts);
+
+ int hlen = asprintf(&header, "Every %.1fs: ", interval);
+
+ // the rules:
+ // width < tsl : print nothing
+ // width < tsl + hlen + 1: print ts
+ // width = tsl + hlen + 1: print header, ts
+ // width < tsl + hlen + 4: print header, ..., ts
+ // width < tsl + hlen + wcommand_columns: print header, truncated wcommand, ..., ts
+ // width > "": print header, wcomand, ts
+ // this is slightly different from how it used to be
+ if(width >= tsl) {
+ if(width >= tsl + hlen + 1) {
+ mvaddstr(0, 0, header);
+ if(width >= tsl + hlen + 2) {
+ if(width < tsl + hlen + 4) {
+ mvaddstr(0, width - tsl - 4, "... ");
+ }else{
+ if(width < tsl + hlen + wcommand_columns) {
+ // print truncated
+ int avail_columns = width - tsl - hlen;
+ int using_columns = wcommand_columns;
+ int using_characters = wcommand_characters;
+ while(using_columns > avail_columns - 4) {
+ using_characters--;
+ using_columns = wcswidth(wcommand, using_characters);
+ }
+ mvaddnwstr(0, hlen, wcommand, using_characters);
+ mvaddstr(0, width - tsl - 4, "... ");
+ }else{
+ mvaddwstr(0, hlen, wcommand);
+ }
+ }
+ }
+ }
+ mvaddstr(0, width - tsl + 1, ts);
+ }
+
free(header);
}
@@ -268,47 +350,62 @@
for (y = show_title; y < height; y++) {
int eolseen = 0, tabpending = 0;
+ wint_t carry = WEOF;
for (x = 0; x < width; x++) {
- int c = ' ';
+ wint_t c = L' ';
int attr = 0;
if (!eolseen) {
/* if there is a tab pending, just spit spaces until the
next stop instead of reading characters */
if (!tabpending)
- do
- c = getc(p);
- while (c != EOF && !isprint(c)
- && c != '\n'
- && c != '\t');
- if (c == '\n')
+ do {
+ if(carry == WEOF) {
+ c = my_getwc(p);
+ }else{
+ c = carry;
+ carry = WEOF;
+ }
+ }while (c != WEOF && !isprint(c) && c<128
+ && wcwidth(c) == 0
+ && c != L'\n'
+ && c != L'\t');
+ if (c == L'\n')
if (!oldeolseen && x == 0) {
x = -1;
continue;
} else
eolseen = 1;
- else if (c == '\t')
+ else if (c == L'\t')
tabpending = 1;
- if (c == EOF || c == '\n' || c == '\t')
- c = ' ';
+ if (x==width-1 && wcwidth(c)==2) {
+ y++;
+ x = -1; //process this double-width
+ carry = c; //character on the next line
+ continue; //because it won't fit here
+ }
+ if (c == WEOF || c == L'\n' || c == L'\t')
+ c = L' ';
if (tabpending && (((x + 1) % 8) == 0))
tabpending = 0;
}
move(y, x);
if (option_differences) {
- int oldch = inch();
- char oldc = oldch & A_CHARTEXT;
+ cchar_t oldc;
+ in_wch(&oldc);
attr = !first_screen
- && (c != oldc
+ && ((wchar_t)c != oldc.chars[0]
||
(option_differences_cumulative
- && (oldch & A_ATTRIBUTES)));
+ && (oldc.attr & A_ATTRIBUTES)));
}
if (attr)
standout();
- addch(c);
+ addnwstr((wchar_t*)&c,1);
if (attr)
standend();
+ if(wcwidth(c) == 0) { x--; }
+ if(wcwidth(c) == 2) { x++; }
}
oldeolseen = eolseen;
}
diff -du -r procps-3.2.8-orig/AUTHORS procps-3.2.8/AUTHORS
--- procps-3.2.8-orig/AUTHORS 2002-10-10 22:14:34.000000000 +0100
+++ procps-3.2.8/AUTHORS 2008-10-04 14:50:10.000000000 +0100
@@ -47,4 +47,5 @@
watch:
