Andrew Benton wrote: > On Sat, 04 Feb 2012 09:49:37 -0800 Bryan Kadzban > <[email protected]> wrote: > >> Andrew Benton wrote: >>> On Fri, 03 Feb 2012 21:44:34 -0800 Bryan Kadzban >>> <[email protected]> wrote: >>> >>>> Do you get __USE_ISOC11 #define'd (to what value?) or >>>> #undef'ed? What about __cplusplus (again, what value)? >>>> >>> andy@eccles:/mnt/lfs/sources$ /tools/bin/x86_64-lfs-linux-gnu-gcc >>> -E -D_GNU_SOURCE -Wp,-dU /tools/include/stdio.h | grep >>> __USE_ISOC11 >> Ah, sorry, no -- g++, not gcc. :-) > > The first pass of gcc only compiles a C compiler, so there is no > x86_64-lfs-linux-gnu-g++
Bah, right. Well, there's a C++ compiler *somewhere*, that the pass2 gcc is using to compile libstdc++ when it runs into this error. :-) Maybe it'd be a better idea to do it that way: Run a build until it breaks, then repeat the command that fails (with all flags intact), adding -E and -Wp,-dU, removing -o <file>, and replacing any C or C++ source file with whatever stdio.h is being used. (Probably the one in /tools/include, but it might be worth double-checking. You can do that with another compiler run just adding the -E flag, and grepping through the output file for stdio. It'll be on a #file line somewhere.) The full output of that -E -Wp,-dU run is one of the few things that'll be useful for debugging this I think. (Along with whatever the command was, I think.) >> Actually, the full output will be useful as well. > > The full output of what? The full output of this: > /tools/bin/x86_64-lfs-linux-gnu-gcc -E -D_GNU_SOURCE -Wp,-dU > /tools/include/stdio.h > dump.txt is here: > http://www.linuxfromscratch.org/~andy/dump.txt Yeah, that's still C mode, not C++ mode (which is where the change is supposed to have happened). :-) I think the info above should help.
signature.asc
Description: OpenPGP digital signature
-- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
