[Bug c++/110051] New: error: writing 1 byte into a region of size 0

[Bug c++/110051] New: error: writing 1 byte into a region of size 0

[pkurash at cern dot ch]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110051

            Bug ID: 110051
           Summary: error: writing 1 byte into a region of size 0
           Product: gcc
           Version: 13.1.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: pkurash at cern dot ch
  Target Milestone: ---

Created attachment 55220
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=55220&action=edit
build log

Hello,

I am compiling some software with use of GCC, there are many errors which are
not dependent on the code being compiled, since the logs seem to point at the
errors in the files within /usr directory:

/usr/lib/gcc/x86_64-pc-linux-gnu/13/include/g++-v13/bits/stl_algobase.h:398:17:
error: writing 1 byte into a region of size 0 [-Werror=stringop-overflow=]
  398 |         { *__to = *__from; }
      |           ~~~~~~^~~~~~~~~

I use Gentoo Linux, my current gcc version  is 13.1.1_p20230520 which I
installed with the system package manager (portage).

I also attach a full log of compilation of a package (I need to build multiple
packages in a chain, when I skip this one there arise similar errors at
compilation process of subsequent packages).

[Bug c++/110051] error: writing 1 byte into a region of size 0

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110051

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |diagnostic

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
The diagnostics are from template instantiations of the standard template
library which we now consider useful to the user.  They will necessarily
point to implementation details of that library, a better reporting is
difficult to implement.

[Bug c++/110051] error: writing 1 byte into a region of size 0

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110051

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2023-05-31
             Status|UNCONFIRMED                 |WAITING
     Ever confirmed|0                           |1

--- Comment #2 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to Felix from comment #0)
> I am compiling some software with use of GCC, there are many errors which

Those are warnings. You requested to turn them into errors with -Werror.

> are not dependent on the code being compiled, since the logs seem to point
> at the errors in the files within /usr directory:

This is an incorrect assumption.

If you write the following code, you'll get diagnostics pointing to the
standard library code, but the bug is still in your code:

  char s1[] = "abcdefghij";
  char s2[5];
  std::copy(std::begin(s1), std::end(s1), s2);

In your specific case, it might be a false positive warning. But you can't
assume that just because the diagnostics show code in the std::lib.

The build log is useless though. All it shows is there are some warnings,
nobody can tell if they're valid or not without the code. Please read
https://gcc.gnu.org/bugs again and provide the requested info. Also please
search the (many!) existing bugs about -Wstringop-overflow warnings and check
if yours is another duplicate of one of those.

[Bug tree-optimization/110051] error: writing 1 byte into a region of size 0

[pkurash at cern dot ch]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110051

--- Comment #3 from Felix <pkurash at cern dot ch> ---
Yes, you're right, the warning is turned to error.

Without this option the compilation runs normally

[Bug tree-optimization/110051] error: writing 1 byte into a region of size 0

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110051

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |INVALID
             Status|WAITING                     |RESOLVED

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
No testcase in over 6 months so closing as invalid.