* [Bug libstdc++/110196] stacktrace_entry does not print origin binary filename
2023-06-09 15:42 [Bug libstdc++/110196] New: stacktrace_entry does not print origin binary filename fabian_kessler at gmx dot de
@ 2023-06-09 15:56 ` pinskia at gcc dot gnu.org
2023-06-09 15:58 ` fabian_kessler at gmx dot de
` (10 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: pinskia at gcc dot gnu.org @ 2023-06-09 15:56 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110196
--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I am trying to understand the issue here?
Is what you are asking is that the " at :0" should not be printed out? Or the
issue is the source line cannot be found.
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/110196] stacktrace_entry does not print origin binary filename
2023-06-09 15:42 [Bug libstdc++/110196] New: stacktrace_entry does not print origin binary filename fabian_kessler at gmx dot de
2023-06-09 15:56 ` [Bug libstdc++/110196] " pinskia at gcc dot gnu.org
@ 2023-06-09 15:58 ` fabian_kessler at gmx dot de
2023-06-09 16:07 ` fabian_kessler at gmx dot de
` (9 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: fabian_kessler at gmx dot de @ 2023-06-09 15:58 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110196
--- Comment #2 from fabian_kessler at gmx dot de <fabian_kessler at gmx dot de> ---
(In reply to Andrew Pinski from comment #1)
> I am trying to understand the issue here?
> Is what you are asking is that the " at :0" should not be printed out? Or
> the issue is the source line cannot be found.
The issue is, that the origin library, in this case libc.so/gtk.so/glib.so is
not shown. In all other lines, xournalpp as executable is missing.
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/110196] stacktrace_entry does not print origin binary filename
2023-06-09 15:42 [Bug libstdc++/110196] New: stacktrace_entry does not print origin binary filename fabian_kessler at gmx dot de
2023-06-09 15:56 ` [Bug libstdc++/110196] " pinskia at gcc dot gnu.org
2023-06-09 15:58 ` fabian_kessler at gmx dot de
@ 2023-06-09 16:07 ` fabian_kessler at gmx dot de
2023-06-09 16:09 ` [Bug libstdc++/110196] stacktrace_entry does not print which library the function was from pinskia at gcc dot gnu.org
` (8 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: fabian_kessler at gmx dot de @ 2023-06-09 16:07 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110196
--- Comment #3 from fabian_kessler at gmx dot de <fabian_kessler at gmx dot de> ---
It may be, because backtrace_create_state is called without a filename.
Also, backtrace_create_state is ment to be called just once per execution,
because it allocates unfreeable memory. Maybe, a singleton would be the way to
go here.
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/110196] stacktrace_entry does not print which library the function was from
2023-06-09 15:42 [Bug libstdc++/110196] New: stacktrace_entry does not print origin binary filename fabian_kessler at gmx dot de
` (2 preceding siblings ...)
2023-06-09 16:07 ` fabian_kessler at gmx dot de
@ 2023-06-09 16:09 ` pinskia at gcc dot gnu.org
2023-06-09 16:16 ` fabian_kessler at gmx dot de
` (7 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: pinskia at gcc dot gnu.org @ 2023-06-09 16:09 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110196
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Severity|normal |enhancement
Summary|stacktrace_entry does not |stacktrace_entry does not
|print origin binary |print which library the
|filename |function was from
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/110196] stacktrace_entry does not print which library the function was from
2023-06-09 15:42 [Bug libstdc++/110196] New: stacktrace_entry does not print origin binary filename fabian_kessler at gmx dot de
` (3 preceding siblings ...)
2023-06-09 16:09 ` [Bug libstdc++/110196] stacktrace_entry does not print which library the function was from pinskia at gcc dot gnu.org
@ 2023-06-09 16:16 ` fabian_kessler at gmx dot de
2023-06-09 16:17 ` fabian_kessler at gmx dot de
` (6 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: fabian_kessler at gmx dot de @ 2023-06-09 16:16 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110196
--- Comment #4 from fabian_kessler at gmx dot de <fabian_kessler at gmx dot de> ---
Indeed, changing _S_init to the following did the trick:
```
static backtrace_state *_S_init() {
static backtrace_state *__state = []() {
auto getpath = []() -> std::string {
char buf[PATH_MAX + 1];
if (readlink("/proc/self/exe", buf, sizeof(buf) - 1) == -1) {
return "";
}
std::string str(buf);
return str.substr(0, str.rfind('/'));
};
auto exec_filename = getpath();
return backtrace_create_state(exec_filename.c_str(), 1, _S_err_handler,
nullptr);
}();
return __state;
}
```
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/110196] stacktrace_entry does not print which library the function was from
2023-06-09 15:42 [Bug libstdc++/110196] New: stacktrace_entry does not print origin binary filename fabian_kessler at gmx dot de
` (4 preceding siblings ...)