Tony Rems <re...@unisoft.com>
Mike Coleman <m...@acm.org>
+Jarrod Lowe <pro...@rrod.net>
diff -du -r procps-3.2.8-orig/Makefile procps-3.2.8/Makefile
--- procps-3.2.8-orig/Makefile 2006-06-24 10:02:25.000000000 +0100
+++ procps-3.2.8/Makefile 2008-10-04 15:05:50.000000000 +0100
@@ -68,6 +68,7 @@
_TARFILES := Makefile
CURSES := -lncurses
+CURSESW := -lncursesw
# This seems about right for the dynamic library stuff.
# Something like this is probably needed to make the SE Linux
@@ -119,7 +120,7 @@
# Unlike the kernel one, this check_gcc goes all the way to
# producing an executable. There might be a -m64 that works
# until you go looking for a 64-bit curses library.
-check_gcc = $(shell if $(CC) $(ALL_CPPFLAGS) $(ALL_CFLAGS) dummy.c $(ALL_LDFLAGS) $(1) -o /dev/null $(CURSES) > /dev/null 2>&1; then echo "$(1)"; else echo "$(2)"; fi ;)
+check_gcc = $(shell if $(CC) $(ALL_CPPFLAGS) $(ALL_CFLAGS) dummy.c $(ALL_LDFLAGS) $(1) -o /dev/null $(CURSES) $(CURSESW) > /dev/null 2>&1; then echo "$(1)"; else echo "$(2)"; fi ;)
# Be 64-bit if at all possible. In a cross-compiling situation, one may
# do "make m64=-m32 lib64=lib" to produce 32-bit executables. DO NOT
@@ -250,7 +251,7 @@
$(CC) $(ALL_CFLAGS) $^ $(ALL_LDFLAGS) -o $@ $(CURSES)
watch: % : %.o
- $(CC) $(ALL_CFLAGS) $^ $(ALL_LDFLAGS) -o $@ $(CURSES)
+ $(CC) $(ALL_CFLAGS) $^ $(ALL_LDFLAGS) -o $@ $(CURSESW)
############ progX --> progY
diff -du -r procps-3.2.8-orig/watch.1 procps-3.2.8/watch.1
--- procps-3.2.8-orig/watch.1 2003-02-09 07:05:25.000000000 +0000
+++ procps-3.2.8/watch.1 2008-10-04 14:46:46.000000000 +0100
@@ -79,9 +79,21 @@
.PP
Non-printing characters are stripped from program output. Use "cat -v" as
part of the command pipeline if you want to see them.
+.PP
+Combining Characters that are supposed to display on the character at the
+last column on the screen may display one column early, or they may not
+display at all.
+.PP
+Combining Characters never count as different in
+.I --differences
+mode. Only the base character counts.
+.PP
+Blank lines directly after a line which ends in the last column do not
+display.
.SH AUTHORS
The original
.B watch
was written by Tony Rems <re...@unisoft.com> in 1991, with mods and
corrections by Francois Pinard. It was reworked and new features added by
-Mike Coleman <m...@acm.org> in 1999.
+Mike Coleman <m...@acm.org> in 1999. Unicode support was added in 2009
+by Jarrod Lowe <pro...@rrod.net>.
diff -du -r procps-3.2.8-orig/watch.c procps-3.2.8/watch.c
--- procps-3.2.8-orig/watch.c 2006-06-17 10:18:38.000000000 +0100
+++ procps-3.2.8/watch.c 2008-10-04 15:06:09.000000000 +0100
@@ -8,14 +8,17 @@
* Mike Coleman <m...@acm.org>.
*
* Changes by Albert Cahalan, 2002-2003.
+ *
+ * Unicode Support added by Jarrod Lowe <pro...@rrod.net> in 2009.
*/
-#define VERSION "0.2.0"
+#define VERSION "0.3.0"
+#include <wchar.h>
#include <ctype.h>
#include <getopt.h>
#include <signal.h>
-#include <ncurses.h>
+#include <ncursesw/ncurses.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
@@ -25,6 +28,7 @@
#include <termios.h>
#include <locale.h>
#include "proc/procps.h"
+#include <errno.h>
#ifdef FORCE_8BIT
#undef isprint
@@ -134,6 +138,32 @@
}
}
+// read a wide character from a popen'd stream
+#define MAX_ENC_BYTES 16
+wint_t my_getwc(FILE *s);
+wint_t my_getwc(FILE *s) {
+ char i[MAX_ENC_BYTES]; //assuming no encoding ever consumes more than 16 bytes
+ int byte = 0;
+ int convert;
+ int x;
+ wchar_t rval;
+ while(1) {
+ i[byte] = getc(s);
+ if (i[byte]==EOF) { return WEOF; }
+ byte++;
+ errno = 0;
+ mbtowc(NULL, NULL, 0);
+ convert = mbtowc(&rval, i, byte);
+ x = errno;
+ if(convert > 0) { return rval; } //legal conversion
+ if(byte == MAX_ENC_BYTES) {
+ while(byte > 1) { ungetc(i[--byte], s); } //at least *try* to fix up
+ errno = -EILSEQ;
+ return WEOF;
+ }
+ }
+}
+
int
main(int argc, char *argv[])
{
@@ -143,7 +173,10 @@
option_help = 0, option_version = 0;
double interval = 2;
char *command;
+ wchar_t *wcommand = NULL;
