This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
Re: Java bootstrap failure for libjava/java/io/FilePermission.java
- To: bryce at albatross dot co dot nz, pfeifer at dbai dot tuwien dot ac dot at
- Subject: Re: Java bootstrap failure for libjava/java/io/FilePermission.java
- From: "Kaveh R. Ghazi" <ghazi at caip dot rutgers dot edu>
- Date: Mon, 29 Jan 2001 21:57:11 -0500 (EST)
- Cc: gcc-bugs at gcc dot gnu dot org, java at gcc dot gnu dot org
> From: Bryce McKinlay <bryce@albatross.co.nz>
>
> Gerald Pfeifer wrote:
>
> > /cvs/gcc/libjava/java/io/FilePermission.java: In method
> > `java.io.FilePermission.implies(java.security.Permission)':
> > /cvs/gcc/libjava/java/io/FilePermission.java:171: Internal error:
> > Segmentation fault.
>
> Yes, I got this too. It appears to be related to:
>
> 2001-01-28 Kaveh R. Ghazi <ghazi@caip.rutgers.edu>
>
> * stmt.c (estimate_case_costs): Use integer_minus_one_node.
>
> * tree.c (build_common_tree_nodes_2): Set integer_minus_one_node.
>
> * tree.h (tree_index): Add new element TI_INTEGER_MINUS_ONE.
> (integer_minus_one_node): Define.
I assume the segv was caused by an uninitialized integer_minus_one_node?
It appears that tree.c:build_common_tree_nodes_2 (and thus the
initialization of integer_minus_one_node) is only called in c-common.c.
So jc1 isn't setting it. Boom.
My bootstrap/testsuite run prior to installation didn't detect this
because java doesn't appear to be built and tested by default in a
solaris2 bootstrap. (At least that's what happens for me. Dunno why.
Is that expected these days or should I investigate?) In any case,
sorry for the regression.
So... the solution would be to revert the stmt.c change, or initialize
integer_minus_one_node in jc1. What do you prefer?
--Kaveh
--
Kaveh R. Ghazi Engagement Manager / Project Services
ghazi@caip.rutgers.edu Qwest Internet Solutions