2023-06-09 16:16 ` fabian_kessler at gmx dot de
@ 2023-06-09 16:17 ` fabian_kessler at gmx dot de
2023-06-09 16:18 ` fabian_kessler at gmx dot de
` (5 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: fabian_kessler at gmx dot de @ 2023-06-09 16:17 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110196
--- Comment #5 from fabian_kessler at gmx dot de <fabian_kessler at gmx dot de> ---
Indeed, changing _S_init to the following improved it:
```
static backtrace_state *_S_init() {
static backtrace_state *__state = []() {
auto getpath = []() -> std::string {
char buf[PATH_MAX + 1];
if (readlink("/proc/self/exe", buf, sizeof(buf) - 1) == -1) {
return "";
}
std::string str(buf);
return str.substr(0, str.rfind('/'));
};
auto exec_filename = getpath();
return backtrace_create_state(exec_filename.c_str(), 1, _S_err_handler,
nullptr);
}();
return __state;
}
```
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/110196] stacktrace_entry does not print which library the function was from
2023-06-09 15:42 [Bug libstdc++/110196] New: stacktrace_entry does not print origin binary filename fabian_kessler at gmx dot de
` (5 preceding siblings ...)
2023-06-09 16:17 ` fabian_kessler at gmx dot de
@ 2023-06-09 16:18 ` fabian_kessler at gmx dot de
2023-06-09 16:47 ` fabian_kessler at gmx dot de
` (4 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: fabian_kessler at gmx dot de @ 2023-06-09 16:18 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110196
--- Comment #6 from fabian_kessler at gmx dot de <fabian_kessler at gmx dot de> ---
output is now
```
0# main at /mnt/c/Users/Febbe/workspace/test/test/main.cpp:7
1# __libc_start_call_main at ../sysdeps/nptl/libc_start_call_main.h:58
2# __libc_start_main_impl at ../csu/libc-start.c:392
3# _start at :0
4#
```
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/110196] stacktrace_entry does not print which library the function was from
2023-06-09 15:42 [Bug libstdc++/110196] New: stacktrace_entry does not print origin binary filename fabian_kessler at gmx dot de
` (6 preceding siblings ...)
2023-06-09 16:18 ` fabian_kessler at gmx dot de
@ 2023-06-09 16:47 ` fabian_kessler at gmx dot de
2023-06-09 20:56 ` redi at gcc dot gnu.org
` (3 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: fabian_kessler at gmx dot de @ 2023-06-09 16:47 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110196
--- Comment #7 from fabian_kessler at gmx dot de <fabian_kessler at gmx dot de> ---
To fix this some sort of
char ** __backtrace_symbols (void *const *array, int size)
must be used internally.
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/110196] stacktrace_entry does not print which library the function was from
2023-06-09 15:42 [Bug libstdc++/110196] New: stacktrace_entry does not print origin binary filename fabian_kessler at gmx dot de
` (7 preceding siblings ...)
2023-06-09 16:47 ` fabian_kessler at gmx dot de
@ 2023-06-09 20:56 ` redi at gcc dot gnu.org
2023-06-09 21:38 ` redi at gcc dot gnu.org
` (2 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: redi at gcc dot gnu.org @ 2023-06-09 20:56 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110196
Jonathan Wakely <redi at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Ever confirmed|0 |1
Status|UNCONFIRMED |ASSIGNED
Assignee|unassigned at gcc dot gnu.org |redi at gcc dot gnu.org
Last reconfirmed| |2023-06-09
--- Comment #8 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to fabian_kessler@gmx.de from comment #3)
> Also, backtrace_create_state is ment to be called just once per execution,
> because it allocates unfreeable memory. Maybe, a singleton would be the way
> to go here.
It is only called once, isn't it?
