[Bug c++/26122] New: gdb fails to demangle a name

[Bug c++/26122] New: gdb fails to demangle a name

[mgaudet at mozilla dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=26122

            Bug ID: 26122
           Summary: gdb fails to demangle a name
           Product: gdb
           Version: 9.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c++
          Assignee: unassigned at sourceware dot org
          Reporter: mgaudet at mozilla dot com
  Target Milestone: ---

Using a locally compiled GDB 9.2, I'm trying to use `skip -rfunction`,
however,according to `set debug skip on`, the regexp is being applied to the
mangled name, not the demangled name: 

```
(gdb) step 
skip: checking if function
_ZN2JS6HandleINS_11PropertyKeyEEC2IS1_EERKNS_6RootedIT_EENSt9enable_ifIXsr3stdE16is_convertible_vIS5_S1_EEiE4typeE
matches regex .*JS::Handle.*...no.
skip: checking if function
_ZN2JS6HandleINS_11PropertyKeyEEC2IS1_EERKNS_6RootedIT_EENSt9enable_ifIXsr3stdE16is_convertible_vIS5_S1_EEiE4typeE
matches regex .*JS::Handle.*...no.
```

This appears to be because the name cannot be demangled by GDB: 

```
(gdb) demangle
_ZN2JS6HandleINS_11PropertyKeyEEC2IS1_EERKNS_6RootedIT_EENSt9enable_ifIXsr3stdE16is_convertible_vIS5_S1_EEiE4typeE 
Can't demangle
"_ZN2JS6HandleINS_11PropertyKeyEEC2IS1_EERKNS_6RootedIT_EENSt9enable_ifIXsr3stdE16is_convertible_vIS5_S1_EEiE4typeE"
```

Gimli-rs cpp_demangle [1] appears to demangle it correctly: 


```
$ /home/matthew/.cargo/bin/cppfilt 
_ZN2JS6HandleINS_11PropertyKeyEEC2IS1_EERKNS_6RootedIT_EENSt9enable_ifIXsr3stdE16is_convertible_vIS5_S1_EEiE4typeE 
JS::Handle<JS::PropertyKey>::Handle<JS::PropertyKey>(std::enable_if<std::is_convertible_v<JS::PropertyKey,
JS::PropertyKey>, int>::type) 
```

[1]: https://github.com/gimli-rs/cpp_demangle

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug c++/26122] gdb fails to demangle a name

[trass3r at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=26122

trass3r <trass3r at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |trass3r at gmail dot com

--- Comment #1 from trass3r <trass3r at gmail dot com> ---
Confirmed on trunk but it has nothing to do with not being able to demangle.
It happens with a simple std::make_unique<int>(5) call and regex ^std:: too.

The problem is that in infrun.c
function_name_is_marked_for_skip (ecs->stop_func_name,
actually gets the mangled name unlike in other calls to that function.

https://sourceware.org/git/?p=binutils-gdb.git;a=blob;f=gdb/infrun.c;h=938bc081a1c96ab0551046b6cccc777bd40dd78e;hb=HEAD#l6859

The only workaround is to match the mangled name, 'skip -rfu ^_ZSt' in my case.

-- 
You are receiving this mail because:
You are on the CC list for the bug.