int command_length = 0; /* not including final \0 */
+ int wcommand_columns = 0; /* not including final \0 */
+ int wcommand_characters = 0; /* not including final \0 */
setlocale(LC_ALL, "");
progname = argv[0];
@@ -216,6 +249,23 @@
command[command_length] = '\0';
}
+ // convert to wide for printing purposes
+ //mbstowcs(NULL, NULL, 0);
+ wcommand_characters = mbstowcs(NULL, command, 0);
+ if(wcommand_characters < 0) {
+ fprintf(stderr, "Unicode Handling Error\n");
+ exit(1);
+ }
+ wcommand = (wchar_t*)malloc((wcommand_characters+1) * sizeof(wcommand));
+ if(wcommand == NULL) {
+ fprintf(stderr, "Unicode Handling Error (malloc)\n");
+ exit(1);
+ }
+ mbstowcs(wcommand, command, wcommand_characters+1);
+ wcommand_columns = wcswidth(wcommand, -1);
+
+
+
get_terminal_size();
/* Catch keyboard interrupts so we can put tty back in a sane state. */
@@ -252,12 +302,44 @@
if (show_title) {
// left justify interval and command,
// right justify time, clipping all to fit window width
- asprintf(&header, "Every %.1fs: %.*s",
- interval, min(width - 1, command_length), command);
- mvaddstr(0, 0, header);
- if (strlen(header) > (size_t) (width - tsl - 1))
- mvaddstr(0, width - tsl - 4, "... ");
- mvaddstr(0, width - tsl + 1, ts);
+
+ int hlen = asprintf(&header, "Every %.1fs: ", interval);
+
+ // the rules:
+ // width < tsl : print nothing
+ // width < tsl + hlen + 1: print ts
+ // width = tsl + hlen + 1: print header, ts
+ // width < tsl + hlen + 4: print header, ..., ts
+ // width < tsl + hlen + wcommand_columns: print header, truncated wcommand, ..., ts
+ // width > "": print header, wcomand, ts
+ // this is slightly different from how it used to be
+ if(width >= tsl) {
+ if(width >= tsl + hlen + 1) {
+ mvaddstr(0, 0, header);
+ if(width >= tsl + hlen + 2) {
+ if(width < tsl + hlen + 4) {
+ mvaddstr(0, width - tsl - 4, "... ");
+ }else{
+ if(width < tsl + hlen + wcommand_columns) {
+ // print truncated
+ int avail_columns = width - tsl - hlen;
+ int using_columns = wcommand_columns;
+ int using_characters = wcommand_characters;
+ while(using_columns > avail_columns - 4) {
+ using_characters--;
+ using_columns = wcswidth(wcommand, using_characters);
+ }
+ mvaddnwstr(0, hlen, wcommand, using_characters);
+ mvaddstr(0, width - tsl - 4, "... ");
+ }else{
+ mvaddwstr(0, hlen, wcommand);
+ }
+ }
+ }
+ }
+ mvaddstr(0, width - tsl + 1, ts);
+ }
+
free(header);
}
@@ -268,47 +350,62 @@
for (y = show_title; y < height; y++) {
int eolseen = 0, tabpending = 0;
+ wint_t carry = WEOF;
for (x = 0; x < width; x++) {
- int c = ' ';
+ wint_t c = L' ';
int attr = 0;
if (!eolseen) {
/* if there is a tab pending, just spit spaces until the
next stop instead of reading characters */
if (!tabpending)
- do
- c = getc(p);
- while (c != EOF && !isprint(c)
- && c != '\n'
- && c != '\t');
- if (c == '\n')
+ do {
+ if(carry == WEOF) {
+ c = my_getwc(p);
+ }else{
+ c = carry;
+ carry = WEOF;
+ }
+ }while (c != WEOF && !isprint(c) && c<128
+ && wcwidth(c) == 0
+ && c != L'\n'
+ && c != L'\t');
+ if (c == L'\n')
if (!oldeolseen && x == 0) {
x = -1;
continue;
} else
eolseen = 1;
- else if (c == '\t')
+ else if (c == L'\t')
tabpending = 1;
- if (c == EOF || c == '\n' || c == '\t')
- c = ' ';
+ if (x==width-1 && wcwidth(c)==2) {
+ y++;
+ x = -1; //process this double-width
+ carry = c; //character on the next line
+ continue; //because it won't fit here
+ }
+ if (c == WEOF || c == L'\n' || c == L'\t')
+ c = L' ';
if (tabpending && (((x + 1) % 8) == 0))
tabpending = 0;
}
move(y, x);
if (option_differences) {
- chtype oldch = inch();
- char oldc = oldch & A_CHARTEXT;
+ cchar_t oldc;
+ in_wch(&oldc);
attr = !first_screen
- && ((char)c != oldc
+ && ((wchar_t)c != oldc.chars[0]
||
(option_differences_cumulative
- && (oldch & A_ATTRIBUTES)));
+ && (oldc.attr & A_ATTRIBUTES)));
}
if (attr)
standout();
- addch(c);
+ addnwstr((wchar_t*)&c,1);
if (attr)
standend();
+ if(wcwidth(c) == 0) { x--; }
+ if(wcwidth(c) == 2) { x++; }
}
oldeolseen = eolseen;
}
