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]

Re: forgotten C/OBJ parser update breaks cc1obj


On Sat, 18 Dec 1999 07:54:43 -0800, "David S. Miller" wrote:
>
>The following change:
>
>1999-12-17 13:21 -0800  Zack Weinberg  <zack@rabi.columbia.edu>
>
> ...
>
>	* c-parse.in (string action): Do not warn about ANSI string
>	concatenation in system headers.  Affects C parser only.
>	* c-parse.y, c-parse.c, c-parse.h: Rebuild.
>
>Indicates that only the C parser should be rebuilt, this is completely
>wrong because objc is now expecting different identifier codes than
>what the rest of the C front-end it shares does.

My apologies.  I went by the makefile, which didn't choose to rebuild
objc-parse.c, and I should have realized that something was wrong when
all the objc tests failed.

A corrected version has now been checked in.  I used bison 1.28 as this is
the only version I have and it shouldn't really matter - if policy is to use
1.25, please let me know.

[Right now the stage1 compiler - cc1 or cc1obj - ICEs in reload when asked
to compile *anything* with optimization, but I do not believe this is my
fault.  Symptoms:

c-torture/code_quality/920608-1.c: In function `bar':
c-torture/code_quality/920608-1.c:25: Internal compiler error in
	`refers_to_regno_for_reload_p', at reload.c:5722
Please submit a full bug report.
See <URL:http://www.gnu.org/software/gcc/faq.html#bugreport> for
instructions.]

zw


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