* [PATCH v3] [gdb/tui] Handle unicode chars in prompt
@ 2025-07-06 7:40 Tom de Vries
2025-07-06 7:47 ` Tom de Vries
0 siblings, 1 reply; 3+ messages in thread
From: Tom de Vries @ 2025-07-06 7:40 UTC (permalink / raw)
To: gdb-patches
Let's try to set the prompt using a unicode character, say '❯', aka U+276F
(heavy right-pointing angle quotation mark ornament).
This works fine on an xterm with CLI (with X marking the position of the
blinking cursor):
...
$ gdb -q -ex "set prompt GDB❯ "
GDB❯ X
...
but with TUI:
...
$ gdb -q -tui -ex "set prompt GDB❯ "
...
we get instead:
...
GDB GDB X
...
We can use the test-case gdb.tui/unicode-prompt.exp to get more details, using
tuiterm.
With Term::dump_screen we have:
...
16 (gdb) set prompt GDB❯
17 GDB❯ GDB❯ GDB❯ set prompt (gdb)
18 (gdb)
...
and with Term::dump_screen_with_attrs (summarizing using attribute sets <attrs1>
and <attrs2>):
...
16 (gdb) set prompt GDB❯
17 GDB<attrs1>❯<attrs2> GDB<attrs1>❯<attrs2> GDB<attrs1>❯<attrs2> set prompt (gdb)
18 (gdb)
...
where:
...
<attrs1> == <reverse:1><invisible:1><blinking:1><intensity:bold>
<attrs2> == <reverse:0><invisible:0><blinking:0><intensity:normal>
...
This explains why we didn't see the unicode char on xterm: it's hidden
because the invisible attribute is set.
So, there seem to be two problems:
- the attributes are incorrect, and
- the prompt is repeated a couple of times.
In TUI, the prompt is written out by tui_puts_internal, which outputs one byte
at a time using waddch, which apparently breaks multi-byte char support.
Fix this by detecting multi-byte chars in tui_puts_internal, and printing them using
waddnstr.
Tested on x86_64-linux and x86_64-freebsd.
Reported-By: wuzy01@qq.com
PR tui/28800
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=28800
---
gdb/charset.c | 20 +++++
gdb/charset.h | 7 ++
gdb/testsuite/gdb.tui/unicode-prompt.exp | 71 ++++++++++++++++
gdb/tui/tui-io.c | 100 +++++++++++++++++++----
4 files changed, 183 insertions(+), 15 deletions(-)
create mode 100644 gdb/testsuite/gdb.tui/unicode-prompt.exp
diff --git a/gdb/charset.c b/gdb/charset.c
index 259362563b2..9c0df83c0d3 100644
--- a/gdb/charset.c
+++ b/gdb/charset.c
@@ -690,6 +690,26 @@ wchar_iterator::iterate (enum wchar_iterate_result *out_result,
return -1;
}
+/* See charset.h. */
+
+void
+wchar_iterator::skip (size_t len)
+{
+ m_input += len;
+
+ gdb_assert (len <= m_bytes);
+ m_bytes -= len;
+}
+
+/* See charset.h. */
+
+void
+wchar_iterator::reset (const gdb_byte *input, size_t bytes)
+{
+ m_input = input;
+ m_bytes = bytes;
+}
+
struct charset_vector
{
~charset_vector ()
diff --git a/gdb/charset.h b/gdb/charset.h
index a0f109da5ee..4d68e61f35a 100644
--- a/gdb/charset.h
+++ b/gdb/charset.h
@@ -126,6 +126,13 @@ class wchar_iterator
int iterate (enum wchar_iterate_result *out_result, gdb_wchar_t **out_chars,
const gdb_byte **ptr, size_t *len);
+ /* Increase the input buffer pointer by LEN bytes. */
+ void skip (size_t len);
+
+ /* Reset the input buffer pointer to INPUT and the number of bytes in the
+ input buffer to BYTES. */
+ void reset (const gdb_byte *input, size_t bytes);
+
private:
/* The underlying iconv descriptor. */
diff --git a/gdb/testsuite/gdb.tui/unicode-prompt.exp b/gdb/testsuite/gdb.tui/unicode-prompt.exp
new file mode 100644
index 00000000000..ac2b6202c04
--- /dev/null
+++ b/gdb/testsuite/gdb.tui/unicode-prompt.exp
@@ -0,0 +1,71 @@
+# Copyright 2025 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+require allow_tui_tests
+
+tuiterm_env
+
+save_vars { env(LC_ALL) } {
+ # Override "C" settings from default_gdb_init.
