On Thu, Aug 11, 2016 at 2:36 PM, Jakub Jelinek <ja...@redhat.com> wrote: > Hi! > > On Wed, Aug 10, 2016 at 12:23:06PM +0200, Richard Biener wrote: >> > Introducing another attribute that does the same thing as existing >> > attribute >> > would be way too ugly. In theory the reference class could be added to >> > DW_AT_string_length, I can ask on DWARF workgroup (but it might be too late >> > for DWARF 5), but that still doesn't solve the issue of the indirect >> > params. >> > >> > How do you want to handle the debug info without ammending the >> > early-generated >> > DWARF though? Just by using it as abstract origins of everything and >> > ammending in copies? >> >> Yes. > > I've filed an enhancement request and got one positive feedback, but it is > going to be multiple months at best. > So, at least for non-LTO I'd strongly prefer to go with what the current > patch does, and for LTO we'd then have to ask GDB to implement the reference > class for DW_AT_string_length and then just use that instead, depending on > flag_lto or similar, or perhaps for dwarf4 and earlier don't emit for LTO > variable string length and keep it only for dwarf5+ (if the change is > approved). For the indirect parms and LTO, I guess we'd need to create some > artificial VAR_DECL at function scope with DECL_VALUE_EXPR of *parm, > DECL_ARTIFICIAL, DECL_NAMELESS and reference that instead of the parm > itself. > With this, do you object to the current patch?
No, I still hope it avoids generating references to possibly optimized out stuff early - I didn't thorougly review it. I guess I'll find out soon ;) Richard. > Jakub