(In reply to fabian_kessler@gmx.de from comment #5)
> Indeed, changing _S_init to the following improved it:
>
> ```
> static backtrace_state *_S_init() {
>
> static backtrace_state *__state = []() {
> auto getpath = []() -> std::string {
> char buf[PATH_MAX + 1];
> if (readlink("/proc/self/exe", buf, sizeof(buf) - 1) == -1) {
This isn't portable, not all systems have /proc/self/exe, or readlink.
I think it would be better to move this function into the library, and use
std::filesystem::read_symlink.
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/110196] stacktrace_entry does not print which library the function was from
2023-06-09 15:42 [Bug libstdc++/110196] New: stacktrace_entry does not print origin binary filename fabian_kessler at gmx dot de
` (8 preceding siblings ...)
2023-06-09 20:56 ` redi at gcc dot gnu.org
@ 2023-06-09 21:38 ` redi at gcc dot gnu.org
2023-06-10 10:21 ` fabian_kessler at gmx dot de
2023-06-10 12:57 ` fabian_kessler at gmx dot de
11 siblings, 0 replies; 13+ messages in thread
From: redi at gcc dot gnu.org @ 2023-06-09 21:38 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110196
--- Comment #9 from Jonathan Wakely <redi at gcc dot gnu.org> ---
libbacktrace is supposed to have code to do this, see fileline_initialize in
libbacktrace/fileline.c
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/110196] stacktrace_entry does not print which library the function was from
2023-06-09 15:42 [Bug libstdc++/110196] New: stacktrace_entry does not print origin binary filename fabian_kessler at gmx dot de
` (9 preceding siblings ...)
2023-06-09 21:38 ` redi at gcc dot gnu.org
@ 2023-06-10 10:21 ` fabian_kessler at gmx dot de
2023-06-10 12:57 ` fabian_kessler at gmx dot de
11 siblings, 0 replies; 13+ messages in thread
From: fabian_kessler at gmx dot de @ 2023-06-10 10:21 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110196
--- Comment #10 from fabian_kessler at gmx dot de <fabian_kessler at gmx dot de> ---
> It is only called once, isn't it?
Actually, _S_init is even called in a loop. And the singleton pattern is also a
code extension from me. Therefore, backtrace_create_state is also called in a
loop.
>
> This isn't portable, not all systems have /proc/self/exe, or readlink.
>
> I think it would be better to move this function into the library, and use
> std::filesystem::read_symlink.
Yes, that code snipped was just to test my assumption, which was only partially
correct. Adding the path improved the stacktrace, but did not add the
originating library.
I don't feel experienced, to contribute STL code for more than x86_64.
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/110196] stacktrace_entry does not print which library the function was from
2023-06-09 15:42 [Bug libstdc++/110196] New: stacktrace_entry does not print origin binary filename fabian_kessler at gmx dot de
` (10 preceding siblings ...)
2023-06-10 10:21 ` fabian_kessler at gmx dot de
@ 2023-06-10 12:57 ` fabian_kessler at gmx dot de
11 siblings, 0 replies; 13+ messages in thread
From: fabian_kessler at gmx dot de @ 2023-06-10 12:57 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110196
--- Comment #11 from fabian_kessler at gmx dot de <fabian_kessler at gmx dot de> ---
With the following code extension, I get what I want, but this may also be
platform dependent:
class stacktrace_entry{
// ...
std::string _symbol(uintptr_t const pc) const {
Dl_info info;
int status;
link_map *map;
status = dladdr1((void const *)(pc), &info, reinterpret_cast<void
**>(&map),
RTLD_DL_LINKMAP);
return std::string(info.dli_fname ?: "");
}
// ...
}
// ...
inline std::ostream &operator<<(std::ostream &__os,
const stacktrace_entry &__f) {
std::string __desc, __file;
int __line;
if (__f._M_get_info(&__desc, &__file, &__line)) {
__os.width(4);
__os << __f._symbol(__f._M_pc) << "(" << __desc << ") at " << __file << ':'
<< __line;
}
return __os;
}
Outputs:
0# /mnt/c/Users/Febbe/workspace/test/build/test_stacktrace_20(main) at
/mnt/c/Users/Febbe/workspace/test/test/main.cpp:7
1# /lib/x86_64-linux-gnu/libc.so.6(__libc_start_call_main) at
../sysdeps/nptl/libc_start_call_main.h:58
2# /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main_impl) at
../csu/libc-start.c:360
3# /mnt/c/Users/Febbe/workspace/test/build/test_stacktrace_20(_start) at :0
^ permalink raw reply [flat|nested] 13+ messages in thread