+ setenv LC_ALL "C.UTF-8"
+
+ Term::clean_restart 24 80
+}
+
+if {![Term::enter_tui]} {
+ unsupported "TUI not supported"
+ return
+}
+
+set unicode_char "\u276F"
+set unicode_char_unsupported "❯"
+
+set color_on [string cat {\033} "\[" "31m"]
+set color_off [string cat {\033} "\[" "0m"]
+
+set prompt "GDB$color_on$unicode_char$color_off "
+set prompt_no_color "GDB$unicode_char "
+set prompt_no_color_re [string_to_regexp $prompt_no_color]
+
+if { [ishost *-*-*bsd*] } {
+ set issue_eol "\r\n"
+} else {
+ set issue_eol "\n"
+}
+
+# Set new prompt.
+send_gdb "set prompt $prompt$issue_eol"
+# Set old prompt back.
+send_gdb "set prompt (gdb) $issue_eol"
+
+gdb_assert \
+ { [Term::wait_for "^${prompt_no_color_re}set prompt $gdb_prompt "] } \
+ "prompt with unicode char"
+
+set line [Term::get_line_with_attrs [expr $Term::_cur_row - 1]]
+verbose -log "line with attrs: '$line'"
+
+set prompt_with_attrs_re "GDB<fg:red>$unicode_char<fg:default> "
+set prompt_unsupported_re "GDB$unicode_char_unsupported "
+
+set test "colored unicode char"
+if { [regexp "^${prompt_with_attrs_re}set prompt .*$" $line] } {
+ pass $test
+} elseif { [regexp "^${prompt_unsupported_re}set prompt .*$" $line] } {
+ # I get this on freebsd: no color, and unicode char not recognized.
+ unsupported $test
+} else {
+ fail $test
+}
diff --git a/gdb/tui/tui-io.c b/gdb/tui/tui-io.c
index 1b4cc82cce8..5c39e42362a 100644
--- a/gdb/tui/tui-io.c
+++ b/gdb/tui/tui-io.c
@@ -45,6 +45,7 @@
#include "gdbsupport/unordered_map.h"
#include "pager.h"
#include "gdbsupport/gdb-checked-static-cast.h"
+#include "charset.h"
/* This redefines CTRL if it is not already defined, so it must come
after terminal state related include files like <term.h> and
@@ -539,30 +540,99 @@ tui_puts_internal (WINDOW *w, const char *string, int *height)
char c;
int prev_col = 0;
bool saw_nl = false;
+ size_t skip = 0;
+ wchar_iterator it ((gdb_byte *)string, strlen (string), host_charset (), 1);
- while ((c = *string++) != 0)
+ while (true)
{
- if (c == '\1' || c == '\2')
- {
- /* Ignore these, they are readline escape-marking
- sequences. */
- continue;
- }
+ bool handled = false;
+
+ /* Get iterator in sync with string. */
+ it.skip (skip);
+ skip = 0;
+
+ /* Detect and handle multibyte chars. */
+ {
+ enum wchar_iterate_result res2;
+ gdb_wchar_t *dummy1;
+ const gdb_byte *dummy2;
+ size_t len;
+ int res = it.iterate (&res2, &dummy1, &dummy2, &len);
+ if (res < 0)
+ {
+ /* End of string. */
+ gdb_assert (res2 == wchar_iterate_eof);
+ break;
+ }
+
+ if (res == 0)
+ {
+ if (res2 == wchar_iterate_invalid)
+ {
+ /* Let single-byte char code handle it. */
+ gdb_assert (len == 1);
+ }
+ else if (res2 == wchar_iterate_incomplete)
+ {
+ /* Iterator has been setup to return end-of-string on next
+ call to iterate. Make that an advance-by-one instead, and
+ let single-byte char code handle it. */
+ it.reset ((gdb_byte *)(string + 1), strlen (string + 1));
+ }
+ else
+ gdb_assert_not_reached ("");
+ }
+ else
+ {
+ /* res > 0. */
+ gdb_assert (res2 == wchar_iterate_ok);
+ if (len > 1)
+ {
+ /* Multi-byte char. Handle it. */
+ waddnstr (w, string, len);
