Re: svn 260311 breaks gcc builds on releng9 ?

On 17.01.2014 16:07, Luigi Rizzo wrote:
>
>
>
> On Fri, Jan 17, 2014 at 12:50 PM, Pedro Giffuni <pfg_at_freebsd.org 
> <mailto:pfg_at_freebsd.org>> wrote:
>
>     Hello Luigi;
>
>
>     On 17.01.2014 15:18, Luigi Rizzo wrote:
>
>         Hi,
>         I am seeing an odd problem which seems to be triggered by
>         svn260311
>
>         I have two machines running snapshots of stable/9 from last fall
>         (one 255898 sep.26, the other 258126 nov.14). All is amd64
>
>         Build a recent head (260311 and newer) with gcc fails on the
>         sep.26 machine:
>
>                 ...
>                 19:51:22 ===> gnu/usr.bin/cc/cc1plus (all)
>                 PROG CC      
>         /media/bsd10/usr/home/luigi/FreeBSD/head/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/main.c
>                 ...
>                 PROG CC      
>         /media/bsd10/usr/home/luigi/FreeBSD/head/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/except.c
>                 parser.o: In function
>         `cp_parser_objc_tentative_protocol_refs_opt':
>                 parser.c:(.text+0x14665): undefined reference to
>         `cp_objc_protocol_id_list'
>                 *** [cc1plus-dummy] Error code 1
>                 1 error
>
>         but works on the newer one.
>
>         svn 260310 builds on both.
>
>         I am a bit unclear on what is going on because the missing symbol
>         (cp_objc_protocol_id_list) only appears in the source file
>         parser.c
>         and in a ChangeLog.apple file, so maybe it is generated and
>         the actual failure is at some earlier stage.
>
>     Thanks for the report.
>
>     Of course gcc can also have bugs but it would look like you may
>     not be doing a clean build or that some how object code from two
>     different builds got mixed up.
>
>
> i thought so too, but i am building in a completely new dir
> and i retried some 10 times rm -rf on the output dir,
> resyncing the repo and checking diffs, etc.
>
> Also it is very suspicious that going back to 260310 makes
> the problem disappear, and 260311 brings it back in.
>
> Anyways, it is very confusing indeed.
>
>
>
>     JIC, the cp_obj_protocol_id_list stuff is obviously objc stuff
>     that we don't need at all and that crept in along with other
>     changes from Apple so I will remove it to see if it helps somehow.
>
>
> ok thanks.
> let me know if you want me to test a patch before committing,
> since I have a way to reproduce the error (i am not upgrading
> the offending system on purpose).
>

The Apple change was incomplete so I guess the compiler in head is not 
being strict enough. I removed the dead code and rebuilt.
Should be fixed as r260831.

Thank you for the report!

Pedro.