[Bug 1866604] [NEW] gdbtui source window does't display Unicode text

[Лъчезар Георгиев]

Public bug reported:

In Ubuntu 20.04, I discovered that the gdb 9 TUI source code window
doesn't display properly the UTF-8 characters (Chinese, Cyrillic, etc.)
but displays correctly only the ASCII ones. For example, the following
source code

printf("Здравей, свят!\n");

appears in the source code window as

printf("^^P^ ^P^ ^Q^┗↑@^^P^ ^P^ ^P^ ^P^  ^^Q^┗↑╝^^P^ ^Q^ ^Q^ \n");

The command windows is not affected (when I type non-ASCII characters
into it, they appear OK).

Additionally, I found out the following:

(1) This bug was introduced in version 8.3;
(2) It affects only some CPU architectures.

In more detail, I tested version "Debian 8.3.1-1" on two 64-bit machines
with the same version of Debian (bullseye/sid), one with PowerPC, and
the other with SPARC. The bug shows up only on the SPARC machine, not
the PowerPC one. I then built the generic version 8.3.1 on a 64-bit ARM
machine. The bug didn't show up there. And then, I built it on a 64-bit
MIPS (big endian) machine, where the bug did show up. Repeated this with
version 8.3. The bug showed up again. When I built the generic version
8.2.1 on the same MIPS machine, the bug didn't show up. A "diff -u
gdb-8.2.1/gdb/tui gdb-8.3/gdb/tui" shows a lot of changes made in the
TUI code of version 8.3 to add some colourisation to the TUI, and the
bug was probably introduced during this process. Sorry that I can't help
more to isolate the bug. But it's quite strange that it manifests itself
only on x86-64 (as I had found out previously using Ubuntu 20-04), SPARC
and MIPS, but not on ARM and PowerPC. I suppose that because of this, it
will be rather hard to pinpoint.

This bug is similar to the already fixed bug 1275210 where I also
described it but then realised that it has a different origin and is
therefore a different bug, so I report it as a new one now. Someone
already reported this bug as bug 25296 in the gdb Bugzilla which I
confirmed there.

** Affects: gdb (Ubuntu)
     Importance: Undecided
         Status: New


** Tags: gdb gdbtui utf-8

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to gdb in Ubuntu.
https://bugs.launchpad.net/bugs/1866604

Title:
  gdbtui source window does't display Unicode text

Status in gdb package in Ubuntu:
  New

Bug description:
  In Ubuntu 20.04, I discovered that the gdb 9 TUI source code window
  doesn't display properly the UTF-8 characters (Chinese, Cyrillic,
  etc.) but displays correctly only the ASCII ones. For example, the
  following source code

  printf("Здравей, свят!\n");

  appears in the source code window as

  printf("^^P^ ^P^ ^Q^┗↑@^^P^ ^P^ ^P^ ^P^  ^^Q^┗↑╝^^P^ ^Q^ ^Q^ \n");

  The command windows is not affected (when I type non-ASCII characters
  into it, they appear OK).

  Additionally, I found out the following:

  (1) This bug was introduced in version 8.3;
  (2) It affects only some CPU architectures.

  In more detail, I tested version "Debian 8.3.1-1" on two 64-bit
  machines with the same version of Debian (bullseye/sid), one with
  PowerPC, and the other with SPARC. The bug shows up only on the SPARC
  machine, not the PowerPC one. I then built the generic version 8.3.1
  on a 64-bit ARM machine. The bug didn't show up there. And then, I
  built it on a 64-bit MIPS (big endian) machine, where the bug did show
  up. Repeated this with version 8.3. The bug showed up again. When I
  built the generic version 8.2.1 on the same MIPS machine, the bug
  didn't show up. A "diff -u gdb-8.2.1/gdb/tui gdb-8.3/gdb/tui" shows a
  lot of changes made in the TUI code of version 8.3 to add some
  colourisation to the TUI, and the bug was probably introduced during
  this process. Sorry that I can't help more to isolate the bug. But
  it's quite strange that it manifests itself only on x86-64 (as I had
  found out previously using Ubuntu 20-04), SPARC and MIPS, but not on
  ARM and PowerPC. I suppose that because of this, it will be rather
  hard to pinpoint.

  This bug is similar to the already fixed bug 1275210 where I also
  described it but then realised that it has a different origin and is
  therefore a different bug, so I report it as a new one now. Someone
  already reported this bug as bug 25296 in the gdb Bugzilla which I
  confirmed there.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/gdb/+bug/1866604/+subscriptions