+ string += len;
+ handled = true;
+ }
+ else
+ {
+ /* Single-byte char. Let single-byte char code handle it. */
+ gdb_assert (len == 1);
+ }
+ }
+ }
- if (c == '\033')
+ if (!handled)
{
- size_t bytes_read = apply_ansi_escape (w, string - 1);
- if (bytes_read > 0)
+ c = *string++;
+ if (c == '\0')
+ {
+ /* End of string. */
+ break;
+ }
+
+ if (c == '\1' || c == '\2')
{
- string = string + bytes_read - 1;
+ /* Ignore these, they are readline escape-marking
+ sequences. */
continue;
}
- }
- if (c == '\n')
- saw_nl = true;
+ if (c == '\033')
+ {
+ size_t bytes_read = apply_ansi_escape (w, string - 1);
+ if (bytes_read > 0)
+ {
+ skip = bytes_read - 1;
+ string += skip;
+ continue;
+ }
+ }
+
+ if (c == '\n')
+ saw_nl = true;
- do_tui_putc (w, c);
+ do_tui_putc (w, c);
+ }
if (height != nullptr)
{
base-commit: 87f5e2edca1412326ae40489e2780821093481cb
--
2.43.0
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH v3] [gdb/tui] Handle unicode chars in prompt
2025-07-06 7:40 [PATCH v3] [gdb/tui] Handle unicode chars in prompt Tom de Vries
@ 2025-07-06 7:47 ` Tom de Vries
2025-07-18 12:35 ` Tom de Vries
0 siblings, 1 reply; 3+ messages in thread
From: Tom de Vries @ 2025-07-06 7:47 UTC (permalink / raw)
To: gdb-patches
On 7/6/25 09:40, Tom de Vries wrote:
> Let's try to set the prompt using a unicode character, say '❯', aka U+276F
> (heavy right-pointing angle quotation mark ornament).
>
> This works fine on an xterm with CLI (with X marking the position of the
> blinking cursor):
> ...
> $ gdb -q -ex "set prompt GDB❯ "
> GDB❯ X
> ...
> but with TUI:
> ...
> $ gdb -q -tui -ex "set prompt GDB❯ "
> ...
> we get instead:
> ...
> GDB GDB X
> ...
>
> We can use the test-case gdb.tui/unicode-prompt.exp to get more details, using
> tuiterm.
>
> With Term::dump_screen we have:
> ...
> 16 (gdb) set prompt GDB❯
> 17 GDB❯ GDB❯ GDB❯ set prompt (gdb)
> 18 (gdb)
> ...
> and with Term::dump_screen_with_attrs (summarizing using attribute sets <attrs1>
> and <attrs2>):
> ...
> 16 (gdb) set prompt GDB❯
> 17 GDB<attrs1>❯<attrs2> GDB<attrs1>❯<attrs2> GDB<attrs1>❯<attrs2> set prompt (gdb)
> 18 (gdb)
> ...
> where:
> ...
> <attrs1> == <reverse:1><invisible:1><blinking:1><intensity:bold>
> <attrs2> == <reverse:0><invisible:0><blinking:0><intensity:normal>
> ...
>
> This explains why we didn't see the unicode char on xterm: it's hidden
> because the invisible attribute is set.
>
> So, there seem to be two problems:
> - the attributes are incorrect, and
> - the prompt is repeated a couple of times.
>
> In TUI, the prompt is written out by tui_puts_internal, which outputs one byte
> at a time using waddch, which apparently breaks multi-byte char support.
>
> Fix this by detecting multi-byte chars in tui_puts_internal, and printing them using
> waddnstr.
>
This v3 is roughly the same as what I've posted here (
https://sourceware.org/pipermail/gdb-patches/2023-June/200296.html ).
Changes compared to that post:
- rebased on current trunk
- updated test-case for freebsd
- updated test-case copyright year to 2025
- fixed formatting issue in tui-io.c
Thanks,
- Tom
> Tested on x86_64-linux and x86_64-freebsd.
