Hi Aaron, On Wed, Aug 27, 2025 at 06:51:36PM -0400, Aaron Merey wrote: > On Tue, Aug 26, 2025 at 6:46 PM Frank Ch. Eigler <f...@redhat.com> wrote: > > > RFC: Is the option name --ar-member clear? Do we also want a variant > > > (--all-ar-members) that checks whether all members match? Do we want > > > this to work recursively? Should it print the member names that match > > > instead of the archive file (is that useful?). > > > > Suggest "--any-ar-member" to clarify the name, and I wouldn't bother > > with the other things. > > I second this.
OK. Lets go with that name then. > Also I noticed the following gcc error when building with the patch applied: > > ‘__builtin___sprintf_chk’ argument 5 overlaps destination object > ‘current_path’ [-Werror=restrict] > [...] > elfclassify.c: In function ‘check_ar_members’: > elfclassify.c:47:14: note: destination object referenced by > ‘restrict’-qualified argument 1 was declared here > 47 | static char *current_path; > > gcc thinks one of the arguments to sprintf (ar_path) is an alias of > the destination string (current_path). It's a false positive since > current_path is reassigned before the sprintf call but the code can be > adjusted to avoid this error. That is odd. Which gcc version is this with and which configure options? Locally (GCC 15.2.1) seems fine. It clearly is a false positive indeed. I'll try to work around it by using some local temporary. Thanks, Mark
