Thread: Re: Fix potential overflow risks from wcscpy and sprintf

On 06.06.25 22:50, Yan Haibo wrote:
> This change stems from a recent static code analysis, which identified a 
> minor potential overflow issue. I would appreciate it if someone could 
> review the fix at their convenience.

Please provide more detail in each case what the issue is and how you 
are proposing to fix it.

Note that PostgreSQL code is regularly run through various static 
analysis tools, so if the issue were super-obvious, it would probably 
have been noticed by now.

Yan Haibo <haibo.yan@hotmail.com> writes:
> Thank you. Peter. It seems the patch may have been lost during our earlier communication, so I¡¯ve reattached it
here.
> I hope it comes through correctly this time.

Thanks for the patch.

Using wcsncpy in search_locale_enum() seems fine, assuming it exists
on Windows (note that code is Windows-only, possibly explaining why
we've not seen other static-analysis reports).  I doubt there's any
actual bug there, since we're relying on Windows' own
LOCALE_NAME_MAX_LENGTH constant; but I agree the chain of reasoning
is kind of long.  (But shouldn't you write LOCALE_NAME_MAX_LENGTH
not LOCALE_NAME_MAX_LENGTH - 1?)

I'm unexcited about the guc.c changes.  There is visibly no bug
there.  The only reason to change it would be if we were going
to introduce a strict project policy against using sprintf(),
which we're not likely to given that there are hundreds of other
occurrences in our code base.  I don't see a reason to think
that these three are more pressing than the others.

            regards, tom lane