>
> Reported-By: wuzy01@qq.com
>
> PR tui/28800
> Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=28800
> ---
> gdb/charset.c | 20 +++++
> gdb/charset.h | 7 ++
> gdb/testsuite/gdb.tui/unicode-prompt.exp | 71 ++++++++++++++++
> gdb/tui/tui-io.c | 100 +++++++++++++++++++----
> 4 files changed, 183 insertions(+), 15 deletions(-)
> create mode 100644 gdb/testsuite/gdb.tui/unicode-prompt.exp
>
> diff --git a/gdb/charset.c b/gdb/charset.c
> index 259362563b2..9c0df83c0d3 100644
> --- a/gdb/charset.c
> +++ b/gdb/charset.c
> @@ -690,6 +690,26 @@ wchar_iterator::iterate (enum wchar_iterate_result *out_result,
> return -1;
> }
>
> +/* See charset.h. */
> +
> +void
> +wchar_iterator::skip (size_t len)
> +{
> + m_input += len;
> +
> + gdb_assert (len <= m_bytes);
> + m_bytes -= len;
> +}
> +
> +/* See charset.h. */
> +
> +void
> +wchar_iterator::reset (const gdb_byte *input, size_t bytes)
> +{
> + m_input = input;
> + m_bytes = bytes;
> +}
> +
> struct charset_vector
> {
> ~charset_vector ()
> diff --git a/gdb/charset.h b/gdb/charset.h
> index a0f109da5ee..4d68e61f35a 100644
> --- a/gdb/charset.h
> +++ b/gdb/charset.h
> @@ -126,6 +126,13 @@ class wchar_iterator
> int iterate (enum wchar_iterate_result *out_result, gdb_wchar_t **out_chars,
> const gdb_byte **ptr, size_t *len);
>
> + /* Increase the input buffer pointer by LEN bytes. */
> + void skip (size_t len);
> +
> + /* Reset the input buffer pointer to INPUT and the number of bytes in the
> + input buffer to BYTES. */
> + void reset (const gdb_byte *input, size_t bytes);
> +
> private:
>
> /* The underlying iconv descriptor. */
> diff --git a/gdb/testsuite/gdb.tui/unicode-prompt.exp b/gdb/testsuite/gdb.tui/unicode-prompt.exp
> new file mode 100644
> index 00000000000..ac2b6202c04
> --- /dev/null
> +++ b/gdb/testsuite/gdb.tui/unicode-prompt.exp
> @@ -0,0 +1,71 @@
> +# Copyright 2025 Free Software Foundation, Inc.
> +
> +# This program is free software; you can redistribute it and/or modify
> +# it under the terms of the GNU General Public License as published by
> +# the Free Software Foundation; either version 3 of the License, or
> +# (at your option) any later version.
> +#
> +# This program is distributed in the hope that it will be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> +# GNU General Public License for more details.
> +#
> +# You should have received a copy of the GNU General Public License
> +# along with this program. If not, see <http://www.gnu.org/licenses/>.
> +
> +require allow_tui_tests
> +
> +tuiterm_env
> +
> +save_vars { env(LC_ALL) } {
> + # Override "C" settings from default_gdb_init.
> + setenv LC_ALL "C.UTF-8"
> +
> + Term::clean_restart 24 80
> +}
> +
> +if {![Term::enter_tui]} {
> + unsupported "TUI not supported"
> + return
> +}
> +
> +set unicode_char "\u276F"
> +set unicode_char_unsupported "❯"
> +
> +set color_on [string cat {\033} "\[" "31m"]
> +set color_off [string cat {\033} "\[" "0m"]
> +
> +set prompt "GDB$color_on$unicode_char$color_off "
> +set prompt_no_color "GDB$unicode_char "
> +set prompt_no_color_re [string_to_regexp $prompt_no_color]
> +
> +if { [ishost *-*-*bsd*] } {
> + set issue_eol "\r\n"
> +} else {
> + set issue_eol "\n"
> +}
> +
> +# Set new prompt.
> +send_gdb "set prompt $prompt$issue_eol"
> +# Set old prompt back.
> +send_gdb "set prompt (gdb) $issue_eol"
> +
> +gdb_assert \
> + { [Term::wait_for "^${prompt_no_color_re}set prompt $gdb_prompt "] } \
> + "prompt with unicode char"
> +
> +set line [Term::get_line_with_attrs [expr $Term::_cur_row - 1]]
> +verbose -log "line with attrs: '$line'"
> +
> +set prompt_with_attrs_re "GDB<fg:red>$unicode_char<fg:default> "
> +set prompt_unsupported_re "GDB$unicode_char_unsupported "
> +
> +set test "colored unicode char"
> +if { [regexp "^${prompt_with_attrs_re}set prompt .*$" $line] } {
> + pass $test
> +} elseif { [regexp "^${prompt_unsupported_re}set prompt .*$" $line] } {
> + # I get this on freebsd: no color, and unicode char not recognized.
