Change History (11)
comment:1 by , 19 years ago
Type: | task → enhancement |
---|
comment:2 by , 19 years ago
Looks that way - I've tested this a few times myself, and never had a problem with headers in the wrong place (never installed more than C and C++ compilers though).
comment:3 by , 19 years ago
Looks like it only affects --enable-libgcj. So, this is more of a BLFS issue. I'm guessing a simple sed on libjava/Makefile.in will take care of it.
comment:4 by , 19 years ago
http://kerneltrap.org/node/6271 also discusses the Java failure. http://gcc.gnu.org/gcc-4.1/changes.html discusses changes to the ABI and memory model.
Note a recent tar is required but already in the book. Note a more recent binutils 2.16.91 is required but not already in the book.
I do not fully understand the implications of the memory model change on attempting to run precompiled code (e.g. Nvidia closed source driver). However because of this change I encourage treading very carefully into 4.1.0.
comment:5 by , 19 years ago
binutils-2.16.91 needed? I'm not sure about that. The changes file indicates that "Binutils 2.16.91 or newer are required for compiling medium model now". You can still build a functional gcc-4.1 with binutils-2.16.1. DIY-Linux (http://www.diy-linux.org/x86-reference-build/) has this arrangement.
comment:6 by , 19 years ago
Milestone: | → Future |
---|
comment:7 by , 19 years ago
Type: | enhancement → task |
---|
comment:9 by , 18 years ago
Milestone: | Future → 6.3 |
---|
comment:10 by , 18 years ago
Resolution: | → fixed |
---|---|
Status: | new → closed |
For anyone wanting to test this version out, please watch out for headers being installed in the wrong directory as described in http://gcc.gnu.org/ml/gcc/2006-02/msg00720.html. I think we're pretty safe from this, as it only looks to be Java stuff that's not installed correctly.