On 5/2/19 12:30 PM, Jakub Jelinek wrote: > On Mon, Apr 01, 2019 at 02:11:17PM +0200, Martin Liška wrote: >> 2019-04-01 Martin Liska <mli...@suse.cz> >> >> * gcc.c (process_command): Add dummy file only >> if n_infiles == 0. >> * opts-global.c (decode_options): Pass lang_mask. >> * opts.c (print_help): New function. >> (finish_options): Print --help if help_option_argument >> is set. >> (common_handle_option): Factor out content of OPT__help_ >> into print_help. >> * opts.h (finish_options): Add new argument. > > As reported by Ulrich Drepper on IRC, this broke > --enable-as-accelerator-for=... build, tree-streamer-in.c calls > finish_options too and that caller has not been adjusted.
You are right, it's guarded in #ifdef ACCEL_COMPILER so I haven't seen the compilation error. > > Any reason why you've called print_help from finish_options rather than > decode_options after it calls finish_options? Yes, I need that as I need option propagation, as described in: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=89885#c0 > > Or should tree-streamer-in.c pass CL_LTO as the lang_mask, or say CL_DRIVER > so that print_help does nothing in that case? Yes, that would fix it. Can you please provide a patch that he can test? Thanks, Martin > > Jakub >