> + unsupported $test
> +} else {
> + fail $test
> +}
> diff --git a/gdb/tui/tui-io.c b/gdb/tui/tui-io.c
> index 1b4cc82cce8..5c39e42362a 100644
> --- a/gdb/tui/tui-io.c
> +++ b/gdb/tui/tui-io.c
> @@ -45,6 +45,7 @@
> #include "gdbsupport/unordered_map.h"
> #include "pager.h"
> #include "gdbsupport/gdb-checked-static-cast.h"
> +#include "charset.h"
>
> /* This redefines CTRL if it is not already defined, so it must come
> after terminal state related include files like <term.h> and
> @@ -539,30 +540,99 @@ tui_puts_internal (WINDOW *w, const char *string, int *height)
> char c;
> int prev_col = 0;
> bool saw_nl = false;
> + size_t skip = 0;
> + wchar_iterator it ((gdb_byte *)string, strlen (string), host_charset (), 1);
>
> - while ((c = *string++) != 0)
> + while (true)
> {
> - if (c == '\1' || c == '\2')
> - {
> - /* Ignore these, they are readline escape-marking
> - sequences. */
> - continue;
> - }
> + bool handled = false;
> +
> + /* Get iterator in sync with string. */
> + it.skip (skip);
> + skip = 0;
> +
> + /* Detect and handle multibyte chars. */
> + {
> + enum wchar_iterate_result res2;
> + gdb_wchar_t *dummy1;
> + const gdb_byte *dummy2;
> + size_t len;
> + int res = it.iterate (&res2, &dummy1, &dummy2, &len);
> + if (res < 0)
> + {
> + /* End of string. */
> + gdb_assert (res2 == wchar_iterate_eof);
> + break;
> + }
> +
> + if (res == 0)
> + {
> + if (res2 == wchar_iterate_invalid)
> + {
> + /* Let single-byte char code handle it. */
> + gdb_assert (len == 1);
> + }
> + else if (res2 == wchar_iterate_incomplete)
> + {
> + /* Iterator has been setup to return end-of-string on next
> + call to iterate. Make that an advance-by-one instead, and
> + let single-byte char code handle it. */
> + it.reset ((gdb_byte *)(string + 1), strlen (string + 1));
> + }
> + else
> + gdb_assert_not_reached ("");
> + }
> + else
> + {
> + /* res > 0. */
> + gdb_assert (res2 == wchar_iterate_ok);
> + if (len > 1)
> + {
> + /* Multi-byte char. Handle it. */
> + waddnstr (w, string, len);
> + string += len;
> + handled = true;
> + }
> + else
> + {
> + /* Single-byte char. Let single-byte char code handle it. */
> + gdb_assert (len == 1);
> + }
> + }
> + }
>
> - if (c == '\033')
> + if (!handled)
> {
> - size_t bytes_read = apply_ansi_escape (w, string - 1);
> - if (bytes_read > 0)
> + c = *string++;
> + if (c == '\0')
> + {
> + /* End of string. */
> + break;
> + }
> +
> + if (c == '\1' || c == '\2')
> {
> - string = string + bytes_read - 1;
> + /* Ignore these, they are readline escape-marking
> + sequences. */
> continue;
> }
> - }
>
> - if (c == '\n')
> - saw_nl = true;
> + if (c == '\033')
> + {
> + size_t bytes_read = apply_ansi_escape (w, string - 1);
> + if (bytes_read > 0)
> + {
> + skip = bytes_read - 1;
> + string += skip;
> + continue;
> + }
> + }
> +
> + if (c == '\n')
> + saw_nl = true;
>
> - do_tui_putc (w, c);
> + do_tui_putc (w, c);
> + }
>
> if (height != nullptr)
> {
>
> base-commit: 87f5e2edca1412326ae40489e2780821093481cb
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH v3] [gdb/tui] Handle unicode chars in prompt
2025-07-06 7:47 ` Tom de Vries
@ 2025-07-18 12:35 ` Tom de Vries
0 siblings, 0 replies; 3+ messages in thread
From: Tom de Vries @ 2025-07-18 12:35 UTC (permalink / raw)
To: gdb-patches
On 7/6/25 09:47, Tom de Vries wrote:
>
> This v3 is roughly the same as what I've posted here ( https://
> sourceware.org/pipermail/gdb-patches/2023-June/200296.html ).
>
> Changes compared to that post:
> - rebased on current trunk
> - updated test-case for freebsd
> - updated test-case copyright year to 2025
> - fixed formatting issue in tui-io.c
I've submitted a v4.
Changes:
- dropped freebsd-specific changes. I'm about to submit a series that
make those changes superfluous.
- adding missing "require {have_host_locale C.UTF-8}"
Thanks,
- Tom
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2025-07-18 12:36 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2025-07-06 7:40 [PATCH v3] [gdb/tui] Handle unicode chars in prompt Tom de Vries
2025-07-06 7:47 ` Tom de Vries
2025-07-18 12:35 ` Tom de Vries
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox