This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [lto] PATCH: fill in code to merge declarations


Mark Mitchell wrote:

else
DECL_EXTERNAL (result) = 1;
+ if (inlined == DW_INL_declared_inlined
+ || inlined == DW_INL_declared_not_inlined)
+ DECL_DECLARED_INLINE_P (result) = 1;
+

Shouldn't the second case be DW_INL_inlined? But, I think you only want the first case -- the compiler is always free to inline things anyhow, and the bit you're setting is about what the user has explicitly declared.

No, I double-checked this against the DWARF spec and dwarf2out.c, and it is correct as written. DW_INL_declared_not_inlined means it was declared inline by the user, although not actually inlined by the compiler. Should I be setting DECL_INLINE as well to indicate whether or not the function is actually inlined, or is that something that is computed by some later compiler pass?


-Sandra


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]